Chapter 1 — Definition, Landscape and Stakes

An offer is the specific, purchasable promise a business puts in front of a customer — what is done, in which package, with which commitments, on which terms — and designing it deliberately is how a strategy becomes something a customer can actually buy.

What You'll Be Able to Do

• Define an offer as a specific, purchasable combination of what is done, what is promised, what is added, what the customer sees and the terms on which it is sold.
• Separate the offer from its neighbors in this library — the value proposition of Book A1, the positioning of Book A3, the price of Book A7 and the message of Book B6 — and say what question each of them answers.
• Name the five layers of an offer and place a business's own elements in them, from the core at the center to the terms on the outside.
• Take an offer inventory in five steps, dating and costing every element the customer receives, including the ones that were given without ever being written down.
• Recognize the three signals of an undesigned offer — who says no and why, what customers ask for that nothing in the offer contains, and what the business delivers without pricing it.
• State the rules that frame what an offer may promise in Canada and Québec, and know when a guarantee, a warranty or an advertising claim brings a legal obligation with it.

Why It Matters

A strategy is a decision about where to play and how to win, and it earns nothing until a customer can buy it. The distance between the two is the offer. A business can hold a clear position, serve a well-chosen segment and understand its customer better than any competitor, and still lose the sale, because what it puts in front of the buyer is a description of its capabilities rather than a thing that can be accepted, refused, compared or signed. The offer is the point at which everything the business has decided becomes a document with a scope, a commitment and a set of terms, and it is the last place where a good strategy can be lost cheaply.

The business followed through this book shows how quietly an offer can drift away from the strategy that produced it. Nordec Entretien inc. provides preventive maintenance and on-call technical service to commercial buildings on Montréal's north shore, and one year after it signed the strategy of Book A3 it holds 46 buildings under contract, an average contract of $712 a month and a contract run-rate of $393,024 a year. Its renewal rate is 95.5%, 96% of its urgent contract calls are answered on site within ninety minutes, and every contract building has a building health record. Those are the numbers of a business that is working. Underneath them, the contract has grown from two elements to six, each addition was given free, and the cost of serving a building in a territory has risen from $282.73 a month to $341.23 while the price rose from $691 to $712. The margin of a territory building fell from 59.08% to 52.08%, and $20,700 a year of added cost was never priced at all.

That is what is at stake in this book, and it is not an accounting problem. The offer decides who can say yes: a contract designed for a 4,500-square-metre building is more than a 2,500-square-metre building needs, and two of the four proposals Nordec lost on price went to the trades billed by the call, both of them small buildings whose owners said the contract was more than the building required. The offer decides what the business must deliver: every commitment written into it becomes an obligation with a cost, and every element delivered without being written becomes a habit nobody can price. And the offer decides what the business earns per hour of the work it sells, because the cost to serve is set by what the offer contains. Designing the offer deliberately is how a business takes control of all three at once, and it is the subject of the twelve chapters that follow.

In This Chapter

Section 1.1 defines the offer and separates it from the value proposition, the positioning, the price, the message and the proposal. Section 1.2 examines the mixture of goods and services inside every offer and the argument that nothing is a commodity. Section 1.3 sets out the five layers of an offer — the core, the promise, the supplements, the evidence and the terms — and shows where offers grow without anyone deciding. Section 1.4 describes the landscape of 2026: goods, services and experiences, tiers, subscriptions, productized services and the uneven spread of guarantees. Section 1.5 sets out what is at stake for a small business and the symptoms of an offer nobody designed. Section 1.6 gives the vocabulary the rest of the book uses. Section 1.7 states the rules that frame what an offer may promise in Canada, Québec and the United States. Section 1.8 introduces the case and the plan of Chapters 2 to 12.

1.1 What an Offer Is, and What It Is Not

Every business sells something, and almost every business describes what it sells in a sentence that is too large to buy. "We do commercial maintenance." "We build websites." "We advise small companies on their finances." Those sentences are true, and none of them is an offer. An offer is the narrower, harder and more useful thing: the specific combination a customer can accept, at stated terms, for a stated amount, on a stated day. It says what will be done, what the customer will receive, what the business commits to, what is excluded, how long it lasts and what happens at the end. A business that can say what it does but cannot say what a customer buys has a capability, not an offer.

The distinction matters because the two are confused constantly, and the confusion is expensive in a particular way. A capability is described from the inside, in the language of the work. An offer is described from the outside, in the language of the purchase. The technician who says "we look after the mechanical systems" is describing a capability. The contract that says "a scheduled inspection of the mechanical, plumbing and door systems once a month, urgent calls answered on site within ninety minutes day and night, up to two hours of work included in each urgent call, and work beyond that scope billed at a stated hourly rate plus parts" is an offer. Both statements are honest. Only the second one can be accepted, refused, compared with an alternative, or priced.

A written offer contains four things, and a business can test its own offer by looking for each of them. The first is what it is: the thing being sold, named plainly enough that a buyer could repeat it to a colleague. The second is what the customer receives and what the customer does not receive, because an offer that lists only what is included leaves the buyer to guess at the boundary, and the guess is always more generous than the contract. The third is the commitments: the response times, the inclusions, the guarantees, the standards the business binds itself to. The fourth is the terms: the price, the duration, the renewal, the notice, the exclusions and the way out. An offer missing any of the four is not a small offer; it is an incomplete one, and the missing part will be supplied by the customer's expectation rather than by the business's intention.

The offer is not the value proposition. Book A1 (From Idea to Business Model) treats the value proposition as the answer to a larger question: what does this business sell, to whom, and why is it better for them than the alternatives? That answer can be true of a business that has never written an offer, and it can remain true while the offer changes several times. The value proposition is the reason a customer should care about the business at all. The offer is the thing the customer is handed when they decide to act on that reason.

The offer is not the positioning either. Book A3 (Strategy and Competitive Positioning) chose where Nordec would play — mid-size buildings of 2,000 to 9,000 square metres, owner-operators and small portfolios, in two territories — and how it would win there, with three deliberate refusals and a value curve that raised response time, knowledge of the building and monthly reporting. Positioning is a decision about the shape of the business. The offer is the instrument that carries that decision to a buyer, and a positioning can be sound while the offer contradicts it, which is exactly what happens when a business designs a contract for the segment it wants and then sells it to a building it did not choose.

The offer is not the price. Book A7 (Pricing Strategy and Revenue Models) puts a number on what this book designs, and the two books meet at a single seam: this book decides what is inside the offer and what it costs to deliver, and Book A7 decides what the customer pays for it. The seam matters because the two are usually discussed together and decided separately. Nordec's tier prices were set with the method of Book A7; what each tier contains, what it excludes and what it promises is designed here. A business that changes a price without changing the contents, or changes the contents without revisiting the price, is moving one half of a pair.

The offer is not the message, and it is not the proposal. Book B6 (Copywriting and Persuasive Communication) and Book B1 (Marketing Fundamentals for Business Owners) put the offer into words for a market, and the words can be excellent while the offer underneath them is vague. Book E5 (Proposals, Quotes, Business Cases and Contracts) puts the offer in front of one buyer, with terms and a signature line, and a proposal is the offer in its most formal dress. The order is fixed: the offer is designed first, and the message and the proposal carry it. A business that writes the message first is writing about something that does not yet exist.

The following table places the offer beside its five neighbors in this library. Read it as a map of questions rather than as a list of documents: each row names a concept, the single question it answers, and the book where that question is treated in full. The offer's row is the one this book occupies, and the four rows around it are the ones a reader will most often confuse with it.

Table 1.1 — The offer and its neighbors in this library

DEFINITION Offer, core, supplement, promise and terms

An offer is the specific, purchasable combination a business puts in front of a customer: what is done, what the customer receives and does not receive, the commitments the business makes, and the terms on which the whole thing is sold. The core is what is actually done for the customer, the work itself. The promise is what the business commits to about that work — a response time, an inclusion, a guarantee — and it is what turns a capability into an obligation. A supplement is a service delivered around the core, such as a report, a record or a planning note, and it is where most offers grow without anyone deciding. The evidence is what the customer sees and keeps: the document, the visit sheet, the report, the arrival of the technician at the promised hour. The terms are the price, the duration, the exclusions, the renewal and the way out. The five together are the offer, and a customer buys all five whether or not the business has written them down.

1.2 Product, Service and Offer

Almost no offer is purely a good or purely a service, and the habit of sorting businesses into one category or the other hides the thing that matters. A restaurant sells food and the evening around it. A software business sells a license and the support behind it. A maintenance contractor sells a technician's hours and the knowledge of a building that makes those hours short. The useful question is not whether an offer is a product or a service, but how much of it is handed over and how much of it is performed with the customer present.

A service is performed rather than delivered, and that single difference changes what an offer can promise. A good can be inspected before it is bought, held in inventory, returned and resold. A service is produced at the moment it is consumed, in the presence of the customer or on the customer's property, and it cannot be stored. A maintenance visit that is late cannot be replaced by an earlier one. A report that arrives after the meeting it was written for has no value at all. This is why the promise layer of an offer carries more weight in a service than in a good: the customer cannot examine the thing before buying it, so the offer has to carry the assurance instead.

Theodore Levitt argued in 1980 that there is no such thing as a commodity (Levitt, 1980). The article's thesis was that all goods and services can be differentiated, and that industrial products and services offer more room for differentiation than consumer goods, not less. That argument is the foundation of this book. A business that believes it sells a commodity will compete on price, because price is the only variable it has left; a business that looks at the same work and finds the elements a customer would pay for has an offer to design.

The scale of the question is easy to underestimate. Services are about 80% of United States GDP, and yet innovation in services has been less disciplined than innovation in manufacturing (Bitner et al., 2008). The authors make that observation as the reason for their work on service blueprinting, and it applies with full force to the offer. A manufacturer designs a product deliberately, with drawings, tolerances and a specification. A service business often assembles its offer by responding to requests, and the result is a set of commitments that nobody chose and nobody can describe in one page.

The link between the customer and the offer is the value proposition canvas, which places a customer profile — the jobs customers try to get done, their pains and their gains — beside a value map of the pain relievers and gain creators an offer provides, and asks whether the two fit (Osterwalder et al., 2014). The canvas is useful here for one reason above the others: it insists that the fit be tested before an expensive launch, and it treats the value proposition as something embedded in a profitable business model rather than as a slogan. Chapter 2 of this book uses the same logic in the form of value interviews, and Chapter 8 uses it in the form of a pilot.

KEY CONCEPT The customer buys the whole offer, not the service inside it

A customer never buys a service in the abstract. The customer buys a package of things at once: the work that will be done, the commitments attached to it, the services delivered around it, the documents and moments that make it visible, and the terms that govern the whole arrangement. A maintenance contract is not bought because a technician visits; it is bought because the visit happens on a schedule, because an urgent call is answered within a stated time, because a report arrives each month, because a note tells the owner what will need replacing, and because the price and the exclusions are known in advance. When a business improves one of those and leaves the others alone, the customer notices the whole. When a business adds one of them without deciding what it is worth, the customer receives it and the business pays for it. The practical consequence is that an offer is designed as a set, and that a change to any layer is a change to the offer.

1.3 The Layers of an Offer

This book uses a single framework for the contents of an offer, and it has five layers. The framework is this book's own; it is not attributed to any source, and it exists to give a business a place to put each element it delivers. The layers run from the inside out, and each one is a different kind of decision.

The core is what is done. It is the work itself, the reason the customer is buying: the monthly preventive visit and the urgent calls in Nordec's contract, the haircut in a salon, the audit in an accounting practice. The core is where a business usually starts, and it is the layer most businesses can describe.

The promise is what the business commits to about the core. Nordec's promise is a ninety-minute response on urgent contract calls, day and night, with up to two hours of work included in each call. The promise is what turns the core into an obligation, and it is the layer where the cost of an offer is most easily underestimated, because a commitment made in a sales conversation becomes a cost in every month that follows.

The supplements are the services delivered around the core. Nordec's supplements are the building health record, the one-page monthly report and the capital planning note. Supplements are where offers grow without anyone deciding, because each one is small, each one is useful, and each one is usually given away before anyone has asked what it costs. Chapter 6 treats them in full.

The evidence is what the customer sees and keeps. It is the visit sheet emailed after each visit, the monthly report, the capital planning note, the technician arriving inside the promised window, the invoice that matches the quote. Evidence is the layer a business most often neglects, because it is not work and it feels like administration. It is also the layer that decides whether the customer believes the other four exist.

The terms are the price, the duration, the exclusions, the renewal and the way out. Nordec's contract excludes parts, work beyond two hours per incident, roof and envelope, elevators, and anything requiring a licensed trade the company does not hold. The terms are where an offer is bounded, and a boundary that is not written is a boundary that will be argued about later.

Figure 1.1 shows the five layers as five concentric rings. At the center sits the core, the monthly preventive visit and the urgent calls. Around it sits the promise, the ninety-minute response and the two included hours. Around that sit the supplements, the building health record, the one-page monthly report and the capital planning note. Around that sits the evidence, the visit sheet, the report, the note and the technician's arrival. On the outside sit the terms, the price, the duration, the exclusions and the renewal. Each ring is labeled with the example of a maintenance contract, and the figure is read from the center outward: the customer buys the outermost ring, and everything inside it is what the business must deliver to honor it.

Figure 1.1 — The five layers of an offer. The core is what is done; the promise, the supplements, the evidence and the terms are what the customer actually buys with it, and where an offer grows without anyone deciding.

The reason the framework earns its place is that the outer three layers are where offers drift. A core rarely changes by accident; a business knows when it has started doing something new. A promise changes slowly, as a response time is relaxed in a busy month and never restored. A supplement is added in an afternoon because a customer asked and the answer was yes. Evidence is added because a technician started leaving a sheet behind and the customers liked it. Terms are changed by a sentence in an email. None of those five events is a decision about the offer, and together they are the offer.

METHOD An offer inventory in five steps

First, list every element the customer receives, including the ones given without being written: the visit, the response commitment, the report, the record, the note, the sheet left behind, the phone call returned. Second, date each element, because the date tells the business whether it was part of the original design or an addition made later, and an addition made later is the one most likely to be unpriced. Third, cost each element per customer per month, using the loaded cost of the people who deliver it and the consumables it consumes, so that every element has a number beside it. Fourth, mark what is priced and what was given, and total the two columns separately, because the difference between them is the amount the business is spending without having decided to. Fifth, write the three signals: who says no and for what reason, what customers ask for that nothing in the offer contains, and what the business delivers that no line prices. Deliverable: one dated, costed inventory of the offer. Pitfall: listing only what is written in the contract, which leaves out the elements the customer actually values most.

1.4 The Landscape in 2026

The forms an offer can take have widened, and a business choosing a form in 2026 is choosing among more options than its counterpart twenty years ago. The clearest account of the widening is the one Pine and Gilmore gave in 1998, tracing a birthday cake through the economy: the ingredients bought as commodities, a cake mix bought as a good, a cake bought from a bakery as a service, and a party staged by a business as an experience (Pine & Gilmore, 1998). Their argument is that an experience occurs when a company uses services as the stage and goods as props to engage a person in a memorable event, and that it is a fourth kind of economic offering, after commodities, goods and services. The cake is the same cake in all four cases; what changes is what the customer is buying.

Most small businesses now sell a mixture of the four without having chosen the mixture. A contractor sells a service and stages a small experience in the way the technician arrives, greets the tenant and leaves the work area. A retailer sells goods and adds a service at the counter. A consultant sells a service and produces documents that function as goods. The useful reading of the experience argument is not that every business should stage parties; it is that the customer's memory of the encounter is part of what was bought, and that the memory can be designed.

Three forms dominate the offers a small business is likely to consider in 2026. The first is the tiered offer, in which the same core is sold at two or three levels of inclusion, so that a customer can choose the level that fits. The second is the subscription, in which the offer is sold as a recurring period rather than a one-time purchase, which changes the business's cash flow and its obligation to keep delivering. The third is the productized service, in which a service that was previously scoped for each customer is defined once, with a fixed scope, a fixed price and a fixed duration, so that it can be sold repeatedly without redesign. Chapter 4 treats tiers, bundles and add-ons; Chapter 7 treats productized services and their mirror image, the servitized product.

The spread of guarantees across the landscape is uneven, and the unevenness is an opportunity rather than a fact to accept. Products such as cars, cameras and toaster ovens come with warranties, while services such as auto repair, wedding photography or catered dinners almost never do, even though a service cannot be inspected before it is bought (Hart, 1988). The author's argument is that the strongest guarantees are unconditional, meaningful to the customer, easy to understand and to invoke, and easy to collect, and that a well-designed guarantee forces the business to define good service in the customer's terms, builds a customer-oriented culture and produces the failure data the business needs to improve. Chapter 5 of this book treats guarantees, service levels and risk reversal in full.

The population of businesses that will read this book is large and its survival curve is steep. Canada had 1,079,188 small businesses in December 2024, 98.2% of employer businesses, employing 5.8 million people, or 46.6% of the private-sector labor force; survival was 68.0% at five years, 48.2% at ten and 24.2% at twenty-one (Innovation, Science and Economic Development Canada, 2026a). The same source defines a small business as one with 1 to 99 paid employees and an SME as one with 1 to 499. Those figures are the reason this book is written for the small end of the market and not for the large end: the offer is one of the few instruments a business of seven people can design in a week and use for a year.

QUÉBEC & CANADA NOTE How many small businesses there are, and how long they last

The federal small business statistics give the size of the field this book is written for. A small business has 1 to 99 paid employees and an SME 1 to 499; there were 1,079,188 small businesses in December 2024, 98.2% of employer businesses, employing 5.8 million people, 46.6% of the private-sector labor force. Survival was 68.0% at five years, 48.2% at ten years and 24.2% at twenty-one years, and services-producing firms survived at 67.5%, 47.2% and 23.0% over the same horizons. Firms that started with 20 to 99 employees survived five years at 75.0%, which is a reminder that the size a business begins at shapes what it can carry. Small and medium-sized businesses produced 47.8% of private-sector GDP on average in 2018–2022, and small businesses alone 34.2% (Innovation, Science and Economic Development Canada, 2026a). The same department publishes a free financial performance tool, covering more than 1,000 industries with more than 30 benchmarks, with which a business can compare its own revenue and expense figures with those of its industry before it decides what its offer can afford (Innovation, Science and Economic Development Canada, 2026b). The figures are read here for one purpose: to place the offer in a market where most businesses are small, where most do not reach their tenth year, and where the difference between a designed offer and an accumulated one is available to any of them.

WHAT CHANGED From goods to services to experiences

The 1998 argument that an experience is a fourth kind of economic offering, after commodities, goods and services, changed what a business could sell (Pine & Gilmore, 1998). The birthday cake traced through the economy makes the shift concrete: the same cake is a commodity, a good, a service or an experience depending on what the customer is buying, and the price rises at each step. Two decades later, the observation that services are about 80% of United States GDP, and that innovation in services has been less disciplined than innovation in manufacturing, describes where most of the economy now sits (Bitner et al., 2008). Read together, the two statements set the landscape this book works in. Most of what a small business sells is a service, most of it is designed less carefully than a manufactured product, and the customer's memory of the encounter is part of what was bought. The offer is the place where a business decides what mixture of the four it is selling, and the decision is available to a business of any size.

1.5 What Is at Stake for a Small Business

An offer decides three things at once, and a small business feels all three in the same month. The first is who can say yes. An offer is built for a customer, and the customer it is built for is the one who can accept it without asking for it to be changed. Nordec's contract was designed for a building of 4,500 square metres, and when it was put in front of a 2,500-square-metre building the owner said the contract was more than the building needed. That is not a sales failure; it is an offer that was never designed for the buyer in the room, and the two proposals Nordec lost to the trades billed by the call were both lost that way.

The second thing an offer decides is what the business must deliver. Every commitment in the promise layer becomes an obligation with a cost, and every element in the supplements layer becomes work that somebody has to do. Nordec's ninety-minute response commitment is delivered by a dispatcher, four technicians and two vehicles, and it is honored on 96% of urgent contract calls at an average response of 64.4 minutes. The building health record is kept for every contract building. The capital planning note has been delivered 24 times. None of those is free, and all of them were given away before anyone costed them.

The third thing an offer decides is what the business earns per hour of the work it sells. The cost to serve is set by what the offer contains, and the price is set separately, which means the two can drift apart for a long time before anyone notices. Nordec's preventive visit for a building in a territory costs $245.60 a month: 3.5 hours of preventive work plus 0.4 hour of travel at a loaded technician cost of $54 an hour, plus $35 of consumables. The urgent-call element costs $37.13: 5.5 calls a building a year at 1.5 hours each, at the same rate, divided by twelve. Those two elements were the whole offer in 2022, and they cost $282.73 a month against a price of $691, a margin of 59.08%.

An offer that nobody designed shows itself in symptoms a business can recognize without any measurement at all. The first symptom is discounting everywhere: the price is negotiated on every sale, because the offer does not give the buyer a reason to accept it as written. The second is options nobody chooses: a tier or an option that exists on the page and is never selected, which usually means it was designed for the business's convenience rather than the customer's need. The third is requests with nothing to buy: customers ask for something the business already does informally, and there is no line to sell it on. Book E1 (The Sales Process: From Prospect to Signed Contract) treats the sales side of the same problem, and the two books meet at the point where a salesperson has to improvise because the offer does not cover what the buyer asked for.

The quiet way an offer degrades is accretion, and it is quiet because every step of it is reasonable. A customer asks whether the business could send a short report each month, and the answer is yes. A technician starts leaving a sheet behind, and the customers like it. A renewal conversation produces a note about the equipment that will need replacing, and the owner finds it useful. Each addition is small, each is given free, and each is delivered by somebody who was already being paid. Together they change the cost of the offer without changing its price, and the business discovers the change in a margin rather than in a decision.

WATCH OUT Adding without subtracting

Each addition to an offer feels small and free, and that feeling is what makes accretion dangerous. A quarter-hour a month of record-keeping, twenty minutes of a technician and fifteen minutes of an administrator for a report, three hours a year for a planning note — none of those is a decision anyone would hesitate over, and none of them appears on a statement. Together they are a different offer. At Nordec, four additions cost $58.50 a month per building, which is $32,292 a year across 46 buildings, while the price rose by $21 a month, which brings $11,592. The gap is $20,700 a year, and it was never priced because it was never decided. The remedy is not to refuse additions; it is to make each one a decision with a date and a number, and to ask the two questions together: what does this add to the customer's offer, and what does it add to the cost of serving them?

DO THE MATH Six elements, one price

Illustrative figures; the case is fictional. Step 1 — The four elements added after the founding cost $13.50 for the building health record's upkeep, $27.00 for the one-page monthly report, $13.50 for the capital planning note at renewal and $4.50 for the mid-contract note, which is $58.50 a month per building. Step 2 — Across 46 buildings and twelve months, that is 58.50 × 12 × 46 = $32,292 a year of added cost. Step 3 — The price rose from $691 to $712, which is $21 a month, or 21 × 12 × 46 = $11,592 a year of added revenue. Step 4 — The difference, $32,292 − $11,592 = $20,700 a year, is cost the offer carries and no line prices. Step 5 — The margin of a territory building fell from 59.08% at $691 with a cost of $282.73 to 52.08% at $712 with a cost of $341.23, which is a loss of seven points of margin on every building in the two territories.

WHO DOES THIS The owner owns the offer until someone else is named

In a small business, the offer belongs to the owner by default, and the default is the problem. The owner decides what is included, what is promised and what is given away, usually in a conversation with a customer, and those decisions accumulate without a document. In Nordec's case, Sophie Nadeau founded the business in 2022 after fourteen years as a building systems technician, and she is the person who set the contract's contents, its exclusions and its response commitment. What she did not have, until the first annual refresh, was a written inventory of what the offer contained and what each element cost. Naming the offer as somebody's responsibility is the first act of design. In a business of seven people the owner can hold it, provided the inventory exists and the reviews happen; in a larger business the work passes to a manager, and the owner's role becomes the decision rather than the maintenance. Chapter 10 treats expertise and roles, and Chapter 9 treats the cadence that keeps the inventory true.

1.6 The Vocabulary of Offer Design

The rest of this book uses a small vocabulary, and the words are worth fixing now because several of them are used loosely in the trade. The table below gives fifteen terms, the meaning each one carries in this book, and the element of Nordec's offer that illustrates it. Read it as a glossary rather than as a chapter: the terms are used from Chapter 2 onward, and a reader who returns to this table will find that each one has acquired a chapter behind it.

Table 1.2 — The vocabulary of offer design

The vocabulary is not decoration. Each term names a decision that a business either makes deliberately or makes by accident. A tier is a decision about who can buy what. A fence is a decision about who cannot. A guarantee is a decision about what happens when the business fails. A take rate is the measurement that tells the business whether the first three decisions were right. A business that has no words for these things still makes all of them, and it makes them in conversations that nobody records.

TOOLS OF THE TRADE What an offer inventory needs

The inventory described in section 1.3 is carried by four ordinary things, and none of them is a specialist product. The first is a spreadsheet with one line per element, with columns for the element, the date it was added, the person who delivers it, the time it takes, the cost per customer per month and whether it is priced. The second is the contracts and proposals of the last two years, which are the record of what the business actually promised, as distinct from what it believes it promises. The third is the invoices, which show what was billed and, more usefully, what was delivered and never billed. The fourth is the time records of the people who deliver the offer, which are the only source for the cost of an element that nobody has ever costed. No brand name belongs on this list, and no price; the requirement is the function, and a business that has the four things has everything an offer inventory needs.

1.7 The Rules That Apply

An offer is a promise, and a promise made to the public is regulated. The rules that follow are the ones this book refers to, and they are stated here in outline because Chapters 5 and 6 return to them in detail. This book explains these laws; it does not give legal advice, and a business that is designing a guarantee, a warranty or an advertising claim should take advice from a lawyer on its own facts.

In Canada, the Competition Act applies in every province and to sales to businesses as well as to consumers. Section 52 makes it a criminal offense to knowingly or recklessly make a representation to the public that is false or misleading in a material respect, by any means, and the Crown does not have to prove that anyone was deceived; advertising a price that leaves out a mandatory charge other than one imposed by law is deemed misleading. Section 74.01(1) treats the same conduct as civil reviewable conduct, and adds two provisions that bear directly on offer design: a representation about the performance, efficacy or length of life of a product must be based on an adequate and proper test, with the burden of proving the test resting on the person who makes the claim, and a representation in the form of a warranty or guarantee must not be materially misleading and must have a reasonable prospect of being carried out.

For offers made to consumers in Québec, the Consumer Protection Act adds its own requirements. Goods must be fit for the purposes for which goods of that kind are ordinarily used, and they must be durable in normal use for a reasonable length of time given their price, the terms of the contract and the conditions of use. Goods and services must conform to the statements and advertising of the merchant or manufacturer, which bind them, and a warranty mentioned in a statement or an advertisement binds even if it is not written into the contract. The Act also governs contracts of additional warranty, which is what the trade calls an extended warranty, and, from October 5, 2026, new goods of listed kinds sold or leased to consumers carry a warranty of good working order whose duration is set by regulation, covering parts and labor, with the duration stated prominently near the advertised price.

Nordec's contracts are between two businesses, and in Québec those are governed by the Civil Code of Québec rather than by consumer protection law. Article 2100 provides that a contractor or provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice and, where applicable, must ensure that the work or service conforms to the contract; when bound to an obligation of result, the provider may be relieved of liability only by proving superior force. That article is the reason a promise written into an offer is not a marketing statement: it is the standard against which the delivery will be measured. A reader who sells to consumers rather than to businesses must know that a different regime applies, and Book H4 (Consumer Protection, Advertising Law, Competition and Language Requirements) treats it in full.

In the United States, the Magnuson-Moss Warranty—Federal Trade Commission Improvement Act governs written warranties on consumer products. It defines a consumer product as tangible personal property normally used for personal, family or household purposes, and a written warranty as a written affirmation or promise, made in connection with the sale, about the nature of the material or workmanship or that the product is defect-free or will meet a specified level of performance over a specified period. It sets the minimum standards of a full warranty — remedy within a reasonable time and without charge, no limit on the duration of implied warranties, an exclusion of consequential damages only if conspicuous, and a refund or replacement after a reasonable number of failed attempts to repair — and a warranty that does not meet them must be called limited.

The following table gathers the rules that frame an offer in this book, with what each one requires, whom it protects and the chapter where it is treated. Read it as a boundary rather than as a procedure: the rules do not tell a business what to promise, and they do tell it what a promise costs once it is made.

Table 1.3 — The rules that frame an offer in this book

THE RULE What an offer promises must be true and deliverable

Under the Competition Act, s. 52, it is a criminal offense to knowingly or recklessly make a representation to the public that is false or misleading in a material respect, by any means, and the Crown does not have to prove that anyone was deceived. Under s. 74.01(1)(a), the same representation is civil reviewable conduct, and under s. 74.01(1)(c), a representation in the form of a warranty or guarantee must not be materially misleading and must have a reasonable prospect of being carried out. The Act applies in every province and to sales to businesses as well as to consumers. The rule for an offer designer is therefore short: a promise written into an offer must be true, and it must be something the business can actually deliver. This book explains the law and does not replace the advice of a lawyer.

1.8 The Case and the Plan of This Book

The business followed through this book is Nordec Entretien inc., a fictional company created for this encyclopedia. It provides preventive maintenance and on-call technical service to commercial buildings on Montréal's north shore, in light HVAC, minor plumbing, service electrical and loading doors. Sophie Nadeau founded it in 2022 after fourteen years as a building systems technician. It has seven people — Sophie, four technicians, one of them the lead technician, a dispatcher and an administrator — and two vehicles. It sells a monthly preventive contract with a ninety-minute response commitment on urgent contract calls, day and night, and bills work beyond the scope at $96 an hour plus parts. Its market is buildings of 2,000 to 9,000 square metres, of which about 1,400 lie within forty kilometres of its depot, in two territories.

Book A2 (The Business Plan) and Book A3 (Strategy and Competitive Positioning) established the business and its strategy, and this book begins one year after that strategy was signed, at the first annual refresh. Nordec holds 46 buildings under contract, 38 of them in the two territories and 8 scattered, at an average contract of $712 a month and a contract run-rate of $393,024 a year. Its renewal rate is 95.5%, with 22 renewals and one building lost. It answers 96% of urgent contract calls on site within ninety minutes, at an average response of 64.4 minutes. Every contract building has a building health record, and 24 capital planning notes have been delivered. Three property management firms hold 14 of the buildings, or 30.4% of the total. At the refresh, the team decided to add a second capital planning note, mid-contract, for the buildings whose equipment the first note had flagged. In the strategy year, 19 proposals were decided: 11 were won, 4 were lost on price, 2 were lost on after-hours coverage and 2 were left without a decision.

The offer inventory at the refresh found six elements and three signals. The six elements are set out in the table below, with the date each was added, its cost per building per month and what the customer sees. Read the table from top to bottom: the first two rows are the offer as it was designed in 2022, the next three were added during the strategy year, and the last was decided at the refresh and has not yet been announced. The total column is the cost of serving one building in a territory, and the difference between the total and the price is the margin.

Table 1.4 — Nordec's offer, element by element

The three signals are the reason the book exists. The first is who says no: two of the four proposals lost on price went to the trades billed by the call, and both were small buildings under 3,500 square metres whose owners said the contract was more than the building needed. The second is what customers ask for that nothing in the offer contains: 5 customers asked for a quarterly review meeting, 4 asked for help with replacement projects, and 3 property managers asked for a single report across their buildings, which is 12 requests with nothing to buy. The third is what the business delivers without pricing it: the $20,700 a year of added cost described in section 1.5. An offer that produces all three signals at once is an offer that has outgrown its design.

CASE IN POINT Nordec Entretien inc.: the contract that grew by itself

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The contract Sophie Nadeau wrote in 2022 had two elements: a monthly preventive visit and urgent calls answered within ninety minutes with up to two hours of work included. Together they cost $282.73 a month to deliver and sold for $691, a margin of 59.08%. Over the strategy year, four elements were added, each in answer to a customer or a good idea: the upkeep of the building health record, a one-page monthly report, a capital planning note at renewal, and a mid-contract note for the buildings the first note had flagged. Each was given free. Together they cost $58.50 a month per building, or $32,292 a year across 46 buildings, while the price rose from $691 to $712, which brings $11,592. The margin of a territory building is now 52.08%. Nothing went wrong in any of those decisions, and that is the point: the offer grew by itself, and the only record of it is a margin that fell seven points.

The plan of the book follows the case. Chapter 2 treats the understanding of value, through the twenty value interviews that asked customers and prospects what they actually buy. Chapter 3 treats service design and the customer journey, through the blueprint of an urgent call. Chapter 4 treats packaging, through the three tiers and two add-ons. Chapter 5 treats guarantees, service levels and risk reversal, through the ninety-minute guarantee and the first-ninety-days guarantee. Chapter 6 treats value-added services and complementary offers, through the service menu of eight candidates and the decisions taken on each. Chapter 7 treats productizing services and servitizing products, through the Building Check. Chapter 8 treats testing and iterating the offer, through the sixty-day pilot and its decision rules. Chapter 9 treats the typical tasks, cadence and workflows that keep an offer true. Chapter 10 treats expertise, roles and when to get help. Chapter 11 treats the tools, by function rather than by brand. Chapter 12 gathers the practical advice, the pitfalls and the action plan, and reports where the case stands six months after the refresh.

IN LA MACHINE Where an offer can be recorded

Two pages of the application copy read for this book hold an offer in the form of a package (La Machine, 2026). Business → Session Packages, at productivity/packages and also reached at business/packages, defines packages of sessions or minutes: each has a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost — the page shows the profit and the margin per sale as the package is edited —, a description and a promotional image, and a package can be added, edited, deactivated or deleted. Business → Subscription Packages, at business/subscription-packages, defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown; a package that is disabled lets existing subscribers keep their access until the end of their current billing period. What the platform does not have belongs in the same breath: there are no tiers, bundles or add-ons for the user's own packages, since each package is a single, flat item, and there is no guarantee or warranty register. The package pages record what an offer contains and what it costs; the design of the offer is done outside them.

Costly Mistakes

• Describing a capability instead of an offer. A sentence such as "we do commercial maintenance" cannot be accepted, refused, compared or priced, and a business that sells from one will find that every sale is a negotiation.
• Leaving the boundary of the offer unwritten. An offer that lists only what is included leaves the customer to guess at what is excluded, and the guess is always more generous than the contract.
• Treating the offer as the price. The contents of an offer and the amount charged for it are two decisions, and changing one without revisiting the other is how a margin falls without anyone deciding that it should.
• Adding elements without subtracting or pricing them. Each addition feels small and free, and together they change the cost of the offer while the price stays where it was.
• Designing one offer for every customer. A contract built for a 4,500-square-metre building is more than a 2,500-square-metre building needs, and the owner of the smaller building will say so and buy from somebody else.
• Making a promise the business cannot deliver. A response time, a guarantee or a performance claim written into an offer is an obligation, and under the Competition Act a warranty or guarantee must not be materially misleading and must have a reasonable prospect of being carried out.
• Leaving the offer to the owner's memory. An offer that exists only in conversations has no inventory, no cost and no review, and it will be discovered in a margin rather than in a decision.

Key Takeaways

• An offer is the specific, purchasable combination a business puts in front of a customer: what is done, what the customer receives and does not receive, the commitments the business makes, and the terms on which the whole thing is sold.
• The offer is not the value proposition of Book A1, the positioning of Book A3, the price of Book A7 or the message of Book B6; each of those answers a different question, and the offer is the one a customer can accept.
• Every offer mixes goods and services, and there is no such thing as a commodity: all goods and services can be differentiated, and industrial products and services offer more room for differentiation than consumer goods (Levitt, 1980).
• The five layers of an offer are the core, the promise, the supplements, the evidence and the terms, and the outer three are where offers grow without anyone deciding.
• An undesigned offer shows itself in three signals: who says no and why, what customers ask for that nothing in the offer contains, and what the business delivers without pricing it — at Nordec, $20,700 a year of added cost that no line prices.
• What an offer promises must be true and deliverable: the Competition Act applies in every province and to sales to businesses as well as consumers, and a warranty or guarantee must not be materially misleading and must have a reasonable prospect of being carried out.

Your Action Plan

1. Write down what a customer actually buys. Take the last three contracts or proposals the business signed and list every element the customer receives, including the ones given without being written into the document. Deliverable: one list of elements, in the customer's language rather than the business's. Verification: a colleague who does not do the work can read the list and say what the customer gets.
2. Place each element in one of the five layers. Sort the list into the core, the promise, the supplements, the evidence and the terms, and mark any element that does not fit, because an element that fits nowhere is usually a habit rather than an offer. Deliverable: a five-layer inventory of the offer. Verification: every element has a layer, and every layer has at least one element.
3. Date and cost every element. Add the date each element was introduced and its cost per customer per month, using the loaded cost of the people who deliver it and the consumables it consumes. Deliverable: a dated, costed inventory. Verification: the total cost column and the price column are both filled in, and the difference between them is stated as a margin.
4. Write the three signals. Record who said no in the last year and for what reason, what customers asked for that the offer does not contain, and what the business delivers that no line prices. Deliverable: three short lists with dates and names attached. Verification: each signal has at least one entry, or the business can state that it has looked and found none.
5. Name the owner of the offer and set the first review. Decide who holds the inventory, and place a date within the next quarter to read it against the price and the cost to serve. Deliverable: a named owner and a dated review. Verification: the owner can say what the offer contains without opening the contract, and the review date is on the calendar.

In La Machine

Two pages of the application copy read for this book record an offer in the form of a package (La Machine, 2026). Business → Session Packages, at productivity/packages and also reached at business/packages, defines packages of sessions or minutes, each with a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image; the page shows the profit and the margin per sale as the package is edited, and a package can be added, edited, deactivated or deleted. The same page lists each client's credit balance and its history, and credits added by hand are recorded as free gifts or adjustments, with no payment. Business → Subscription Packages, at business/subscription-packages, defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown; a package that is disabled lets existing subscribers keep their access until the end of their current billing period. Read together, the two pages hold the contents and the internal cost of an offer, which is the pair this chapter asks a business to write down.

What the platform does not do belongs in the same breath. There are no tiers, bundles or add-ons for the user's own packages, because each package is a single, flat item; there is no guarantee or warranty register, no service-level tracking and no service credits, no refund workflow and no free trial, and no service blueprint, customer journey map or survey tool beyond a general form builder. The design of an offer, in the sense of this chapter, is therefore done outside the platform and recorded inside it. A business that wants to sell three tiers records three packages, and the fence that keeps the lowest tier from being bought by the wrong customer is a rule the business applies rather than a setting the platform enforces. Book N1 (Getting Started with La Machine) maps the modules the platform does cover, and Book A7 (Pricing Strategy and Revenue Models) treats the setting of the prices those packages carry.

Self-Check

1. What is an offer, and how does it differ from the value proposition, the positioning, the price and the message?
2. What are the five layers of an offer, and which of them is where an offer most often grows without anyone deciding?
3. Why does the argument that there is no such thing as a commodity matter to a small service business?
4. What are the three signals of an undesigned offer, and what did each of them show at Nordec?
5. What does the Competition Act require of a warranty or a guarantee written into an offer, and to whom does the Act apply?

Answers. 1. An offer is the specific, purchasable combination a business puts in front of a customer — what is done, what the customer receives and does not receive, the commitments the business makes and the terms on which it is sold — while the value proposition of Book A1 answers what the business sells and to whom and why it is better, the positioning of Book A3 answers where the business plays and how it wins, the price of Book A7 answers what the customer pays and how that amount is set, and the message of Book B6 answers how the offer is put into words for the market. 2. The five layers are the core, which is what is done; the promise, which is what the business commits to about the core; the supplements, which are the services delivered around it; the evidence, which is what the customer sees and keeps; and the terms, which are the price, duration, exclusions, renewal and way out. The supplements and the evidence are where an offer most often grows without anyone deciding, because each addition is small, useful and given free. 3. Because a business that believes it sells a commodity has only price left to compete on, while the same work examined for the elements a customer would pay for becomes an offer that can be designed; Levitt (1980) argued that all goods and services can be differentiated and that industrial products and services offer more room for differentiation than consumer goods. 4. The three signals are who says no and why, what customers ask for that nothing in the offer contains, and what the business delivers without pricing it; at Nordec, two of the four proposals lost on price went to the trades billed by the call and both were small buildings under 3,500 square metres, 12 requests arrived with nothing to buy, and $20,700 a year of added cost was never priced. 5. Under s. 74.01(1)(c) of the Competition Act, a representation in the form of a warranty or guarantee must not be materially misleading and must have a reasonable prospect of being carried out, and under s. 52 a representation to the public that is false or misleading in a material respect is a criminal offense whether or not anyone was deceived; the Act applies in every province and to sales to businesses as well as to consumers.

Chapter References

Bitner, M. J., Ostrom, A. L., & Morgan, F. N. (2008). Service blueprinting: A practical technique for service innovation. California Management Review, 50(3), 66–94. https://doi.org/10.2307/41166446

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

Hart, C. W. L. (1988, July). The power of unconditional service guarantees. Harvard Business Review. https://hbr.org/1988/07/the-power-of-unconditional-service-guarantees

Innovation, Science and Economic Development Canada. (2026a). Key small business statistics 2025 (page modified April 23, 2026). Retrieved September 22, 2026, from https://ised-isde.canada.ca/site/sme-research-statistics/en/key-small-business-statistics/key-small-business-statistics-2025

Innovation, Science and Economic Development Canada. (2026b). Financial performance data (page modified July 2, 2026). Retrieved September 22, 2026, from https://ised-isde.canada.ca/site/financial-performance-data/en

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages and business/subscription-packages.

Levitt, T. (1980, January). Marketing success through differentiation—of anything. Harvard Business Review. https://hbr.org/1980/01/marketing-success-through-differentiation-of-anything

Magnuson-Moss Warranty—Federal Trade Commission Improvement Act, 15 U.S.C. §§ 2301 and 2304 (United States). Legal Information Institute, Cornell Law School — read September 22, 2026.

Osterwalder, A., Pigneur, Y., Bernarda, G., & Smith, A. (2014). Value proposition design: How to create products and services customers want. Wiley.

Pine, B. J., II, & Gilmore, J. H. (1998, July–August). Welcome to the experience economy. Harvard Business Review. https://hbr.org/1998/07/welcome-to-the-experience-economy

Chapter 2 — Understanding Value

Value is what customers believe they receive for what they give up, judged against their next best alternative, and a business can design an offer well only once it knows, in the customers' own words, which parts of what it does create that value and for whom.

What You'll Be Able to Do

• Explain why value is a judgment the customer makes rather than a quantity the seller delivers, and why that judgment is always made against an alternative.
• Separate the families of value — functional, emotional and economic — and say which of them a given offer actually carries for a given segment.
• Describe a customer's job to be done, with its functional, social and emotional dimensions, and connect it to the pains and gains a value map answers.
• Build a value equation that puts the customer's alternative and the offer side by side in dollars, and state plainly what the equation leaves out.
• Run a value interview that asks about the last time the problem happened rather than about the product the customer imagines, and count the answers across groups.
• Order the elements of an offer from the must-be through the performance elements to the attractive ones, and recognize which of them are painkillers and which are vitamins.

Why It Matters

Every offer is a claim about value, and most offers fail not because the claim is dishonest but because it was never checked. A business knows what it does: the visits it makes, the hours it keeps, the parts it stocks, the reports it writes. The customer knows something different: what the problem cost them the last time it happened, what they did about it, and what they would have done if the business did not exist. Those two knowledges are not the same, and the distance between them is where offers are designed badly. A seller who has never asked the customer what the alternative costs is guessing at the value of their own work, and a guess about value becomes a guess about price, which becomes a negotiation the seller cannot win.

The stakes are concrete for a small business, because the alternative is usually not nothing. It is a patchwork of trades, a competitor, an internal arrangement, or simply doing without and absorbing the consequences. In the case followed through this book, a building of about 4,500 square metres arriving from a patchwork of trades would spend $3,900 a year on preventive work and pay for eleven unplanned interventions at an average of $560, or $10,060 a year, against the contract's $8,544. The contract is cheaper, but nobody had ever written that down, and the business had been selling the contract for four years on the strength of what it does rather than on the strength of what the customer avoids. The same arithmetic, run on a smaller building, produces the opposite answer: $6,620 for the alternative against $7,080 for the contract, a loss of $460 for the customer, which is exactly what two lost owners had said in fewer words when they described the contract as more than the building needed.

Understanding value also protects a business from a quieter failure, which is adding things nobody asked for and calling them value. The contract in this case had grown by accretion: a monthly visit, urgent calls, a building health record, a one-page monthly report, a capital planning note at renewal, and a mid-contract note for the buildings the first note had flagged. Each addition was reasonable on the day it was made and each was given away free. Together they cost $58.50 a month per building, or $32,292 a year across forty-six buildings, while the price rose from $691 to $712 and brought $11,592. The margin of a territory building fell from 59.08% to 52.08%. None of that is a pricing problem in the first instance; it is a value problem, because the business had never established which of the six elements its customers would actually miss. This chapter is the work that comes before the packaging, the guarantees and the tiers of the chapters that follow.

In This Chapter

Section 2.1 establishes that value is perceived rather than delivered, and that it is judged against the customer's next best alternative. Section 2.2 sets out the families of value in consumer and business markets and separates the economic part a buyer can put in dollars. Section 2.3 describes the job the customer is trying to get done, with its functional, social and emotional dimensions, and the pains and gains that surround it. Section 2.4 shows how to use the elements of value without turning them into a checklist. Section 2.5 builds the value equation this book uses and applies it to two buildings. Section 2.6 covers how to measure what customers value, from the interview to the theory of attractive quality. Section 2.7 orders the elements of an offer and distinguishes the painkiller from the vitamins. Section 2.8 tells the story of the twenty interviews the case ran and the conclusion they produced.

2.1 Value Is Perceived, Not Delivered

Value is what a customer believes they receive for what they give up. The sentence is short, and it moves the whole subject away from the seller. A business can deliver a service exactly as designed, on time, at the agreed standard, with every promise kept, and the customer can still judge the offer to be poor value. Nothing has gone wrong in the delivery. What has happened is that the customer has weighed what they received against what they gave up, and the weighing did not come out in the seller's favor. This is why value cannot be installed, shipped or guaranteed in the way a part can. It is formed in the customer's mind, from their circumstances, their history and their alternatives, and the seller's only access to it is through asking.

What the customer gives up is broader than the price. It includes the money, certainly, and the money is the part that appears on an invoice. It also includes the time spent arranging the work, the attention spent supervising it, the risk that the work will not be done, the effort of managing a supplier, the cost of switching away from whatever they use today, and the discomfort of a decision they may regret. A maintenance contract asks an owner to give up a monthly amount, a degree of control over who enters the building, and the freedom to call whichever trade they like on any given morning. Those are real costs, and a seller who counts only the invoice will overstate the value of the offer by exactly the amount of the rest.

What the customer receives is equally broad. It includes the outcome, which is a building that works. It includes the reliability of the outcome, which is the difference between a promise and a habit. It includes the feeling that comes with not having to think about the equipment, and the proof that can be shown to somebody else. In the case followed through this book, the two jobs the customers described in their own words carry all of these at once. The owner-operator wants to keep the building running and the tenants quiet without becoming a maintenance expert, and to know a year ahead what will have to be replaced. The property manager wants to show the owners they report to that their buildings are under control, and never to be the last to know about a failure. Neither of those sentences is about a monthly visit, and both of them are about value.

The research tradition behind this way of thinking treats value as the end of a chain rather than a single quantity. The means-end model relates price, perceived quality and perceived value, and draws implications for managers who decide price, quality and value (Zeithaml, 1988). The model is useful here for one reason and it should be used for only that reason: it places price and quality on the same path toward a judgment the customer makes, rather than treating them as two separate arguments. A seller who lowers a price without changing what the customer perceives has moved one number and left the judgment where it was. A seller who raises perceived quality without touching the price has moved the other. The judgment at the end of the path is what the customer acts on.

The most important consequence of perceived value is that it is always comparative. A customer does not evaluate an offer in isolation; they evaluate it against what they would otherwise do. That alternative may be another supplier, a bundle sold by a larger firm, a set of trades billed by the call, an internal person, or doing nothing at all and accepting the consequences. The alternative is the benchmark, and it is the benchmark that makes value measurable. A contract that costs $8,544 a year is expensive or cheap depending entirely on what the customer would spend without it, and there is no way to answer the question from the seller's side of the table.

This is where the seller's cost becomes a distraction. The cost of serving a building in a territory is $245.60 a month for the preventive visit alone, and the full cost of the current contract is $336.73 a month. Those figures matter enormously to the business, because they decide whether the work is worth doing. They matter not at all to the customer, who has never seen them and has no reason to care. A seller who argues from cost is asking the customer to accept the seller's problem as the basis for the price, and customers do not do that. They compare the offer with their alternative, and the comparison is made in their own numbers.

The gap between what a business thinks it sells and what a customer thinks they buy is therefore not a failure of communication. It is a structural feature of every transaction, and it widens with the complexity of the offer. Sophie Nadeau founded Nordec Entretien in 2022 and had worked fourteen years as a building systems technician, which means she knows precisely what a preventive visit contains: the inspection of mechanical, plumbing and door systems, the checklist, the parts, the judgment about what is about to fail. Her customers know none of that. They know that the elevator room flooded once, that a tenant complained for three weeks, and that somebody now comes every month. The seller's knowledge is deep and narrow; the buyer's knowledge is shallow and wide, and the offer has to be designed for the buyer's version.

The practical consequence for this book is a rule about sequence. A business cannot package, tier, bundle or guarantee an offer until it knows which parts of what it does create value and for whom. Chapter 4 designs three tiers for Nordec, and it can do so only because this chapter establishes that two kinds of customer want different things from the same contract. Chapter 5 writes guarantees, and it can do so only because this chapter establishes which promises the customers would actually miss. Chapter 6 decides which services to add and which to decline, and it can do so only because this chapter counts who asked for what. The order is not a preference; it is the only order in which the later decisions have anything to stand on.

DEFINITION Perceived value, economic value and the next best alternative

Perceived value is the customer's own judgment of what an offer gives them, weighed against everything the offer asks them to give up: the money, the time, the attention, the risk and the effort of managing it. It is formed in the customer's mind and cannot be delivered by the seller. Economic value is the part of that judgment a buyer can express in dollars, usually by comparing what the offer costs with what the alternative costs over the same period. It is the part that can be written into a value equation and shown to a buyer, and it is never the whole of perceived value. The next best alternative is what the customer would actually do if the offer did not exist: another supplier, a bundle, a set of trades billed by the call, an internal arrangement, or nothing at all. It is the benchmark against which value is judged, and it is the reason a seller's own cost has no place in the customer's calculation.

KEY CONCEPT Customers compare offers with their alternative, not with the seller's cost

The single idea that reorganizes how a business thinks about its offer is this: the customer is not comparing the price with what the work costs the seller. The customer is comparing the whole offer with the next best alternative, and the alternative is usually a mixture of money, inconvenience and risk. A building arriving from a patchwork of trades spends $3,900 a year on preventive work and pays for eleven unplanned interventions at an average of $560, which is $10,060 a year, and it also absorbs the disruption of eleven failures, the phone calls, the waiting and the arguments about who should have caught the problem. A contract at $8,544 a year is cheaper in money and far cheaper in disruption, and the seller who can say so has moved the conversation from a price to a comparison. The seller who cannot say so is left arguing about a monthly amount against a customer who has never added up the alternative.

2.2 Functional, Emotional and Economic Value

Value arrives in families, and a seller who can name the family they are selling in is a seller who can design for it. The most useful map of those families comes from a body of work on the elements of value, built from consumer research and arranged as a pyramid of four levels modeled on Maslow's hierarchy of needs. The thirty elements run from functional at the base, through emotional, to life-changing and social impact at the top, and combinations of elements are associated with stronger loyalty, a greater willingness to try a brand and sustained revenue growth (Almquist et al., 2016). The same research program was later extended to business markets, where forty elements are grouped in five categories: table stakes, functional, ease of doing business, individual and inspirational. Those run from objective factors such as price and specifications to subjective ones such as reducing the buyer's anxiety or enhancing the buyer's reputation, and the authors observe that as offers commoditize, the personal considerations of business customers matter more (Almquist et al., 2018).

The consumer pyramid and the business categories are not two separate subjects. They are the same idea applied to two kinds of buyer, and the difference between them is instructive for anyone selling to a business. A consumer buying a maintenance service for their own home is buying a functional outcome and an emotional relief; a property manager buying the same service for a building they do not own is buying a functional outcome, an ease of doing business, and something about their own standing with the owners they report to. The work is similar. The value is not, and the offer that serves one will not serve the other without being redesigned.

Functional value is the easiest to describe and the easiest to underestimate. It is the equipment that keeps working, the part that is replaced before it fails, the response that arrives within the promised time. A seller tends to assume that functional value is the whole of the offer, because it is the part they can see themselves producing. It is the base of the pyramid and it is necessary, but it is rarely sufficient, and it is rarely the reason a customer chooses one supplier over another when several can deliver it.

Emotional value is the part that shows up in the customer's sentences rather than in the specification. It is the relief of not thinking about the boiler, the confidence of knowing who to call, the absence of the Sunday evening worry that the roof will leak before Monday. In the interviews described later in this chapter, the property manager's job was expressed as never being the last to know about a failure, which is an emotional statement about professional standing, not a functional statement about response times. The seller who hears only the functional half of that sentence will design a faster response and miss the point.

Life-changing and social-impact value sit at the top of the consumer pyramid, and their business equivalents are the individual and inspirational categories. They are the elements that connect the purchase to something the buyer cares about beyond the transaction: their reputation with the people they answer to, their sense of doing the job properly, their contribution to a larger purpose. A capital planning note that lets an owner budget for a replacement is functional in form and individual in effect, because it changes the owner's standing in a conversation with their own board or their own bank. The same document can be sold as a piece of paperwork or as a piece of professional standing, and the second version is worth more.

Economic value is the family that deserves its own paragraph, because it is the one a seller can put in dollars and the one a buyer will test. Economic value is what the offer saves or earns the customer, expressed in the customer's own units, measured against the alternative. It is not the same as price, and it is not the same as the seller's margin. It is the difference the offer makes to the customer's costs or revenues, and it is the only family that can be written into an equation and defended in a meeting. Section 2.5 builds that equation in full, and it is worth saying here that economic value is a component of perceived value rather than a substitute for it. A customer can see a clear saving and still refuse the offer, because the emotional or social elements are missing, or because the effort of switching is larger than the saving.

The following table sets out the families of value side by side as they appear in consumer and business markets, with one example drawn from a service business of the kind this book follows. Read it as a translation table rather than a hierarchy: the same offer can carry elements from several rows at once, and the question a seller has to answer is which rows the segment they are selling to actually cares about.

Table 2.1 — The families of value in consumer and business markets

The table also shows why a seller should resist the temptation to claim every family at once. An offer that promises functional reliability, emotional relief, professional standing and a contribution to the environment is an offer that has not chosen, and a customer who cannot tell which of those is being sold will compare on the only dimension that is easy to compare, which is price. The discipline of choosing the few families that matter to a segment is the subject of section 2.4, and it begins with knowing what job the customer is trying to get done.

2.3 Jobs, Pains and Gains

A customer does not buy a product or a service. A customer hires it to make progress in a circumstance, and the progress is what the seller has to understand. The argument against relying on correlations and customer profiles is that they describe who the customer is rather than what the customer is trying to do, and the question that produces useful answers is which job the customer hires the offer to do. Jobs are rarely only functional; they carry social and emotional dimensions, and a seller who ignores those dimensions is designing for a fraction of the purchase (Christensen et al., 2016). The distinction matters in a small business because it changes what the seller listens for. A profile tells a seller that the customer is an owner-operator of a mid-size commercial building. A job tells the seller what that person is trying to accomplish on a Tuesday morning when the loading door will not close.

The job in the case followed through this book was expressed by the customers themselves, in two sentences that are worth reading slowly. The owner-operator wants to keep the building running and the tenants quiet without becoming a maintenance expert, and to know a year ahead what will have to be replaced. The property manager wants to show the owners they report to that their buildings are under control, and never to be the last to know about a failure. The first sentence contains a functional job, which is keeping the building running; a social job, which is keeping the tenants quiet; and an emotional job, which is not having to become an expert in something that is not their trade. The second contains a functional job, which is control; a social job, which is reporting to owners; and an emotional job, which is never being surprised.

The vocabulary that turns those sentences into a design is the customer profile and the value map. The customer profile describes the jobs customers try to get done, their pains and their gains. The value map describes the pain relievers and gain creators an offer provides, and the fit between the two is what a value proposition is. The method insists on testing the concept before an expensive launch, and it places the value proposition inside a profitable business model rather than beside it (Osterwalder et al., 2014). For a small business, the practical value of the pair is that it separates two questions that are usually asked as one. What is the customer trying to do, and what does the offer do about it? A seller who can answer both has a value proposition. A seller who can answer only the second has a list of features.

Pains are the obstacles and the risks the customer meets on the way to the job, and they are usually more specific than a seller expects. In the case, the pains named by customers included the disruption of an unplanned failure, the argument about who should have caught the problem, the difficulty of budgeting for a replacement that arrives without warning, and the exposure of being the person who did not know. Gains are the outcomes and the benefits the customer wants, and they range from the required, such as a working building, to the expected, such as a response within the promised time, to the desired, such as a year's warning before a major expense, to the unexpected, such as a report that makes the customer look good to somebody else. The four levels of gain are worth keeping apart, because a seller who delivers only the required and expected gains has delivered a service that will be compared on price.

The value map answers the profile with pain relievers and gain creators, and the discipline is to name them in the customer's terms rather than the seller's. A pain reliever for the owner-operator is not "a preventive maintenance program"; it is "you will not have to find a plumber on a Friday afternoon". A gain creator for the property manager is not "monthly reporting"; it is "you will have something to send the owners on the first of the month". The difference is not cosmetic. It is the difference between an offer the customer can evaluate and an offer the customer has to translate, and translation is work the customer will not do.

The fit between the profile and the map is the point of the exercise, and it is where most small businesses discover that they have been answering a question nobody asked. Nordec's contract, before this chapter's work, answered the functional job well and the social and emotional jobs by accident. The urgent response within ninety minutes answered the functional job of keeping the building running and, in part, the emotional job of not being surprised. The monthly preventive visit answered the functional job. The building health record, the one-page monthly report and the capital planning note answered the social and emotional jobs, and nobody had noticed, which is why they had been given away free. The interviews in section 2.8 are what made the fit visible.

WATCH OUT Asking customers what they want instead of what they did

A customer asked what they want will describe the product they imagine, and the description will be shaped by whatever they saw last, by what they think the seller can provide, and by a natural reluctance to admit that they have not thought about it. The answers are polite and nearly useless. The question that produces usable material is about the past: when did the problem last happen, what did it cost, who did you call, how long did it take, what did you do about it, and what would you have done if that option had not existed? Those questions produce facts, and facts can be counted. The difference is visible in the case followed through this book. Asked what they wanted, customers asked for more of what they already had. Asked what happened the last time a system failed, they described the phone calls, the waiting, the tenant complaints and the invoice, and those descriptions are what the value equation in section 2.5 is built from.

2.4 The Elements of Value

The elements of value are a vocabulary, not a checklist, and the distinction decides whether they help. A seller who takes the thirty consumer elements or the forty business elements and tries to deliver all of them will produce an offer that is expensive, unfocused and impossible to explain. The elements are useful in the opposite direction: they give a business a way to name what it is already doing, to see which elements its customers actually notice, and to choose the few that a segment will pay for. The research behind them is explicit that combinations of elements are associated with stronger loyalty and sustained growth (Almquist et al., 2016), which is an argument for deliberate combination rather than for accumulation.

The first step in using them is to establish the table stakes. In business markets, the elements begin with a category the authors call table stakes, and the name is exact: these are the elements a supplier must provide to be considered at all, and providing them well earns no credit. For a maintenance contract, the table stakes include a price the customer considers reasonable, the technical competence to do the work, the licensing and insurance to do it lawfully, and a response that arrives when promised. A business that fails at table stakes is not in the market, whatever else it offers. A business that succeeds at table stakes has earned the right to be compared, and nothing more.

The second step is to choose the few elements that matter to the segment. The interviews described later in this chapter produced a clear ranking for two groups of customers, and the ranking is not the same for both. For the owner-operators, the elements that mattered were urgent response, the monthly preventive visit and the capital planning note. For the property managers, they were urgent response, the one-page monthly report and, less often, the capital planning note. The building health record was named by four people out of twenty. Those counts are the raw material of a decision about what to put in an offer and what to leave out, and they are worth more than any general theory about what customers value.

The third step is to deliver the chosen elements well rather than to add more of them. This is the step small businesses skip, because adding is easier than improving. A monthly report that arrives late, or that contains the same three lines every month, is not an element of value; it is a habit that costs money and teaches the customer to ignore the report. The capital planning note in the case is a good example of the opposite: it is a single document, produced once a year at renewal, listing the equipment approaching the end of its life with a replacement window. It is not elaborate. It is delivered on time, it is specific to the building, and it answers a job the customer described in their own words, which is knowing a year ahead what will have to be replaced.

The fourth step is to verify the table stakes before claiming anything above them. A seller who promises a capital planning note while missing response commitments has built an offer on sand, because the customer's first experience of the business will be the failure. The order matters: table stakes, then the functional elements the segment requires, then the elements that distinguish. A business that reverses the order is selling a distinguishing element to a customer who is still deciding whether the supplier is competent.

Applied to a maintenance contract, the elements produce a short and useful list. The functional base is a building that works and a response that arrives. The ease of doing business is a single point of contact, a dispatcher who knows the building, and an invoice that does not need to be argued about. The individual elements are the report the property manager can forward and the note that lets an owner budget. The inspirational elements are largely absent, and that is correct: a maintenance contract is not a cause, and a seller who dresses it as one will be disbelieved. The value of the vocabulary is that it lets a business say which of those it is selling, and to whom, without pretending to sell all of them to everybody.

WHAT CHANGED From thirty consumer elements to forty business elements

The elements of value were first published for consumer markets in September 2016, as a pyramid of thirty elements in four levels running from functional at the base through emotional and life-changing to social impact at the top, drawn from consumer research and modeled on Maslow's hierarchy of needs (Almquist et al., 2016). In March–April 2018 the same research program was extended to business markets, where forty elements are grouped in five categories: table stakes, functional, ease of doing business, individual and inspirational (Almquist et al., 2018). The extension is not a translation of the consumer list. It adds a category at the bottom, because a business buyer has requirements a consumer does not, and it gives more weight to the personal considerations of the person doing the buying. The authors' observation that these personal considerations matter more as offers commoditize is the part that should change how a small business reads its own market: when several suppliers can meet the table stakes, the elements that decide are the ones about the buyer's own situation.

2.5 Value in Business Markets

Most suppliers in business markets cannot define or measure the value their offers deliver, and the consequence is predictable. Under purchasing pressure, a supplier who cannot demonstrate value is left with a price negotiation (Anderson & Narus, 1998). The remedy the authors recommend is a customer value model, built from field value assessments, that puts value in monetary terms against the customer's alternative. The recommendation is practical and it is uncomfortable, because it requires the supplier to learn the customer's numbers, which means asking questions that a seller trained to talk about their own product will find strange.

The value equation this book uses is a plain version of that idea, and it is worth stating in full before it is applied. The customer's incentive to buy is the offer's value minus its price, compared with the alternative's value minus its price. Written as a sentence rather than a formula, it says that a customer buys when the offer leaves them better off than the alternative does, and that the comparison has to be made in the customer's units rather than the seller's. The equation has three consequences that a seller should keep in view. The first is that lowering the price is not the only way to improve the customer's position; raising the value does the same work. The second is that the alternative's price is as important as the offer's, and a seller who does not know it is arguing blind. The third is that the equation can produce a negative number, and a negative number is information rather than a failure.

The equation also has a limit that must be stated whenever it is used, because a seller who presents it as the whole of value will be caught. The equation prices what can be priced. In the case followed through this book, it prices the preventive work the customer would otherwise buy and the unplanned interventions they would otherwise pay for, and it does not price the capital planning note, the building health record or the one-page monthly report, because those elements have no line in the customer's alternative. They are real, the interviews show that customers notice them, and the equation is silent about them. A seller who presents the equation without saying so is overstating the case; a seller who says so is credible about the part that is priced.

The figure below draws the equation for three situations, and it is worth describing in words for anyone listening rather than reading. Figure 2.1 shows three pairs of horizontal bars. The first pair is the typical building of about 4,500 square metres: the upper bar is what the patchwork of trades costs the customer, and the lower bar is what Nordec's contract costs, with the difference written at the end of the pair as a gain of $1,516. The second pair is the small building of about 2,500 square metres on the current contract: the upper bar is the alternative at $6,620, the lower bar is the contract at $7,080, and the difference is written as a loss of $460. The third pair is the same small building at the Essential price: the upper bar is unchanged at $6,620, the lower bar falls to $5,880, and the difference is written as a gain of $740. The three pairs together make the chapter's argument visible: the same offer, the same alternative, and three different answers depending on the building and the price.

Figure 2.1 — The customer's incentive to buy. Against the patchwork of trades, the contract gives a typical building $1,516 a year and costs a small building $460; the Essential tier turns the small building's loss into a gain of $740.

The following table sets out the same arithmetic in rows, so that each line of the equation can be read on its own. Read it as a working document rather than a summary: the rows are the ones a seller would fill in with a customer, and the last two rows are the ones that decide whether the offer is bought.

Table 2.2 — The value equation for two Nordec buildings

Two features of the table deserve comment. The first is that the typical building's gain of $1,516 is real but modest against a contract of $8,544, which means the offer's case rests partly on the elements the equation cannot price. The second is that the small building's loss of $460 is not a rounding error; it is the arithmetic behind two lost proposals, and the owners who lost them said as much in plain words when they described the contract as more than the building needed. A seller who had run this equation before writing those proposals would have known the answer in advance.

DO THE MATH The value equation for a small building

Illustrative figures; the case is fictional. Step 1 — The small building of about 2,500 square metres buys its preventive work from a patchwork of trades at $2,700 a year. Step 2 — The same building pays for seven unplanned interventions a year at an average of $560 each, which is 7 × $560 = $3,920. Step 3 — The alternative therefore costs the customer $2,700 + $3,920 = $6,620 a year. Step 4 — The current contract at $590 a month costs $590 × 12 = $7,080 a year, which is $7,080 − $6,620 = $460 more than the alternative: the customer loses $460 by buying, before counting the elements the equation cannot price. Step 5 — At the Essential price of $490 a month, the contract costs $490 × 12 = $5,880 a year, and the customer gains $6,620 − $5,880 = $740. The same building, the same alternative, and a different answer at each price.

THE RULE A saving promised to a customer must be proven

A value equation shown to a buyer is a claim about the performance or efficacy of an offer, and in Canada such a claim has a legal standard attached to it. Section 74.01(1)(b) of the Competition Act, R.S.C. 1985, c. C-34, treats as reviewable conduct a representation to the public of the performance, efficacy or length of life of a product that is not based on an adequate and proper test, and the burden of proving that the test was adequate and proper rests on the person who makes the representation. The provision applies in every province and to sales to businesses as well as to consumers. For a seller who presents a value equation, the practical reading is that the numbers must come from somewhere defensible: the customer's own records, published data, or an assessment the seller can describe. The book explains the law and does not replace the advice of a lawyer.

QUÉBEC & CANADA NOTE Benchmarking a customer's costs with federal data

A value equation needs the customer's numbers, and a customer does not always have them. Innovation, Science and Economic Development Canada publishes a free financial performance data tool that covers more than 1,000 industries with more than 30 benchmarks, for businesses with revenues of $30,000 to $5 million and of $5 million to $20 million. Its reports give revenue and expense detail, income statement and balance sheet averages and ratios, and a business can compare its own figures with them; the tool is built on Statistics Canada data (Innovation, Science and Economic Development Canada, 2026b). For a seller, the tool has a specific use: it allows a first estimate of what a typical customer in an industry spends on a category of expense, so that a value equation can begin from public figures rather than from a guess. The estimate is a starting point for a conversation with the customer, not a substitute for the customer's own records, and the seller should say which of the two is being used.

2.6 Measuring What Customers Value

Value is measured by asking, and asking is a skill rather than a conversation. The instrument this book uses is the value interview, and its defining feature is that it asks about the past. A customer describing what they want will describe a product; a customer describing what happened the last time the problem occurred will describe a cost, a delay, a phone call and an invoice, and those are the materials a value equation is built from. The interview also has a second purpose that is easy to miss: it tells the seller which elements the customer would miss if they disappeared, which is a different question from which elements the customer likes.

The question that produces the second answer is the top-three question. Each person is asked to name the three elements of the offer they would miss most if the offer were withdrawn, and the answers are counted across groups. The counting is what turns a set of conversations into a decision, because a single customer's enthusiasm is not evidence and twenty customers' rankings are. The method is deliberately crude, and its crudeness is a feature: it can be performed by an owner and an administrator in a few weeks, without a research budget, and it produces a table that a business can act on.

A theory that helps read the answers to the top-three question is the theory of attractive quality that Noriaki Kano and colleagues published in Japanese in 1984 (Witell et al., 2013). The theory classifies attributes by how their presence relates to satisfaction. Must-be attributes dissatisfy when they are absent and add nothing when they are present, which is why a supplier who meets them earns no credit. One-dimensional attributes are the ones where more is better, and satisfaction rises with performance. Attractive attributes delight when they are present and do not dissatisfy when they are absent, which is why they can be added without penalty and why they lose their power once competitors copy them. Two further categories complete the scheme: indifferent attributes, which the customer does not notice either way, and reverse attributes, which actively reduce satisfaction when present.

The categories explain a pattern that a small business meets constantly. The urgent response within ninety minutes is a must-be attribute for most customers: they expect it, they do not praise it, and they would leave if it stopped. The monthly preventive visit is one-dimensional: more thorough visits are better, and the customer notices the difference. The capital planning note is attractive: it delights the customers who receive it, it costs nothing in satisfaction for the customers who do not, and it is the element most likely to be named by an owner who has read one. The building health record, named by four people out of twenty, is closer to indifferent for most customers and attractive for a few, which is a finding rather than a failure.

Stated importance is not enough on its own, and the reason is that customers rate everything as important when asked directly. A customer asked whether response time matters will say yes; a customer asked whether a monthly report matters will also say yes, and the two answers cannot be compared. The top-three question forces a ranking by making the customer give something up, and the ranking is what a business can use. The same logic explains why the interviews in the case produced a surprise: the one-page monthly report was named by five of the six property managers and by one of the eight owner customers, which is a difference no satisfaction survey would have shown.

The measurement of preferences when an offer has many attributes is the subject of conjoint analysis, an approach that has developed methods for measuring preferences across a large number of attributes (Green & Srinivasan, 1990). It is a more powerful instrument than the interview and it requires more design, more respondents and more analysis, and this book leaves it to Chapter 8, where the testing of an offer is treated. The interview comes first for a practical reason: it produces the attributes that a later study would test, and a business that has not yet had twenty conversations is not ready to design a study.

The following table sets out what the twenty interviews of the case named among each person's top three, counted by group. Read it as the raw material of a packaging decision: the totals in the last column are what a business acts on, and the differences between the columns are what tell it that it is serving more than one kind of customer.

Table 2.3 — What twenty interviews named among their top three

METHOD Five steps to a value interview

First, choose the three groups and invite five to eight people in each: customers who buy the offer, customers who buy a different kind of offer, and people who declined a proposal in the last two years, because the last group is the only one that can explain a refusal. Second, ask about the last time the problem happened and what it cost, before showing anything or describing the offer, so that the answers are about the customer's experience rather than the seller's proposal. Third, ask each person to name the three elements they would miss most if the offer were withdrawn, and write the three down in the order they are given. Fourth, write the answers in the customer's own words on the same day, because a paraphrase written a week later has lost the sentence that mattered. Fifth, count the answers, compare the groups, and write one page of findings that names the elements by group and states which of them the business is not currently selling.

TOOLS OF THE TRADE What a value interview needs

Five functions carry a value interview program, and none of them requires a specialist product. The first is a short interview guide, one page with the questions in order, so that twenty conversations produce comparable material. The second is a way to schedule the conversations, whether by telephone, by message or in person, with a reminder the day before. The third is a recording or a notes template, chosen according to what the customer will accept, with the customer's permission asked before anything is recorded. The fourth is a spreadsheet to count the answers, with one row per person and one column per element, so that the totals in Table 2.3 can be produced by counting rather than by memory. The fifth is a place to keep the customers' words, because the sentences are the material a value proposition is written from and they are lost if they are stored only as a number.

WHO DOES THIS The owner and the people who hear customers

In a small business, the value interviews are run by the owner together with the person who answers the phone, and the pairing is deliberate. The owner has the standing to ask a customer for an hour and the authority to act on the answer; the person who answers the phone has the knowledge of what customers actually say when something goes wrong, which is where the pains live. In the case followed through this book, Sophie Nadeau and the administrator ran the twenty interviews together, which meant that the person writing the answers down already knew which of them were new. In a larger business, the work belongs to whoever holds the customer experience or product role, and the same pairing applies: the person with the authority and the person with the contact. What does not work is delegating the interviews to somebody who has never spoken to the customer, because the follow-up question is the whole instrument.

2.7 Painkillers, Vitamins and the Order of Value

Book A1 (From Idea to Business Model) draws a distinction that is worth carrying into this chapter, because it decides what a business should sell first. A painkiller is an offer that answers a problem the customer already knows they have and is already paying to solve. A vitamin is an offer that makes something better without answering a problem the customer has named. A first venture should almost always be a painkiller, because a customer who is already spending money on a problem is a customer who can be persuaded to spend it differently, and a customer who has not noticed a problem is a customer who has to be educated before they can be sold to. A vitamin is what an established business adds once it has customers who trust it, because trust is what makes a customer willing to accept something they did not ask for.

The distinction is not a judgment about which offers are good. It is a judgment about sequence, and the sequence is what a small business gets wrong. A business that leads with a vitamin is asking a customer to accept a new category of spending, and the customer's answer is usually that they will think about it. A business that leads with a painkiller and adds vitamins afterwards is doing something different: it is deepening an offer the customer has already accepted, and the customer's answer is usually yes, because the marginal decision is small and the trust is already there. The same element can be a painkiller in one market and a vitamin in another, and the seller's job is to know which one they are holding.

At Nordec, the painkiller is the ninety-minute response commitment on urgent contract calls, day and night. It answers a problem the customer already has and already pays for, because a building arriving from a patchwork of trades reports about eleven unplanned interventions a year and pays for each of them. The commitment is the element that changes the customer's position most directly, and it is the element named by eighteen of the twenty people interviewed. The monthly preventive visit is the second painkiller, named by fourteen, and it is the element that prevents some of the interventions from happening at all. Those two elements are the base of the offer, and a business that removed them to add something more interesting would lose the customers it has.

The capital planning note and the one-page monthly report are vitamins, and the interviews show exactly how a vitamin behaves. The note was named by seven of the eight owner customers, by two of the six property managers and by three of the six prospects who had seen a sample, which is a strong result for an element that no customer had asked for before it existed. The report was named by five of the six property managers and by one of the eight owner customers, which is a strong result in one group and a weak one in the other. Neither element answers a problem the customer was already paying to solve. Both make something better, and both are valued by the customers who have already accepted the contract and learned to trust it.

The order in which value should be arranged in an offer follows from the theory of attractive quality reviewed in section 2.6 (Witell et al., 2013). The must-be attributes come first, because their absence makes everything else irrelevant: a response that arrives within the promised time, a technician who knows the building, an invoice that does not have to be argued about. The one-dimensional attributes come second, because they are where the customer compares suppliers: the thoroughness of the preventive visit, the speed of the response, the number of things the visit covers. The attractive attributes come third, because they are what the customer did not expect: the capital planning note, the report that can be forwarded, the record that survives a change of technician. A business that puts the attractive elements first is selling delight to a customer who is still deciding whether the supplier is competent.

The order also has a commercial consequence that a small business should notice. The must-be and one-dimensional elements are the ones a customer will compare across suppliers, which means they are the ones that end up in a price negotiation. The attractive elements are the ones that are not compared, because the customer's other suppliers do not offer them, and they are therefore the elements that protect a price. In the case followed through this book, the capital planning note is the clearest example: it is produced once a year, it costs three hours of work, and it is the element that seven of eight owner customers named among their top three. A business that gives it away free has given away the element that most protects its price, which is precisely what happened before this chapter's work.

2.8 The Case's Value Interviews

The twenty interviews were run by Sophie Nadeau and the administrator over several weeks, in three groups chosen before the first conversation. Eight were owner-operators who held a contract, six were property managers who held contracts across several buildings, and six were owners of small buildings who had declined a proposal in the last two years. The third group was the one that took the longest to arrange and the one that produced the most useful material, because a customer who has said no can explain the no, and a customer who has said yes usually cannot explain why. The invitation was a telephone call, the conversation lasted about an hour, and the questions followed the order set out in the method callout above: the last time the problem happened, what it cost, what they did, and then the three elements they would miss most.

The first question produced the material that the value equation is built from. Asked what happened the last time a system failed, the owner-operators described the phone calls, the waiting for a trade to arrive, the tenant complaints and the invoice. The property managers described the same events with an additional element, which was the conversation with the owners they report to. The prospects described the events and the decision not to buy, usually in the same sentence, and their reason was consistent: the contract was more than the building needed. None of those answers was a request, and all of them were facts.

The top-three question produced the counts in Table 2.3, and the counts are worth reading group by group. Every owner customer and every property manager named three elements, which is twenty-eight answers from fourteen people. The prospects named twelve elements in all, because several of them named the price instead of an element, which is itself a finding: a customer who has declined is telling the seller that the offer's elements did not reach the level where a comparison of elements was worth making. Urgent response within ninety minutes was named by eighteen people out of twenty, which makes it the closest thing to a universal element in the case. The monthly preventive visit was named by fourteen, the capital planning note by twelve, the one-page monthly report by six and the building health record by four.

The surprise was in the distribution rather than in the totals. The monthly report was named by five of the six property managers and by one of the eight owner customers, a difference of five to one between two groups buying the same contract. The capital planning note ran the other way: seven of the eight owner customers named it, against two of the six property managers. Read together, the two lines say that the same contract is carrying two different offers, and that neither group is receiving the one it values most. The owner-operators were receiving a report they did not read; the property managers were receiving a note they had not asked for. The building health record, named by four people, was the element that the business had invested most in creating, at 160 hours and $3,200, and the interviews did not support treating it as a headline element.

The unprompted requests, gathered in the same conversations, pointed in the same direction. Five customers asked for a quarterly review meeting, four asked for help with replacement projects, and three property managers asked for a single report across their buildings, which is twelve requests in all. The requests are not the same as the top-three answers, and the difference is instructive: the top-three answers describe what customers would miss, and the requests describe what customers would like next. A business that confuses the two will add services that nobody would miss and neglect the elements that hold the contract together.

The value equation for the small building is where the interviews and the arithmetic met. The building of about 2,500 square metres spends $2,700 a year on preventive work bought from trades and pays for seven unplanned interventions at an average of $560, which is $3,920, for a total of $6,620. The current contract at $590 a month costs $7,080, which is $460 more than the alternative, and the two owners who had been lost on price had said exactly that in their own words. At the Essential price of $490 a month, the same building pays $5,880 and gains $740. The equation does not price the capital planning note, the building health record or the one-page monthly report, and it says so, which is why the small building's case cannot rest on the arithmetic alone.

The conclusion Sophie drew from the twenty interviews is the conclusion this chapter hands to the rest of the book. Two kinds of customer want different things from the same contract. The owner-operator buys certainty and a year's warning: the response commitment, the preventive visit and the capital planning note. The property manager buys proof to show the owners: the response commitment, the monthly report and, for a portfolio, a consolidated view. And the small building's owner pays for a contract whose value, measured against the alternative, is lower than its price, which is not a failure of selling but a failure of design. Chapter 4 designs the answer, which is three tiers with different contents at different prices, and it can do so only because this chapter established the facts. Nothing in this chapter decides what the tiers should be; it establishes who values what, and it leaves the packaging to the chapter that follows.

CASE IN POINT Nordec Entretien inc.: twenty interviews, two customers

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. Sophie Nadeau and the administrator interviewed twenty people: eight owner-operator customers, six property managers and six owners of small buildings who had declined a proposal in the last two years. Urgent response within ninety minutes was named by eighteen of the twenty; the monthly preventive visit by fourteen; the capital planning note by twelve; the one-page monthly report by six; the building health record by four. The report was named by five of the six property managers and by one of the eight owner customers. The note was named by seven of the eight owner customers and by two of the six property managers. The unprompted requests were a consolidated view across buildings, a quarterly review, a line for tenants' non-urgent requests, a smaller contract and help with replacement. The two jobs, in the customers' words, were to keep the building running and the tenants quiet without becoming a maintenance expert and to know a year ahead what will have to be replaced, and to show the owners that their buildings are under control and never be the last to know about a failure.

IN LA MACHINE Where customers' words can be kept

The application copy read for this book offers a place to run the interviews and a place to keep what they produce (La Machine, 2026). A general form builder at business/marketing/forms can carry an interview sign-up or a short survey, and the answers are kept with the form at business/marketing/forms/responses, which is enough to schedule twenty conversations and to hold the short written answers. One record per customer lives at business/marketing/companies, with the people attached to it at business/marketing/contacts, and both carry notes, which is where the sentences a customer used can be kept beside the customer they came from. What the platform does not have belongs in the same breath: there is no survey tool and no net promoter tool, only the general form builder, so the counting described in this chapter is done in a spreadsheet outside the platform. There is also no customer journey map and no service blueprint, which is why the work of Chapter 3 is drawn elsewhere.

Costly Mistakes

• Assuming the customer values what the business produces. A seller who counts the visits, the hours and the parts will overstate the offer by exactly the amount of everything the customer gives up that does not appear on an invoice.
• Arguing from the seller's cost. The cost of serving a building is the seller's problem, and a customer who has never seen it has no reason to accept it as the basis for a price.
• Never establishing the alternative. A value equation without the customer's alternative is a price with a story attached, and the customer will supply the missing half in the negotiation.
• Asking customers what they want. A customer asked what they want describes the product they imagine; a customer asked what happened the last time the problem occurred describes a cost that can be counted.
• Treating the elements of value as a checklist. An offer that claims every family at once has chosen none of them, and a customer who cannot tell what is being sold will compare on price.
• Leading with a vitamin. An element the customer did not ask for is worth adding to an offer the customer has already accepted, and it is worth very little as the reason to buy.
• Giving away the element that protects the price. The capital planning note was named by seven of eight owner customers and cost three hours a year, and it had been given away free since the day it was created.

Key Takeaways

• Value is the customer's judgment of what they receive against everything they give up, and it is always made against the next best alternative rather than against the seller's cost.
• The families of value run from functional through emotional to life-changing and social impact in consumer markets, and from table stakes through functional, ease of doing business, individual and inspirational in business markets (Almquist et al., 2016; Almquist et al., 2018).
• A customer hires an offer to make progress in a circumstance, and the job carries functional, social and emotional dimensions that a customer profile of jobs, pains and gains is built to capture (Christensen et al., 2016; Osterwalder et al., 2014).
• The value equation this book uses says that the customer's incentive to buy is the offer's value minus its price, compared with the alternative's value minus its price, and it prices only what can be priced.
• The twenty interviews of the case named urgent response eighteen times, the monthly preventive visit fourteen, the capital planning note twelve, the monthly report six and the building health record four, and the report and the note were valued by opposite groups.
• The same contract carries two different offers: the owner-operator buys certainty and a year's warning, and the property manager buys proof to show the owners, which is why Chapter 4 designs three tiers rather than one.

Your Action Plan

1. Write down the alternative your customers actually have. For each segment you serve, describe in one paragraph what the customer would do if your offer did not exist, including the trades, the bundles, the internal arrangement and the option of doing nothing. Deliverable: one paragraph per segment. Verification: a customer in each segment confirms the description in their own words.
2. Build the value equation for one real customer. Fill in the alternative's cost, the offer's cost and the difference, using the customer's records where they exist and published data where they do not. Deliverable: one completed equation with its sources named. Verification: the equation states plainly which elements it does not price.
3. Run the twenty interviews in three groups. Invite five to eight customers, five to eight customers of a different offer and five to eight people who declined in the last two years, and ask about the last time the problem happened before showing anything. Deliverable: twenty sets of notes in the customers' own words. Verification: every conversation produced at least one cost figure and three named elements.
4. Count the top-three answers and compare the groups. Build the table with one row per element and one column per group, and read the differences between the columns rather than only the totals. Deliverable: one counted table and one page of findings. Verification: the findings name at least one element the business is not currently selling and one it is selling to the wrong group.
5. Order the elements of your offer. List the must-be elements, the one-dimensional elements and the attractive elements, and check that the offer leads with the first two and adds the third. Deliverable: one ordered list of elements. Verification: every attractive element is delivered well and none of them is the reason the customer is asked to buy.

In La Machine

Four routes in the application copy read for this book touch the work of this chapter (La Machine, 2026). The general form builder at business/marketing/forms can carry an interview sign-up or a short survey, and the answers are kept with the form at business/marketing/forms/responses, which is enough to invite twenty people and to hold what they write. One record per customer lives at business/marketing/companies, with the people attached to it at business/marketing/contacts, and both carry notes, which is where the sentences a customer used can be kept beside the customer they came from. Read together, the four pages cover the scheduling and the storage of a value interview program, and they cover it in the same place the rest of the customer record lives.

What the platform does not do belongs in the same breath. La Machine has no survey tool and no net promoter tool, only the general form builder, so the counting of the top-three answers is done in a spreadsheet outside the platform. It has no customer journey map and no service blueprint, which is why the work of Chapter 3 is drawn elsewhere, and no experiment or A/B testing tool, which is why the testing of Chapter 8 is designed outside it. The forms page is a general builder rather than a research instrument, and a business that needs branching questions, quotas or weighting will find that it has to do that work by hand. Book N1 maps the modules the platform does cover, and the rule stated there applies here as well: before counting a family as covered, check that the page does what the business needs.

Self-Check

1. Why is value perceived rather than delivered, and what does the customer compare an offer with?
2. What are the families of value in consumer and business markets, and which of them can be put in dollars?
3. What is a job to be done, and what are the two jobs the customers of the case described in their own words?
4. What is the value equation this book uses, and what does it leave out?
5. What did the twenty interviews find, and what conclusion did the case draw from them?

Answers. 1. Value is perceived because it is formed in the customer's mind from their circumstances, their history and their alternatives, and no amount of correct delivery can install it; the customer compares the whole offer with the next best alternative, which may be another supplier, a bundle, a set of trades billed by the call, an internal arrangement or doing nothing, and never with the seller's cost. 2. In consumer markets the elements run from functional at the base through emotional and life-changing to social impact at the top, and in business markets they are grouped as table stakes, functional, ease of doing business, individual and inspirational (Almquist et al., 2016; Almquist et al., 2018); the family that can be put in dollars is economic value, which is the part of the judgment a buyer can express as a difference in cost or revenue against the alternative. 3. A job to be done is the progress a customer tries to make in a circumstance, and it is rarely only functional, carrying social and emotional dimensions as well (Christensen et al., 2016); the owner-operator wants to keep the building running and the tenants quiet without becoming a maintenance expert and to know a year ahead what will have to be replaced, and the property manager wants to show the owners they report to that their buildings are under control and never to be the last to know about a failure. 4. The customer's incentive to buy is the offer's value minus its price, compared with the alternative's value minus its price; the equation prices what can be priced, and in the case it leaves out the capital planning note, the building health record and the one-page monthly report, because those elements have no line in the customer's alternative. 5. The interviews named urgent response within ninety minutes eighteen times out of twenty, the monthly preventive visit fourteen, the capital planning note twelve, the one-page monthly report six and the building health record four, and the report was named by five of six property managers against one of eight owner customers while the note ran the other way; the conclusion was that two kinds of customer want different things from the same contract, that the owner-operator buys certainty and a year's warning while the property manager buys proof to show the owners, and that the small building's owner pays for a contract whose value against the alternative is lower than its price, which sends the reader to Chapter 4.

Chapter References

Almquist, E., Senior, J., & Bloch, N. (2016, September). The elements of value. Harvard Business Review. https://hbr.org/2016/09/the-elements-of-value

Almquist, E., Cleghorn, J., & Sherer, L. (2018, March–April). The B2B elements of value. Harvard Business Review. https://hbr.org/2018/03/the-b2b-elements-of-value

Anderson, J. C., & Narus, J. A. (1998, November–December). Business marketing: Understand what customers value. Harvard Business Review. https://hbr.org/1998/11/business-marketing-understand-what-customers-value

Christensen, C. M., Hall, T., Dillon, K., & Duncan, D. S. (2016, September). Know your customers' "jobs to be done." Harvard Business Review. https://hbr.org/2016/09/know-your-customers-jobs-to-be-done

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

Green, P. E., & Srinivasan, V. (1990). Conjoint analysis in marketing: New developments with implications for research and practice. Journal of Marketing, 54(4), 3–19. https://doi.org/10.1177/002224299005400402

Innovation, Science and Economic Development Canada. (2026b). Financial performance data (page modified July 2, 2026). Retrieved September 22, 2026, from https://ised-isde.canada.ca/site/financial-performance-data/en

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: business/marketing/forms, business/marketing/forms/responses, business/marketing/companies and business/marketing/contacts.

Osterwalder, A., Pigneur, Y., Bernarda, G., & Smith, A. (2014). Value proposition design: How to create products and services customers want. Wiley.

Witell, L., Löfgren, M., & Dahlgaard, J. J. (2013). Theory of attractive quality and the Kano methodology – the past, the present, and the future. Total Quality Management & Business Excellence, 24(11–12), 1241–1252. https://doi.org/10.1080/14783363.2013.791117

Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value: A means-end model and synthesis of evidence. Journal of Marketing, 52(3), 2–22. https://doi.org/10.1177/002224298805200302

Chapter 3 — Service Design and the Customer Journey

A service is experienced as a sequence of moments, most of them invisible to the business that delivers it, and designing the journey and the blueprint behind it is how the promise of an offer survives contact with a Tuesday night, a missing part and a tired technician.

What You'll Be Able to Do

• Draw the customer journey of an offer across the stages before, during and after the purchase, naming for each stage the touchpoint, the thing that can go wrong and the person inside the business who owns it.
• Identify the few moments of truth that decide the customer's judgment of an offer, using interviews, complaints and renewals as the evidence rather than the opinion of the people who deliver the service.
• Build a service blueprint in one afternoon, in five lanes and three lines, so that every promise made in front of the customer is matched to the backstage work and the support process that make it deliverable.
• Find the fail points, the waits and the endings of a service moment, and give each one either a design change that removes it or a recovery that limits the damage when it happens.
• Measure service quality with a small number of measures that follow the customer's judgment rather than the convenience of the person producing the report, and read those measures against a target.
• Design the onboarding of a new customer as a dated checklist with a named owner, so that the first thirty days produce the first moment of truth instead of leaving it to chance.

Why It Matters

An offer is a written promise and a service is the process that keeps it, and the distance between the two is where most customer disappointment is born. Nordec Entretien inc. promises a commercial building owner that a technician will be on site within ninety minutes of an urgent contract call, day or night, and that the monthly preventive visit will keep the building's systems from failing in the first place. Neither promise is delivered by the contract. Both are delivered by a dispatcher who answers the telephone at eleven at night, decides in thirty seconds whether the call is urgent, finds a technician who is not already on a job, checks whether the van holds the part, drives to a building on Montréal's north shore, and writes down what was found before leaving. Every one of those steps is a design decision, whether or not anybody made it deliberately.

The case shows what the averages hide and why this chapter exists. Over the last twelve months, 253 urgent contract calls came in, and 243 of them had a technician on site within ninety minutes, which is 96% against a target of 95%. Read alone, that figure says the promise is kept. Beside it sits a second figure: 199 of the 253 calls were resolved at the first visit, or 78.7% against a target of 88%. That means 54 calls needed a second visit, and the reasons are not mysterious — 37 of them happened because a part was not on the van, and 17 because the work ran past the two hours included in the contract and needed the owner's approval. Nine invoices for beyond-scope work ended in a dispute. The promise of the arrival was kept and the promise of the resolution was not, and no single measure would have shown it.

Design is also the cheapest form of improvement available to a small business, because most of what it changes is a habit rather than a purchase. The second visits cost $81 each in technician time, and the 37 visits caused by a missing part cost $2,997 a year. Putting the twenty parts most often missing onto the two vehicles costs $3,200 once and is expected to avoid 26 of those 37 visits, which is $2,106 a year and a payback of about a year and a half. The rest of the redesign costs nothing at all: a technician who calls the owner with an estimate before passing the two included hours, a dispatcher who says out loud what will happen to a non-urgent request received after hours, and a closing note written into the building record on site rather than in the office the next morning. This chapter is about finding those points, drawing them, and deciding what each one should do.

In This Chapter

Section 3.1 explains why a service must be designed deliberately and what happens when it is not. Section 3.2 sets out the customer journey across the stages before, during and after the purchase, and the table of Nordec's seven stages with their touchpoints and owners. Section 3.3 identifies the moments of truth and shows how to find them in evidence the business already has. Section 3.4 builds the service blueprint in five lanes and three lines, and gives the afternoon procedure for drawing one. Section 3.5 covers fail points, waits and endings, including what behavioral science adds to the psychology of the service encounter. Section 3.6 turns quality into a handful of measures that follow the customer's judgment. Section 3.7 designs the onboarding of a new customer. Section 3.8 returns to the case and draws the blueprint of an urgent call, with its three fail points and their fixes.

3.1 Why Services Must Be Designed

A service is a process the customer takes part in, and that single difference from a product changes everything about how it must be prepared. A product can be built, inspected, stored and photographed before anybody buys it, and the customer meets it in a finished state. A service is produced in the customer's presence, in a sequence of moments that begins long before the purchase and continues long after it, and every one of those moments is a place where the promise of the offer either holds or breaks. Nordec's contract says that a technician will be on site within ninety minutes of an urgent call. Nothing in the document makes that happen. It happens because a dispatcher, four technicians, two vehicles, a dispatch board and a van stock were arranged in a particular way, and the arrangement is either designed or accidental.

Left undesigned, a service is still designed — by whoever takes the call. This is the sentence that makes the rest of the chapter necessary. If nobody has decided what counts as urgent, the dispatcher decides at eleven at night, and decides differently on a busy Tuesday than on a quiet Thursday. If nobody has decided when the owner hears about work that passes the two included hours, the technician decides whether to ask, and the owner decides at the invoice. If nobody has decided what a non-urgent after-hours request means, the owner waits for a technician who is not coming and concludes the night that the contract does not do what it says. None of those outcomes comes from a bad intention. They come from a service that grew instead of being drawn.

The idea that a service can be drawn and engineered like a product is older than most business owners assume (Shostack, 1984). The article that made the argument opened on the everyday failures that nobody had designed: a laundered shirt returned with a broken button, a car repair that does not cure the rattle. It asked why the operations behind those experiences had never been laid out and examined when the manufacture of the object had been examined for a century, and it proposed to design services deliberately, to treat the intangible part of an offering as something that can be specified, drawn and improved. A later article credits that work as the origin of service blueprinting (Bitner et al., 2008), and the authors of that later article make the case for the discipline with a blunt observation of their own: services account for about 80% of United States gross domestic product, and yet innovation in services has remained less disciplined than innovation in manufacturing. The service economy, in other words, is the larger part of the economy and the less carefully engineered part of it.

For a small business, the discipline has a second and more practical justification: the business is the service. A manufacturer can improve a component and let the component carry the improvement to the customer. A maintenance contractor has no component. The customer's experience of the offer is the dispatcher's voice, the technician's arrival time, the state of the van, the invoice, the report and the renewal meeting, and there is nothing else. This is why the book places this chapter between Chapter 2, which listens to what customers value, and Chapter 4, which packages that value into tiers: the journey is where the value that has been identified is actually delivered, and it is the material the packaging has to promise.

DEFINITION Customer journey, touchpoint, moment of truth and service blueprint

The customer journey is the sequence of stages a customer passes through in relation to a business, from the first awareness of it to the decision to renew or to leave, including the stages that happen before any money changes hands and the long stretches between purchases. A touchpoint is any point at which the customer and the business meet: a telephone call, an invoice, a visit, a report, a web page, a referral from another owner. A moment of truth is a touchpoint that carries disproportionate weight, so that the customer's overall judgment of the offer forms there, and there are usually only a few of them in a given journey. A service blueprint is a diagram of one service moment, drawn in horizontal lanes, showing what the customer does, what the employees the customer sees do, what the employees the customer does not see do, what the systems and suppliers behind them contribute, and the physical evidence the customer encounters, together with the lines that separate those lanes.

WHAT CHANGED From the 1984 blueprint to the journey of 2016

Three moments in the published work behind this chapter changed the way a service is drawn. A first moment came in 1984, when an article argued that services could be designed deliberately, in a drawing, rather than left to accumulate from habit (Shostack, 1984), and a later article credited that work as the origin of service blueprinting (Bitner et al., 2008). A second moment came in 2008, with the practical technique of the blueprint itself: a customer-focused drawing with physical evidence, customer actions, onstage and backstage employee actions and support processes, separated by the line of interaction, the line of visibility and the line of internal interaction, with the customer's actions drawn first and each crossing of the line of interaction treated as a moment of truth (Bitner et al., 2008). A third moment came in 2016, which moved the frame outward from the single encounter to the whole journey: customers now interact with a firm through many touchpoints, in several channels and media, their experience has become more social, and delivering a positive one requires several functions of the firm, and sometimes external partners, to work together (Lemon & Verhoef, 2016). The blueprint draws one moment. The journey places that moment among the others.

3.2 The Customer Journey

A journey map is drawn from the customer's side of the counter, and that is the only rule that matters when one is built. It starts before the customer has decided anything, it continues through the purchase, and it extends into the months and years that follow, which is where most of the value of a recurring service is actually produced. The synthesis of the research on customer experience makes the point that the modern customer meets a firm through many touchpoints, in several channels and media, and that the experience has become more social — a referral from another building owner is part of the journey, and so is a remark made in a professional association (Lemon & Verhoef, 2016). The same article argues that delivering a positive experience requires the firm to integrate several business functions and sometimes external partners, which is the quiet reason a journey map is uncomfortable to draw: it shows a customer walking across departments that do not talk to one another.

For Nordec, the journey has seven stages, and the business only began to see them as one sequence after the strategy of Book A3 (Strategy and Competitive Positioning) defined what the company would and would not do. The first stage is discovery, where a building owner hears about the company from another owner, receives a letter about the Building Check described in Chapter 7, or comes through the proposal that followed an enquiry. The second is evaluation, where the owner reads the proposal and, since the strategy year, may be shown a sample capital planning note so that the promise of a yearly note is something they can see rather than imagine. The third is signature. The fourth is onboarding, the first thirty days. The fifth is the monthly routine of the preventive visit and the one-page report. The sixth is the urgent call, which may arrive at any point in the year and is the stage that decides whether the customer believes the contract. The seventh is renewal, where the capital planning note is delivered and the next year is agreed.

Naming the stages is useful and not sufficient. The work of the journey map is to attach to each stage the touchpoint through which the customer experiences it, the thing that can go wrong there, and the person inside the business who answers for it. That last column is the one that changes behavior in a small company, because a touchpoint with no owner is a touchpoint that everyone assumes somebody else is handling. In the case followed through this book, the journey map also exposed a stage the business had never considered: the twelve months between the signature and the renewal, during which the only things the customer receives are the visit sheets, the monthly reports and the capital planning note at the end. A recurring service is sold on a promise about that long middle, and the middle is where the design work of this chapter mostly happens.

The following table sets out Nordec's journey in seven stages, with what the customer does at each one, the touchpoint through which it happens, what can go wrong and the person who owns it. Read the fourth and fifth columns together: the failures that matter are not random, they cluster at the stages where the owner is unclear, and each of them belongs to somebody.

Table 3.1 — The journey of a Nordec customer

When the dispatcher, the lead technician and Sophie drew the journey on a wall in the depot, they produced six stages in under an hour, because those were the stages they worked in every day. It took the administrator to name the seventh. Between the first visit and the renewal there were twelve months in which the customer received visit sheets, a monthly report and nothing else, and in which the only contact from the business came when something broke. That long middle had never been treated as part of the offer, and the twenty interviews described in Chapter 2 showed why it mattered: the capital planning note was named among each person's top three elements by 12 of the 20 people interviewed, and the monthly report by 6, five of whom were property managers. The note and the report were not extras appended to a maintenance contract. They were the content of the middle, and without them the customer experienced a service that consisted of emergencies.

WATCH OUT Drawing the journey from the inside

The most common way a journey map fails is that it is drawn from the organization chart. A map with a row for dispatch, a row for field work, a row for billing and a row for accounts describes the business's departments and tells the customer nothing, because the customer does not experience departments; the customer experiences a sequence in which a person answers, a van arrives, a note is left and an invoice follows. The remedy is a source rule. A journey map is drawn from interviews with customers, from the complaint log, from the disputed invoices and from what happened at renewal, and where those sources disagree with what the team believes, the sources win. Two further tests keep the map honest. The first is to write each stage from the customer's point of view, so that "the owner waits for a call back" appears as a stage rather than as an internal handover. The second is to mark, for every touchpoint, whether the customer can see the person behind it, because the touchpoints where nobody is visible are where confidence leaks away.

3.3 Moments of Truth

Out of the twenty or thirty touchpoints a customer meets in a year, only a few carry weight, and finding those few is the most valuable hour a small business can spend on its offer. A moment of truth is not the moment the business cares most about; it is the moment the customer's judgment forms. In a maintenance contract, the moments that decide whether the owner renews are rarely the ones an operations manager would choose. The arrival within ninety minutes is a promise kept or broken in a way the owner can count, and the customer's judgment of the contract is largely made in the first urgent call, because the entire promise of the ninety minutes exists for that night and cannot be demonstrated on any other.

There are three reliable places to look for moments of truth, and all three are records the business already has. The first is the interviews described in Chapter 2. When 18 of 20 people interviewed named urgent response within ninety minutes among their top three elements, the interview evidence had already located the first moment of truth without anybody calling it that. The second is the complaint log and the disputed invoice. In the last twelve months, nine invoices for beyond-scope work were disputed, which is a small number against 253 calls and a large number against a promise, and the disputes point at a moment in which the owner learns something about money that they should have learned earlier. The third is the renewal, where 22 renewals were decided and one building was lost: a renewal meeting is a moment of truth because the customer is deciding whether the previous year was what they were told it would be.

Nordec's three moments of truth are the first preventive visit, the first urgent call and the renewal meeting. They have three things in common that explain their weight. Each of them happens once, so there is no second chance to make a first impression of it. Each of them is the moment at which a promise made in the proposal becomes either true or false. And each of them is the moment at which the customer decides whether they will bother to read the report in future months, because attention follows confidence. The first visit is where the customer learns what a preventive visit actually consists of, and it is also where the technician either creates the building record or leaves it for the office. The first urgent call is where the ninety-minute promise is tested in public. The renewal meeting is where the year is judged, and where the capital planning note either becomes the reason to plan a replacement or a document nobody discussed.

Designing a moment of truth has a specific shape that differs from designing an ordinary stage, and it comes down to designing the first time something happens. A moment of truth is rehearsed, scripted and checked, because its cost of failure is out of proportion to its frequency. The first preventive visit is planned with a fixed duration and a checklist, and the record is created on the spot. The first urgent call is followed by a call from the dispatcher, not to apologize but to ask whether the building is working, because the customer's confidence is at its lowest in the hours after the technician has left. The renewal meeting opens with the capital planning note and closes with the note, because the meeting's purpose is the year ahead and the price is a consequence of it.

KEY CONCEPT The customer experiences the sequence, not the average

A business measures itself with averages: 96% of urgent calls inside ninety minutes, 64.4 minutes of average response, 78.7% resolved at the first visit. A customer experiences none of those numbers. A customer experiences a sequence, one event after another, and the sequence is judged by the moments that stand out and by what happened last, not by the mean of the year. This is why a service can hold a good average and still lose a customer, and why a single bad night can outweigh eleven good months. The practical consequence is that averages are for managing the business and sequences are for designing the offer. The measures of section 3.6 tell the owner whether the operation is healthy. The journey and the blueprint tell the owner what the customer will remember, which is a different and more decisive question. Design the sequence and measure the average, never the other way round.

WHO DOES THIS The dispatcher, the lead technician and the owner

In a business the size of Nordec, the journey work divides along the lines of who touches each stage. The dispatcher owns the moments in which the customer is waiting: the telephone call, the decision about urgency, the call that says what will happen with a non-urgent request after hours, and the call back after the first urgent visit. The lead technician owns the first preventive visit and the building record, because the record is a technician's document and cannot be dictated to the office from memory at the end of a day. The owner, Sophie, owns the stages where the relationship is decided: discovery, evaluation, the renewal meeting, and the onboarding call made in the first week after signature. The administrator owns the monthly report, the invoice and the renewal notice, and therefore owns most of the touchpoints that arrive by email. Chapter 9, on typical tasks and workflows, turns this distribution into a cadence; Chapter 10, on expertise and roles, asks when a business of another size should get help.

3.4 The Service Blueprint

A journey map shows what the customer goes through. A blueprint shows what has to happen inside the business for each of those stages to be deliverable, and it is the instrument that turns a journey into an operation. The components are well established. The practical technique behind the drawing is customer-focused: it shows the service process, the points of contact and the physical evidence from the customer's point of view, and it connects the support processes behind them (Bitner et al., 2008). Its lanes, from the top, are the physical evidence the customer encounters, the customer's actions, the onstage contact employee actions that the customer can see, the backstage contact employee actions that the customer cannot see, and the support processes that make both possible. Three horizontal lines separate the lanes.

The line of interaction sits between the customer's actions and the onstage employee actions, and each time it is crossed a moment of truth occurs, which is why section 3.3 and this section are the same subject seen from two sides (Bitner et al., 2008). The line of visibility sits between the onstage and the backstage actions: everything above it is seen by the customer, and everything below it is not. The line of internal interaction sits between the backstage actions and the support processes, and it separates the people who serve the customer from the systems, suppliers and back-office functions that supply them. The technique has an order that matters as much as its parts. The service and the customer segment are chosen first, so that one blueprint covers one moment for one kind of customer, and the customer's actions are drawn first, so that the drawing begins on the customer's side.

This book uses one term of its own for the steps where the promise can break: fail points. A fail point is a step in the blueprint at which the customer's expectation and the business's ability to meet it can come apart — a part that is not on the van, an approval that was never asked for, a call that was not returned. Marking them is the reason to draw the blueprint at all, because a fail point that has been marked can be designed out, given a recovery, or at least knowingly accepted. Waits deserve the same treatment, and the practice this book recommends is to look at each step and ask two questions: can this step fail, and if it fails, what does the customer experience next?

Figure 3.1 shows the blueprint of a Nordec urgent call. The figure is drawn as five horizontal lanes running left to right, from the physical evidence at the top to the support processes at the bottom, crossed by three horizontal lines, the line of interaction between the customer and the onstage employees, the line of visibility between the onstage and the backstage work, and the line of internal interaction between the backstage work and the support processes. Along the lanes runs the sequence of the call, from the tenant's report to the property manager, through the dispatcher's questions, the technician's arrival and the work on the equipment, to the closing note written into the building record on site. Three fail points are marked on the drawing: the missing part, the two-hour limit and the after-hours call.

Figure 3.1 — The blueprint of an urgent contract call. The line of visibility separates what the property manager sees from what the dispatch board, the van stock and the building record do; the three fail points are where the promise used to break.

The blueprint earns its place for a reason that is easy to state and hard to accept: most of what the customer experiences is decided below the line of visibility. The ninety-minute arrival is not produced at the doorstep. It is produced by a dispatch board that shows which technician is nearest, by a service van that holds the right part, by a building record that says what equipment is installed, and by a maintenance schedule that keeps a technician from being two hours away when the call comes in. A business that designs only what the customer sees is designing the last five minutes of a two-hour process.

METHOD A service blueprint in one afternoon

First, choose one service moment and one kind of customer, because a blueprint that covers everything covers nothing: the Nordec team drew one urgent call for one property manager. Second, write the customer's actions from left to right across the top of the working area, in the customer's own words, beginning before the call and ending after the technician has left. Third, add the actions of the employees the customer meets above the line of visibility, and the actions of the employees the customer never meets below it: the dispatcher's search of the dispatch board, the technician's check of the van stock. Fourth, add the support processes that make those actions possible, below the line of internal interaction, and the physical evidence the customer encounters in the top lane: the telephone, the van, the equipment, the note, the invoice. Fifth, walk the drawing from left to right and mark every fail point and every wait, and give each one either a design change that removes it or a recovery that limits it. Deliverable: one page, five lanes, three lines, a numbered fail point list with an owner for each.

TOOLS OF THE TRADE What a journey map and a blueprint need

Neither drawing requires software. The first requirement is a surface large enough to hold five lanes across a wall, which may be a wall itself, a sheet of paper taped to one, or a whiteboard application shared on a screen when the people drawing are not in the same room. The second is a supply of sticky notes or cards in several colors, because the value of the exercise comes from moving a step rather than from writing it, and because a card placed in the wrong lane can be picked up and moved by the person who disagrees. The third is a template with the five lanes and the three lines already drawn, so that the afternoon goes into the content instead of the layout. The fourth is the complaint log and the disputed invoices, which are the evidence for the fail points. The fifth is a camera, or a scan of the finished wall, so that the drawing goes into the file that Chapter 9 reviews each quarter. No brand name belongs on this list; the requirement is the function. Book K1 (The SME Technology Stack) covers the inventory of the tools themselves.

3.5 Fail Points, Waits and Endings

A fail point is either designed out or given a recovery, and the difference between the two is worth stating plainly. Designing out means changing the process so that the failure cannot happen: putting the part on the van, writing the closing note on site, saying out loud at the moment of the call what will happen to a non-urgent request. Giving a recovery means accepting that the failure will sometimes happen and deciding in advance what the customer experiences next: the call from the dispatcher after the first urgent visit, the credit that appears on the next invoice without a claim, the owner who is told before the two hours are passed rather than after. A service business with no recoveries is a business that leaves every failure to be improvised by whoever is closest to it.

The psychology behind recoveries and endings is where behavioral science adds something that queue management cannot. The framing is direct: service management has long studied queues and call centers, and the useful addition of behavioral science is what it says about the psychology of the service encounter itself (Chase & Dasu, 2001). What people perceive in waits, sequences and endings can then be managed by design rather than left to arithmetic, as the two studies below suggest. The evidence on endings is the clearest part of it. Pain recorded in real time during two minimally invasive procedures, 154 patients undergoing colonoscopy and 133 undergoing lithotripsy, was strongly related to its peak intensity and to the intensity in the last three minutes when patients were asked what they remembered, and the length of the procedure played little part: long procedures were not remembered as especially bad (Redelmeier & Kahneman, 1996). The same asymmetry appeared in an experiment in which participants held a hand in cold water for 60 seconds in one trial and for 90 seconds in another, the last 30 seconds being slightly less cold, and a majority chose to repeat the longer trial, because remembered evaluations are dominated by the worst and the final moments and duration plays a small role (Kahneman et al., 1993).

Applied to a maintenance contract, the finding changes three designs at once. The first is the urgent call, where the worst moment — the technician leaving without the problem solved, or the discovery that a part must be ordered — is followed by the dispatcher's call, so that the last moment the customer experiences is a person asking whether the building is working rather than a van driving away. The second is the invoice, where the two-hour call moves the unpleasant information out of the ending and into the middle of the visit, when the customer can still decide what to do about it. The third is the renewal meeting, where the capital planning note is placed at the end, because a meeting that closes on a price increase closes on the worst moment, and a meeting that closes on a plan for the next three years closes on something the owner can use.

Waits deserve the same treatment as failures, because a wait that nobody has designed is experienced as a failure of the promise. On an urgent call, the customer waits between the call and the arrival, and the only useful intervention is information: a technician who is coming from three buildings away, and a dispatcher who says so. The wait for a part is a different wait, and it is the one that turns one visit into two. The wait for the invoice is a third, and it is the reason nine invoices for beyond-scope work were disputed in a year, because an invoice for work the owner did not authorize in advance is a surprise arriving weeks later, alone, with no person attached to it.

THE RULE Draw the line of visibility before you design the front

The blueprint, as the practical technique sets it out, shows the service process, the points of contact and the physical evidence from the customer's point of view, connected to the support processes behind them, and it places the line of visibility between the actions the customer can see and those they cannot (Bitner et al., 2008). The rule this book takes from that structure is a design rule, not a drawing rule: every promise made on the customer's side of the line must have a backstage process and a support process that make it deliverable, and the promise and the process are designed together or not at all. A ninety-minute arrival promise without a dispatch board, a van stock and a maintenance schedule is not an offer; it is a hope with a number attached. The same rule read backwards is a diagnostic tool: when a promise keeps failing, the cause is almost always found below the line of visibility, in the part of the process the customer never sees and the business rarely draws.

DO THE MATH What a van stock saves

Illustrative figures; the case is fictional. Step 1 — The last twelve months brought 253 urgent contract calls and 54 second visits; 37 of the second visits happened because a part was not on the van, and each second visit costs one and a half hours of a technician at $54, or $81, which is 37 × $81 = $2,997 a year. Step 2 — The twenty parts most often missing, drawn from the building records and carried on the two vehicles, are expected to avoid 26 of those 37 visits. Step 3 — 26 × $81 = $2,106 a year of technician time that is not spent returning to a building. Step 4 — The stock costs $3,200 across the two vehicles. Step 5 — The payback is $3,200 ÷ $2,106 = 1.5 years, and the customer is spared a second disruption, which the calculation does not price. A customer who has to let a technician into the building twice for one fault experiences the contract as one that did not finish the job, and no arithmetic in this chapter captures that cost.

3.6 Service Quality and Its Dimensions

Quality in a service is not what the business produces; it is what the customer perceives, and the difference between the two is the whole subject of service quality measurement. A technician can perform a perfect preventive visit on a building whose owner never learns what was done, and the customer's perception of quality will be low. The measurement instruments built for this purpose have a long history. A multiple-item instrument for assessing customers' perceptions of service quality in service and retailing organizations was developed with evidence of its reliability, factor structure and validity drawn from four independent samples (Parasuraman et al., 1988). The instrument described in that work is a scale of twenty-two items, and its lesson for a small business is not the scale but the direction of the questions: the customer is asked what they perceived, not what the business believes it delivered.

A small business does not need twenty-two items and does not need a survey platform. It needs a handful of measures, each of which follows something the customer actually judges, and each of which has a target that was set before the year began rather than discovered in it. The test of a good measure is a single question: if this number moves, does the customer notice? A measure that reports the number of preventive visits performed passes the test only if the visits are what the customer experiences; a measure that reports the number of building records delivered within thirty days of signing passes it more directly, because the record is a document the customer receives. The measures should also be few enough that the person who owns them can remember them without reading them, which in a business of seven people means five measures and no more.

Nordec keeps five, and they were written at the first annual refresh of Book A3's strategy, before the year they describe. Two of them follow the urgent call, one follows the invoice, and two follow the monthly routine and the onboarding. Four are already above or near target, and one is not: first-visit resolution sits at 78.7% against a target of 88%, which is the measure the blueprint work of section 3.8 exists to move. Keeping the target at 88% while the figure sits at 78.7% is deliberate, because a target that is quietly lowered to meet the current result stops being a target and becomes a description.

The following table sets out the five measures the case keeps, with what the customer judges, the measure, the target and the last twelve months. Read the first column as the reason the measure exists: each number in the table answers a question a customer is actually asking, and the last column is what the business says when the customer asks it.

Table 3.2 — The quality measures Nordec keeps

Two habits keep a small set of measures honest. The first is that each measure has a named owner, so that the person who reports it is the person who can change it: the dispatcher owns the arrival measure, the lead technician owns first-visit resolution, the administrator owns the two measures about documents arriving on time, and Sophie owns the disputed invoices. The second is that the measures are read at a fixed moment and against the journey rather than in isolation, because a measure read alone tells a story that four measures read together contradict. Chapter 9 places that reading in the quarterly review, and Book J4 (Quality Management and Continuous Improvement) treats the improvement method that follows it.

QUÉBEC & CANADA NOTE What a contractor owes the client

A maintenance contract between two businesses in Québec is governed by the Civil Code of Québec, and two of its articles shape what a blueprint is for. Article 2100 provides that a contractor or a provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice and, where applicable, ensure that the work or service conforms to the contract; where the provider is bound to an obligation of result, they can be relieved of liability only by proving superior force. Article 2102 requires the provider to give the client, before the contract, the useful information on the nature of the task and the time and property it requires. A blueprint is one way a business demonstrates that this duty is met, because it shows what was considered, what was promised and what the process behind the promise is. Clients may also resiliate a service contract unilaterally under articles 2125 and 2129, paying for the work performed and the costs incurred. This book explains the law and does not replace the advice of a lawyer.

3.7 Designing the Onboarding

The first thirty days are where the gap between what was sold and what was delivered is either closed or opened, and in a recurring service they are also the only stage where the customer's attention is at its peak. A building owner who has just signed reads everything: the first visit sheet, the first report, the first invoice. An owner in the fourteenth month of a contract reads very little. Onboarding is therefore the moment of truth that can be designed at the least cost, because it affects one customer at a time and the work is short.

Nordec's onboarding had been implicit until the blueprint work made it visible. The record was supposed to be created within thirty days of signature, and it was: the measure reads 100% over the last twelve months. What the journey map showed was everything around it. The first visit happened when the schedule reached the building, which could be six weeks after signature. The record was created, delivered to the file and not to the owner. Nobody called the customer in the first week. Nothing at all happened after the first urgent call, which meant that the moment at which confidence is lowest was the moment at which the business was silent. The onboarding had one task and four gaps.

Four changes were made, and each of them attaches to a step of the blueprint. The first is a welcome call from Sophie in the first week after signature, whose purpose is to confirm the scope, the exclusions and the two-hour rule before the first invoice can surprise anybody. The second is that the first preventive visit is booked at signature rather than when the schedule allows, which turns a six-week silence into a date the owner can write down. The third is that the building record is delivered to the owner at the end of the first month with a one-page summary, so that a document that used to live in the company's file becomes something the customer has read. The fourth is a call from the dispatcher after the first urgent call, which is the recovery described in section 3.5 placed where it does the most good.

The onboarding is written as a checklist with an owner and a date for each line, because a design that lives in a conversation is a design that disappears in a busy month. The record is created within thirty days of signature and takes six hours of a technician's time, which at a loaded cost of $54 an hour is $324 of work delivered before the first payment settles — a figure that matters in Chapter 5, where the first-ninety-days guarantee is designed, because that onboarding is what the guarantee puts at risk. The checklist also gives the sales side of the business something it can promise in the proposal, and a promise that a customer can date is easier to believe than a promise about the quality of a relationship. Book E7 (Account Management, Customer Success and Expansion) takes the same first thirty days from the account side, and Book J6 (Customer Service and Support Operations) takes the ongoing contact.

IN LA MACHINE What the platform can and cannot hold of a service design

The redesign work of this chapter can live in the application copy read for this book, but only in part (La Machine, 2026). A project at productivity/projects can carry the blueprint work as a piece of work with a start and an end, and the tasks it produces — put twenty parts on the two vans, book the first visit at signature, write the closing note on site — can be held at productivity/tasks, with the meetings that review them at productivity/meetings and their dates in the calendar at productivity/calendar. In the Control center, the page at /taches gives each task a verification criterion, which is where a design change stops being an intention and becomes something that can be checked. What the platform does not have belongs in the same breath: there is no service blueprint and no customer journey map, no service-level tracking, no guarantee or warranty register and no survey or net promoter tool, only the general form builder. The drawing stays on a wall and in a file; the platform holds the tasks that follow from it. Book N2 (Piloting Tasks, Executions and Projects) describes the tasks pages.

3.8 The Case's Blueprint of an Urgent Call

The blueprint was drawn on a Thursday afternoon in the depot, by the three people who know the urgent call best: the dispatcher, the lead technician and Sophie. They took a wall, a roll of paper, a marker and the year's call records, and they spent three hours on it. The dispatcher wrote the customer's lane first, from the tenant's report to the property manager through to the invoice, because the customer's actions are drawn first and because the dispatcher is the person who hears them. The lead technician filled the backstage lane, which is where the afternoon became interesting: the search of the dispatch board, the call to the supplier, the decision about which part to carry, the entry written into the building record. Sophie filled the support lane, where the maintenance schedule, the van stock, the parts account and the software that holds the records sit.

The numbers they wrote underneath the drawing came from the last twelve months, and they had never been laid out in one place. There were 253 urgent contract calls. A technician was on site within ninety minutes for 243 of them, which is 96%, against a target of 95%. Of the 253, 199 were resolved at the first visit, which is 78.7%, against a target of 88%. That left 54 second visits. Of those, 37 happened because a part was not on the van, and 17 because the work passed the two hours included in the contract and needed the owner's approval. Nine invoices for beyond-scope work were disputed. The dispatcher's contribution to the afternoon was the observation that the two-hour cases and the disputed invoices were the same cases seen from two ends of the week.

Three fail points went onto the drawing, and each of them sits at a different line. The first is the part that is not on the van, which fails below the line of visibility, in the backstage lane, where the customer cannot see it until the technician says the word "order". The second is the work that exceeds the two included hours, which fails at the line of interaction, because the moment the technician decides to keep working is the moment the customer's expectation and the invoice begin to diverge. The third is the non-urgent call received after hours, which fails before the blueprint proper even begins, in the first thirty seconds of the telephone call, when the owner expects a night visit and the contract says the next working morning.

The redesign that followed addressed all three. For the missing part, the twenty parts most often needed are now drawn from the building records and carried on both vehicles, at a cost of $3,200, expected to avoid 26 of the 37 second visits and to save $2,106 a year, with a payback of about a year and a half. For the two-hour limit, the technician makes the "two-hour call": before passing the included hours, the technician telephones the owner with an estimate, so that the extra work is authorized in the middle of the visit instead of appearing on an invoice weeks later. For the after-hours call, the dispatcher says what will happen while the call is being taken, in one sentence, so that a non-urgent request received at nine in the evening is understood to be booked for the next working morning rather than refused.

What the blueprint showed that the averages had hidden is the most useful result of the afternoon. The arrival figure of 96% looked like a service that worked, and the second-visit figure of 54 calls is what the same service looked like from the customer's side. One call in five came back. A customer who lets a technician into the building twice for one fault has experienced a contract that did not finish the job, and no measure in the business had been watching that. The target of 88% first-visit resolution and the target of no disputed invoice were set on the same afternoon, which is why they appear in Table 3.2 beside the arrival figure they qualify.

The following table sets out the three fail points, what used to happen at each of them, the fix, the owner and what the fix costs. Read it as the deliverable of the afternoon: a drawing becomes an operation when each fail point acquires a person and a price.

Table 3.3 — The three fail points of an urgent call and their fixes

CASE IN POINT Nordec Entretien inc.: one call in five came back

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The figure that changed the afternoon was not the 96%. It was 54. Twenty-two per cent of the urgent calls in the year had needed a second visit, and the business had never counted that number because it was not the number the promise was about. The promise was ninety minutes and the operation was measured on ninety minutes, so the operation was managed on ninety minutes. Once the second visits were on the wall beside the arrival times, the causes stopped being a matter of opinion: 37 of the 54 were a part, 17 were an approval, and the 9 disputed invoices were the 17 seen from the customer's side a month later. The vans were loaded, the two-hour call was written into the checklist, and the dispatcher was given the sentence about after-hours requests. None of those three changes required a purchase beyond the $3,200 of parts, and each of them was attached to a line on the drawing before the afternoon ended.

Costly Mistakes

• Letting the service design itself by default. When nobody has decided what counts as urgent or when an owner hears about extra hours, the dispatcher and the technician decide it every night, and they decide it differently depending on how tired they are.
• Drawing the journey from the organization chart. A map with a lane for dispatch, a lane for field work and a lane for billing describes the business's departments, and it tells the owner nothing about what a customer actually goes through.
• Treating every touchpoint as equally important. A journey of thirty touchpoints has three or four moments of truth in it, and spending the design effort evenly is how a business perfects the invoice and loses the first urgent call.
• Designing only what the customer can see. The ninety-minute arrival is produced below the line of visibility, by the dispatch board, the van stock and the maintenance schedule, and a promise designed without those is a promise with no process under it.
• Leaving fail points unmarked because nothing has gone wrong yet. A fail point that has not been marked cannot be designed out or given a recovery, and it will be improvised by whoever is closest to it on the night it happens.
• Ending an experience on its worst moment. A technician who leaves without solving the problem, or a renewal meeting that closes on a price, fixes the customer's memory on exactly the wrong point of the year.
• Measuring quality with figures the customer never experiences. A count of visits performed, reports written and parts ordered says what the business produced, and the customer's judgment is about arriving on time, solving the problem and receiving the document that was promised.

Key Takeaways

• A service is a process the customer takes part in, and if it is not designed deliberately it is designed by whoever answers the telephone, which is why the sequence deserves the same care as the contract that describes it.
• The customer journey runs before, during and after the purchase, across many touchpoints in several channels, and delivering it requires several functions of the firm, and sometimes external partners, to work together (Lemon & Verhoef, 2016).
• A service blueprint draws one moment for one customer in five lanes — physical evidence, customer actions, onstage actions, backstage actions and support processes — separated by the line of interaction, the line of visibility and the line of internal interaction (Bitner et al., 2008).
• Every promise made in front of the customer must be matched by a backstage process and a support process that make it deliverable, and that rule read backwards is a diagnostic tool for promises that keep failing.
• People remember an experience by its worst moment and its final moment rather than by its length, which is why a call back after the first urgent visit, a two-hour call before extra work, and a renewal meeting that ends on the capital planning note are design decisions and not courtesies (Redelmeier & Kahneman, 1996; Kahneman et al., 1993).
• Service quality is the customer's perception, which multiple-item instruments such as SERVQUAL were built to measure (Parasuraman et al., 1988); a small business follows it with a small set of measures that match what the customer judges, each with a named owner and a target set before the year began.

Your Action Plan

1. Draw the journey in seven stages and name an owner for each. Take the stages from discovery to renewal, write what the customer does at each one, and put a name in the ownership column before the drawing is finished. Deliverable: one journey map with a named owner for every stage. Verification: no owner appears twice on the same stage, and every owner can say what happens at their stage without reading the map.
2. Locate the moments of truth from evidence rather than opinion. Read the last year's interviews, the complaint log, the disputed invoices and the renewals, and choose the three or four moments that decide the customer's judgment. Deliverable: a short list of moments of truth with the evidence that put each on it. Verification: every moment on the list can be traced to a complaint, a renewal, a lost customer or a named interview answer.
3. Draw the blueprint of one service moment in an afternoon. Choose one moment and one kind of customer, write the customer's actions first, then the onstage and backstage actions, then the support processes and the physical evidence, and mark each fail point and wait as it appears. Deliverable: one blueprint page with five lanes, three lines and a numbered fail point list. Verification: every fail point on the list has a named owner and a date by which a fix or a recovery will be decided.
4. Design the onboarding as a dated checklist. Write the first thirty days line by line, from the welcome call and the booking of the first visit to the delivery of the record and the call that follows the first urgent call. Deliverable: one onboarding checklist with an owner and a date on every line. Verification: the checklist covers the first thirty days with no gap longer than a fortnight and no line without an owner.
5. Write five quality measures with targets and owners, and read them quarterly. Choose measures that follow the customer's judgment, set each target before the year begins, and give each one to the person who can change it. Deliverable: a five-line measure sheet with a target and an owner per line. Verification: every measure answers a question a customer is actually asking, and the quarterly reading is on the calendar with the other reviews of Chapter 9.

In La Machine

The work this chapter describes can be carried in the application copy read for this book, but only in part (La Machine, 2026). A project at productivity/projects holds the redesign of a service moment as a piece of work with a beginning and an end, and the tasks it produces — load the twenty parts on both vehicles, book the first visit at signature, write the closing note into the building record on site — live at productivity/tasks, with the reviews of that work at productivity/meetings and their dates in the calendar at productivity/calendar. What makes the arrangement useful rather than decorative is the Control center page at /taches, where each task carries a verification criterion: a design change entered there can be checked rather than remembered, which is the difference between a blueprint that changed the business and a drawing that was admired on a Thursday.

What the platform does not hold belongs in the same paragraph. There is no service blueprint and no customer journey map, so the five lanes, the three lines and the fail points stay on a wall and in a file; there is no service-level tracking and no service credit mechanism, so the arrival measure and the disputed invoices are counted by hand; there is no guarantee or warranty register, which matters for Chapter 5; and there is no survey or net promoter tool, only the general form builder at business/marketing/forms, whose answers are kept with the form and which can carry a customer interview or a short questionnaire. La Machine has no quote or proposal module either — the page at Business → Motivational Quotes holds motivational quotations — so the proposal that opens the journey is produced outside it, as Book E5 (Proposals, Quotes, Business Cases and Contracts) describes. Book N1 (Getting Started with La Machine) maps the families the platform does cover, and Book N2 (Piloting Tasks, Executions and Projects) describes the tasks pages used here.

Self-Check

1. Why does a service have to be designed deliberately, and what decides its design when nobody designs it?
2. What are the stages of the customer journey, and what three things must be attached to each stage for a journey map to be useful?
3. What is a moment of truth, where are the three reliable places to find one, and which three does the case have?
4. What are the five lanes and the three lines of a service blueprint, and what happens each time the line of interaction is crossed?
5. What did the blueprint of an urgent call show that the averages had hidden, and what three fail points did the redesign address?

Answers. 1. A service is produced in the customer's presence as a sequence of moments, so its quality cannot be built into an object before the sale; where nobody designs it, it is designed by whoever takes the call, which is why the same promise is kept differently on a busy night and a quiet one. 2. The journey runs through discovery, evaluation, signature, onboarding, the monthly routine, the urgent call and the renewal; each stage must carry the touchpoint through which the customer experiences it, the thing that can go wrong there and the person inside the business who owns it. 3. A moment of truth is a touchpoint that carries disproportionate weight, so that the customer's overall judgment of the offer forms there; the three reliable places to find one are the customer interviews, the complaint log and disputed invoices, and the renewal decisions; the case's three are the first preventive visit, the first urgent call and the renewal meeting. 4. The lanes are physical evidence, customer actions, onstage contact employee actions, backstage contact employee actions and support processes, separated by the line of interaction, the line of visibility and the line of internal interaction; each time the line of interaction is crossed, a moment of truth occurs (Bitner et al., 2008). 5. The blueprint showed that the arrival figure of 96% within ninety minutes sat beside 54 second visits, one urgent call in five coming back, with 37 caused by a missing part, 17 by work exceeding the two included hours and nine disputed invoices; the redesign put twenty parts on the two vehicles, added the two-hour call before extra work, and gave the dispatcher the sentence that explains what happens to a non-urgent request received after hours.

Chapter References

Bitner, M. J., Ostrom, A. L., & Morgan, F. N. (2008). Service blueprinting: A practical technique for service innovation. California Management Review, 50(3), 66–94. https://doi.org/10.2307/41166446

Chase, R. B., & Dasu, S. (2001, June). Want to perfect your company's service? Use behavioral science. Harvard Business Review. https://hbr.org/2001/06/want-to-perfect-your-companys-service-use-behavioral-science

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

Kahneman, D., Fredrickson, B. L., Schreiber, C. A., & Redelmeier, D. A. (1993). When more pain is preferred to less: Adding a better end. Psychological Science, 4(6), 401–405. https://doi.org/10.1111/j.1467-9280.1993.tb00589.x

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/projects, productivity/tasks and /taches.

Lemon, K. N., & Verhoef, P. C. (2016). Understanding customer experience throughout the customer journey. Journal of Marketing, 80(6), 69–96. https://doi.org/10.1509/jm.15.0420

Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1988). SERVQUAL: A multiple-item scale for measuring consumer perceptions of service quality. Journal of Retailing, 64(1), 12–40.

Redelmeier, D. A., & Kahneman, D. (1996). Patients' memories of painful medical treatments: Real-time and retrospective evaluations of two minimally invasive procedures. Pain, 66(1), 3–8. https://doi.org/10.1016/0304-3959(96)02994-6

Shostack, G. L. (1984, January). Designing services that deliver. Harvard Business Review. https://hbr.org/1984/01/designing-services-that-deliver

Chapter 4 — Packaging: Tiers, Bundles and Add-ons

Packaging decides what a customer can choose, and a good package is designed from the segments and the value work of Chapter 2 — what each tier contains, what it excludes and whom it is for — before anyone argues about its price, which Book A7 sets.

What You'll Be Able to Do

• Explain why the same list of services has to be arranged into different packages for different buyers, and what goes wrong when a business sells one package to everyone.
• Design a three-tier structure in which each tier is complete for the segment it serves and each step up adds something that segment named in its own words.
• Distinguish price bundling from product bundling, and pure bundling from mixed bundling, and say which of the three an offer is actually using.
• Place an add-on so that it widens a lower tier without emptying the tier above it, and compute what each tier and each add-on leaves behind.
• Read the research on how choices behave — the attraction effect, the compromise effect and the cost of too many options — and treat it as a description of buying rather than as a technique for steering a buyer.
• Migrate existing customers to a new structure without taking away what they already receive, using a fence that opens a new segment instead of cannibalizing the flagship.

Why It Matters

Before packaging, an offer is a list. Nordec Entretien inc. reached the first annual refresh of its strategy with a contract that had grown element by element over two years — the monthly preventive visit, urgent calls answered on site within ninety minutes, the building health record, the one-page monthly report, the capital planning note at renewal — and with a single invitation to the customer: take this, at $712 a month, or take nothing. Twenty interviews conducted in the weeks that followed showed that this one package was doing two opposite things badly at the same time. It was too much for the small building whose owner said the contract was more than the building needed, and too little for the property manager who wanted a single report across a dozen buildings and had no way to buy one. Both of those are packaging failures rather than pricing failures, and neither one can be repaired by moving the price up or down.

Packaging is the step where the value work of Chapter 2 becomes something a buyer can actually choose. The interviews produced counts rather than opinions, and a count is the raw material of a structure. Urgent response within ninety minutes was named among each person's top three elements by 18 of the 20 people interviewed. The monthly preventive visit was named by 14. The capital planning note was named by 12, including 7 of the 8 owner-operator customers and 3 of the 6 owners who had declined a proposal in the previous two years. The monthly report was named by 6, and the building health record by 4. Read one at a time, those numbers are interesting. Read together, they describe groups of people who want different packages: the small building owner wants the response and the visit; the owner-operator wants the response, the visit and the note; the property manager wants the response, the visit, the report, and a way to see several buildings at once. A package is how a business answers each of them without pretending they are the same buyer.

The cost of leaving this undone is measurable in the case, and it shows up in three places. Two of the four proposals that were lost on price in the strategy year went to the independent trades billed by the call, and both were small buildings under 3,500 square metres whose owners said the contract was more than the building needed. At the same time, customers asked for more with nothing to buy: five asked for a quarterly review meeting, four asked for help with replacement projects, and three property managers asked for a single report across their buildings, which is twelve requests that had no product behind them. And the offer carried $20,700 a year of additions — the record, the report, the notes — that nobody had ever priced. Book A7 puts a price on what this chapter designs, and that order is deliberate: a package whose contents are undecided cannot be priced, and a package whose contents are sold at the wrong level cannot be rescued by a discount.

In This Chapter

Section 4.1 explains why an offer has to be packaged at all, what a package decides, and who does the work. Section 4.2 sets out the good-better-best structure, the role of the middle tier as the flagship, and what the research on choice says about three options rather than two. Section 4.3 gives the design rule for the contents of each tier and the table of Nordec's three tiers. Section 4.4 takes up bundling in its three forms — pure, mixed and unbundled — and the difference between bundling prices and bundling products. Section 4.5 covers add-ons and options, and when an add-on protects a tier and when it empties the one above. Section 4.6 addresses choice architecture and the cost of too many options. Section 4.7 covers fences, segments and the migration of existing customers. Section 4.8 follows the working session in which Nordec's three tiers were built, named and costed.

4.1 Why Package an Offer

An unpackaged offer puts the buyer to work. The business publishes what it can do — a monthly preventive visit, an urgent response, a record of the building, a report, a note about equipment — and leaves the customer to assemble a selection and to guess what the total will be. Some customers enjoy that work and most do not, and the ones who do not either buy the whole list at the price asked or walk away with a reason they rarely say out loud. Insurance is the everyday example: the underlying cover could be sold as a set of priced pieces, and it is sold instead as named packages, because the buyer needs to compare rather than to build.

Packaging is the arrangement of an offer's elements into named choices that a buyer can compare. It settles four questions and leaves a fifth to Book A7. The first question is who each choice is for, stated plainly enough that a buyer recognizes themselves in one of them. The second is what each choice contains, in the customer's own vocabulary rather than the technician's. The third is what each choice excludes, written down rather than left to the customer's assumption. The fourth is what each choice is called, which sounds trivial and is not, because a name the customer would use is a name they can repeat to a colleague. The fifth question — what each choice costs the buyer — belongs to Book A7 (Pricing Strategy and Revenue Models), and the sequence matters: contents first, exclusions second, names third, price last.

The reason a single package fails is that the segments of Chapter 2 do not want the same contents, and the difference is not a matter of degree. Take Nordec's two clearest segments. The owner-operator of a single building of about 4,500 square metres wants the building to keep running, the tenants to stay quiet, and a year's notice of what will have to be replaced; the value equation against a patchwork of trades gives that building a gain of $1,516 a year under the contract, before the note, the record and the report, which the equation does not price. The owner of a building of about 2,500 square metres wants the same things at a smaller scale, and the same contract costs that building $7,080 a year against $6,620 for the alternative, which is $460 more — precisely what the two lost owners said when they chose the trades billed by the call. One segment is being served well; the other is being asked to pay for elements it does not value. A package is the instrument that separates them.

The opposite failure is just as common and less visible, because it looks like success. The property manager of fourteen buildings bought the flagship contract, was satisfied with it, and asked for things the business did not sell: a single report across the buildings, a quarterly review, a line for tenants' non-urgent requests. Those requests are the sound a customer makes when the offer stops one step short of what they need, and the business that does not package them into a choice hears the same requests for a year and answers each one by hand. The three property management firms held 14 of the 46 buildings at the refresh, which is 30.4% of the base, and they were the segment asking for the most that could not be bought.

Packaging has a third effect that is easier to miss, and it concerns the business rather than the buyer. A package is a promise about how work will be done, and the shape of the promise determines the shape of the operation behind it. The decision to sell a monthly preventive visit in every tier decides how the technicians' months are scheduled. The decision to include urgent calls up to two hours of work in every tier decides what a first visit must be able to finish. The decision to leave the monthly report out of the lowest tier decides how much of the administrator's time that tier consumes. An offer designed on a whiteboard and handed to operations unchanged will be quietly redesigned by operations, which is why the design of the package and the costing of it belong in the same conversation, and why the costing method itself is treated in Book A7.

Finally, packaging is where a business discovers how much of its own work it has been giving away. Nordec's contract had accreted: the monthly visit cost $245.60 a month per building in a territory, urgent calls $37.13, the building health record $13.50, the one-page monthly report $27.00, the capital planning note $13.50, and a mid-contract note $4.50 that had been decided but never announced. The first two elements came to $282.73 a month, and all six came to $341.23. The price had moved from $691 to $712 in the same period, which brought $11,592 a year across 46 buildings against $32,292 a year of added cost, leaving $20,700 a year of work that was never priced. Writing the elements into tiers is how that arithmetic becomes visible, because a tier is a list of what the customer receives, and a list can be costed.

DEFINITION Tier, bundle, add-on, option and fence

A tier is one of several named versions of the same offer, each complete in itself, arranged in steps so that a buyer can compare them. A bundle is two or more separate products sold in one package; the trade also calls a tier a bundle when it gathers several elements into one price. An add-on is an element of a higher tier sold on its own to a customer of a lower tier, so that the customer can take one step in contents without taking the whole step in price. An option is any element the buyer may include or leave out at the moment of choosing, which makes an add-on the most common kind of option, and which also makes an awkward truth: an option the customer does not understand is not a choice but a question. A fence is the condition that decides who may buy a lower tier — a size, a location, a use, a commitment — and it exists so that the lower tier serves a segment that the higher tier cannot serve, rather than giving the higher tier's customers a cheaper way to buy the same thing.

WHO DOES THIS The owner decides, the team fills the columns

In a business of seven people, the owner decides what is sold and the team supplies the contents. At Nordec the sequence was explicit. The owner ran the twenty value interviews, chose the segments, and set the number of tiers; she is also the person who signs the contract and can therefore make a promise. The lead technician supplied what each element takes on site — the hours of a monthly visit, what can be finished inside two hours, which parts are habitually missing — because a tier whose contents cannot be delivered on a Tuesday morning is a promise the business has already broken. The administrator supplied the time each element takes off site: the reports, the records, the notes, the scheduling of tenant requests, and the minutes those tasks consume in a month. The costing method and the price came afterwards and came from Book A7. What the owner must not do is fill in the technical columns from memory, and what the team must not do is decide the number of tiers. The division is the reason the three tiers matched both the interviews and the van.

4.2 Good, Better, Best

The three-step structure — good, better, best, or in this book's words a lower tier, a flagship and an upper tier — is the shape most offers settle into once their segments are known, and the case is a clean example of it. The lower tier exists for the segment that needs the essentials and cannot pay for more. The flagship is the version the business is built to deliver, which at Nordec is the current contract cleaned up: the full preventive program, with the record, the report and the note at renewal. The upper tier exists for the segment that manages several buildings and needs an instrument for reporting to others. Three is not a magic number; it is the smallest number that lets a buyer see a choice without leaving them to compare a long menu, and Chapter 2's segments happened to produce the three.

The middle tier is the flagship for a reason that has nothing to do with psychology. It is the version the business most wants to sell, because it is the version the operation is built around: the full preventive visit, the disciplines that keep a building record useful, the annual note that gives an owner a year's notice of a replacement. It is also, at Nordec, the version that nearly every current customer already holds, which makes it the natural destination of a migration and the natural default of a renewal conversation. A flagship that is not the version the operation is built around is a tier the business will deliver badly, and the cost of delivering it badly is paid inside the margin rather than on the invoice.

The research on how choices behave gives the three-step structure two findings worth knowing, and both are findings about buyers rather than instructions for sellers. The first is that a choice depends on the set of alternatives being considered, not only on the option being chosen; the authors describe tradeoff contrast, in which an option looks stronger or weaker depending on the tradeoffs inside the set, and extremeness aversion, in which an option is more attractive when it is intermediate in the set and less attractive when it is extreme; those two mechanisms explain the attraction and compromise effects, shown in studies with consumer products (Simonson & Tversky, 1992). The second finding concerns a particular kind of addition: an asymmetrically dominated alternative is one that is dominated by one option of a set but not by another, and adding it can increase the probability of choosing the option that dominates it, a result that challenged standard choice models and suggested implications for product lines (Huber et al., 1982).

Neither finding is a technique to use on a buyer, and the distinction matters to this book. A tier that exists only to make the tier above it look reasonable is a tier nobody should have to read, and a customer who discovers it has been misled about the offer as a whole. What the two findings actually say to a business designing an offer is narrower and more useful: a choice set is not neutral, the position of an option inside the set changes how it is read, and a three-step structure gives a buyer a middle option to hold on to. The ethical use of that knowledge is to make sure the middle option is genuinely the best value for the segment that will choose it, and that each of the three could be defended on its own. Nordec's defense was written down: the flagship is not the most profitable tier per dollar — its margin of 53.56% at $725 is lower than Essential's 55.18% and lower than Portfolio's 55.09% — and it is the tier the business most wants to sell because it is the tier that keeps a building under control.

Figure 4.1 shows the three tiers as they were finally drawn. The figure places three columns side by side, headed Essential, Preventive and Portfolio. Inside each column are the elements that tier contains, and the elements that a tier adds over the one to its left are highlighted, so that a reader can see the steps rather than the whole list. Under each column is one line saying whom the tier is for: the small building that needs the essentials, the owner-operator who wants the full preventive program, and the manager of several buildings. Under the Essential column, in its own band, is the fence that limits it to buildings under 3,500 square metres.

Figure 4.1 — Nordec's three tiers. Each step up adds what the next segment named in the interviews; Essential is fenced to buildings under 3,500 square metres, and Portfolio adds four things a property manager can name.

A word about the fourth column that is not in the figure. Sophie drafted it. It was called Portfolio Plus, it sat above Portfolio, and it contained one thing: a technician available at night for non-urgent requests. It was removed before it reached a customer, for two reasons that section 4.8 sets out in full, and its removal is the reason this offer has three tiers rather than four. A fourth tier is not automatically a mistake, and a business with four real segments will have four real tiers. What the case shows is that a tier is justified by a segment that names it, not by the space above the flagship looking empty.

WHAT CHANGED What the research on choice found, and when

Three findings used in this chapter carry three dates, and the order of the dates is worth keeping. In 1982, a study of asymmetrically dominated alternatives showed that adding an option which one item of a set dominates can raise the probability of choosing the item that dominates it, a result that challenged standard choice models and pointed toward product line strategies (Huber et al., 1982). In 1992, a study of tradeoff contrast and extremeness aversion showed that choices depend on the set of alternatives under consideration, and described the attraction and compromise effects in studies with consumer products (Simonson & Tversky, 1992). In 2000, three studies found that people were more likely to buy jams or chocolates, or to write an optional essay, when offered 6 options rather than 24 or 30, and were more satisfied with what they chose (Iyengar & Lepper, 2000). Read in order, the three findings move from adding an option, to reading a set of options, to having too many of them. That progression, rather than any single result, is what a business designing three tiers is working with.

CASE IN POINT Nordec Entretien inc.: the fourth tier that was removed

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The draft above Portfolio was called Portfolio Plus and contained non-urgent after-hours dispatch: a technician available at night for requests that were not urgent. Two arguments removed it. The first came from Book A3 (Strategy and Competitive Positioning), where Nordec had deliberately drawn a reduce on after-hours coverage, keeping the night hours for genuine emergencies and booking non-urgent after-hours requests for the next working morning; selling night coverage for non-urgent work would have reopened the very trade the strategy had closed. The second argument came from the interviews, and it was the decisive one: the only people who had asked for after-hours coverage were the two owners lost on that point in the strategy year, and neither of them was a segment; they were two buyers who had chosen a different trade-off. A tier with no segment is a column of costs waiting to be sold to somebody it does not fit, and the offer closed with three tiers.

4.3 What Goes in Each Tier

The rule that decides the contents of a tier is short, and it is the same rule at every step: each tier is complete for the segment it serves. Complete means that a customer in that segment can buy the tier and run the building without coming back for an element that should have been in it. It does not mean that the tier contains everything the business sells, and it certainly does not mean that all three tiers contain the same things at different quantities. The tiers of a service offer differ by contents, not by volume, and a business that sells the same contents at three sizes has built one tier with three prices.

The second half of the design is the step. Each step up adds what the next segment named, and only what the next segment named. Nordec's steps came directly from the interview counts. The flagship adds the one-page monthly report over the essentials, because the report was named by 6 people, of whom 5 were property managers, and it exists to be forwarded; it is the element a manager needs and a small owner does not. The upper tier adds four things named by the property managers — the mid-contract capital planning note for buildings the first note flagged, a quarterly review meeting, a tenant request line for non-urgent requests, and one consolidated monthly report across the manager's buildings — at a cost of $67.50 a month per building. Each of those four can be said out loud in a sentence, which is the test of whether a step is a step or a fog.

The third part of the design is the exclusion, and it is the part businesses omit. Every tier of Nordec's offer excludes parts, work beyond two hours per incident, roof and envelope work, elevators, and anything requiring a licensed trade the company does not hold. Individual tiers exclude other things: Essential carries no monthly report and no capital planning note, and the note is available to it as an add-on; Preventive and Portfolio do not carry the tenant request line; no tier carries after-hours dispatch for non-urgent requests. An exclusion written into the contract and stated in the proposal is what allows the business to say yes to the work it does and no to the work it does not, without a negotiation each time. Book A2 (The Business Plan) treats exclusions as part of the definition of what a business sells, and this chapter treats them as part of what a tier is.

The fourth part of the design is the fence on the lowest tier, and it is designed together with the price, which Book A7 sets in its Chapter 5. Nordec's fence is size: Essential is available only for buildings under 3,500 square metres. The fence has one job, and it is not to protect revenue. Its job is to make the lower tier serve a segment that the flagship cannot serve, rather than to give an existing flagship customer a cheaper way to buy the same thing. A building of about 2,500 square metres was genuinely lost to the alternative because the flagship cost $460 more a year than the patchwork of trades; at the Essential price the same building gains $740 a year. The building owner who was being overcharged for elements they did not want is a different customer from the owner-operator of 4,500 square metres who wants the full program, and the fence keeps those two apart.

The fifth part of the design, and the last one before the price, is the check on cost and margin. Each tier is costed from the work it contains before it is offered, using the same cost basis throughout: a loaded technician at $54 an hour, the hours of a monthly visit and the travel in a territory, the consumables, the technician and administrator time of each report, note, review and request line. A tier whose cost is unknown is a tier whose price is a guess, and a price that has been guessed is a price that will be corrected by a customer noticing that the business is reluctant to sell it. Nordec's three tiers were costed before Sophie took them to Book A7's method, and the costs are set out later in this chapter and used in the illustration in section 4.5.

Nothing in the design takes away what a current customer already receives, and that is a rule rather than an outcome of this case. When the tiers were drawn, the one-page monthly report left Essential — the tier that did not exist before — and was replaced there by a visit sheet sent after each visit. No Preventive customer lost the report, because Preventive is the contract they already hold; the elements that preventive customers receive were not reduced by a single line. The mid-contract capital planning note, which had been decided at the refresh and never announced, moved into Portfolio before any customer had grown used to receiving it in the flagship, which is the only moment at which such a move is free. A design that takes something away from an existing customer at the same time as it introduces a new structure gives that customer a reason to re-examine the whole relationship, and the migration rules in section 4.7 are built to avoid exactly that.

The following table sets out what each of Nordec's three tiers contains, element by element, so that the steps and the exclusions are visible on one page. Read it column by column rather than row by row: each column should be complete for the buyer named at its foot, and each cell that reads no should be a thing that buyer does not need rather than a thing they will discover they want in the fourth month.

Table 4.1 — Nordec's three tiers

KEY CONCEPT Each tier is complete for someone

The idea that holds a tier structure together is completeness per segment. A tier is not a slice of a larger offer, and it is not a discount on the tier above; it is a version of the offer that a particular buyer can live inside without improvising. That is why the design starts from the segments rather than from the elements. A business that begins with its list of services will build tiers by distributing the list, and the result is three versions of the same thing, each missing a piece. A business that begins with the segments asks a different question of each element: which buyer cannot do without this, and which buyer would be paying for something they never use? At Nordec the question produced an unusual answer, because the highest tier is not the tier that most protects a building; the flagship is. Portfolio protects a manager's need to report, which is a real need and a different one, and the tiers are complete for three different kinds of buyer rather than ranked by how much protection they buy.

METHOD Building tiers from interview counts

First, list every element the business sells or could sell, and beside each one write the segments that named it in the interviews, with their counts. Second, put into the lowest tier only what its segment cannot do without, and nothing that the segment did not name. Third, work upward, adding at each step what the next segment named, and nothing else; an element named by nobody is an element that belongs in no tier, however proud the business is of it. Fourth, write each tier's exclusions — the common ones that apply to all tiers, and the particular ones that apply to one — and check that the exclusions describe limits the customer would accept if they were read aloud. Fifth, cost each tier from the hours and materials it contains, and check the cost and the margin before Book A7 sets the price; a tier that loses money at every price its segment will pay is a tier to redesign rather than to discount. Deliverable: three columns, each complete, each with its exclusions and its cost. Pitfall: filling the lowest tier with what the business finds cheap to deliver rather than with what its segment named.

4.4 Bundling: Pure, Mixed and Unbundled

Bundling is the sale of two or more separate products in one package, and the word "separate" has a specific meaning in the marketing literature: separate markets exist for the products, so that each of them is bought and sold on its own somewhere. A study that reviewed the field and proposed a synthesis for marketing defines it that way, and separates two things that are often confused under one word. Price bundling sells the separate products in one package at a discount without integrating them, so the bundle adds no value by itself and needs the discount to exist. Product bundling integrates the products, so that the package is worth more than the parts because of what the combination does. The article offers consistent definitions, sets out rules on legality and proposes which strategy fits which context, and it uses familiar examples: season tickets, luggage sets, fixed-price restaurant menus, and bundled internet service (Stremersch & Tellis, 2002).

The difference between the two is not academic, and Nordec's tiers are an example of it. Take the record and the response promise. A building health record sold on its own is a document. Sold with an urgent response commitment, it becomes something else: the record of what is installed in the building is what allows the technician in the van to bring the right part on the first visit, which is what makes the ninety-minute commitment plausible. The elements are integrated, so the tier is a product bundle rather than a price bundle, and the proof is that removing the discount would not destroy it. That distinction is worth stating plainly to a business that is thinking about tiers, because a tier that is only a discount is a tier that will be renegotiated the moment a customer asks for the pieces at the discount separately.

The trade uses three terms for how a package is offered, and this book adopts them in its own words rather than attributing them to any source. Pure bundling sells only the bundle: the customer takes the package or takes nothing, and no part of it is available alone. Mixed bundling sells the bundle and also sells its parts separately, so a customer may assemble their own version at a different total. Unbundled sale puts no package together at all and prices every component on its own, which is where most service businesses begin and where many stay. Each has a use. Pure bundling is the right shape when the parts are only worth having together; mixed bundling is the shape that suits an offer with several segments of different sizes, because it lets a small buyer take a piece of a larger tier without paying for the whole; unbundled sale is honest and puts the assembling work on the buyer, which is exactly the work most buyers will not do.

Nordec landed on mixed bundling, and the reason is visible in the two add-ons. The tiers themselves are bundles, and they are product bundles: Preventive is the full preventive program, and its elements reinforce one another. Two of the elements of higher tiers are also sold alone, to the tier below. The capital planning note, which is part of Preventive and Portfolio, is sold to an Essential building at $420 a year. The quarterly review, which is part of Portfolio, is sold to a Preventive customer at $140 a quarter. The result is a structure in which a customer can buy the tier that fits and add the one or two things they want from above, which is the practical advantage of mixed bundling, and it also produces the risk that section 4.5 takes up: an add-on can protect the tier it crosses or it can empty the tier above it.

There is one more choice inside bundling that a service business has to make, and it concerns who is allowed to buy which bundle. That choice is the fence, and at Nordec it is written into the lowest tier rather than into the bundle above it: Essential is sold only to buildings under 3,500 square metres. A fence placed this way lets the business offer a smaller package without offering a cheaper route to the same package, and it keeps the tiers tied to segments rather than to budgets. The price gap between the tiers — Essential at 67.6% of the Preventive price at $725, and Portfolio at 124.1% of it — is the other half of the same question, and Book A7 sets it in its Chapter 5 with the reasoning that belongs to pricing.

THE RULE A package must say what it excludes

Book A2 (The Business Plan) teaches that a business defines itself as much by what it refuses as by what it sells, and a package is where that definition becomes visible to the buyer. A package whose limits are hidden misleads by omission of its general impression, and in Canada that is not only a matter of manners. Under section 74.01(1)(a) of the Competition Act (R.S.C. 1985, c. C-34), a representation to the public that is false or misleading in a material respect is reviewable conduct, and section 74.01(1)(b) adds that a representation about the performance, efficacy or length of life of a product that is not based on an adequate and proper test is likewise reviewable, with the burden of proving the test resting on the person who makes the representation. The same statute carries the criminal offense at section 52, which does not require proof that anyone was actually deceived. A tier that says "unlimited" and means three visits a month fails this rule whatever its price. The book explains the law and does not replace the advice of a lawyer.

QUÉBEC & CANADA NOTE Selling a package to a consumer

When a package is sold to a consumer in Québec rather than to a business, the Consumer Protection Act (CQLR c P-40.1) applies, and it does three things a package designer must know. Under sections 41 to 43, the goods and services must conform to the statements and the advertising of the merchant or the manufacturer, which bind them, and a warranty mentioned in a statement or an advertisement binds even if it is not written in the contract — so a tier whose description promises a result creates the obligation it describes. Under section 44, an exclusion from a conventional warranty must be set out clearly, in separate successive clauses, which is a stricter form than a line buried in a general terms page. Under sections 38.1, 38.2 and 38.8, in force on October 5, 2026, new goods of listed kinds sold or leased to consumers — among them ranges, refrigerators, washing machines, television sets, computers, cellular telephones, air conditioners and heat pumps — carry a warranty of good working order whose duration is set by regulation, covering parts and labor, and the merchant must state that duration prominently near the advertised price. A guarantee attached to a tier cannot remove the legal warranty. Nordec sells to businesses, so its tiers are governed in Québec by the Civil Code of Québec, and a reader who sells to consumers should treat the rules above as their setting. This book explains the law and does not replace the advice of a lawyer.

4.5 Add-ons and Options

An add-on is an element of a higher tier sold on its own to a customer of a lower tier, and its purpose is to let that customer take one step in contents without taking the whole step in price. The reason to build one is arithmetic: a customer who cannot afford the flagship but needs one of its elements is otherwise a customer who leaves, and the twelve requests that had nothing to buy at Nordec were exactly that customer, asking at the counter. The reason to build one carefully is that an add-on sits on the boundary between two tiers, and everything that happens on that boundary changes how the tiers above and below are read.

Two outcomes are possible, and the difference between them is worth stating as a test. An add-on protects the tier it comes from when the customer who buys it would not otherwise have moved up, and when the total a customer pays for the lower tier plus the add-on remains clearly below the price of the tier above. It empties the tier above when the combination gives a customer nearly everything the higher tier contains for less money, in which case the higher tier has become a longer way of buying the same thing and will be chosen by nobody. The test is not about price alone: the higher tier must keep something the combination does not contain. At Nordec the flagship keeps the monthly report, the capital planning note at renewal, the full monthly visit and the response commitment together, so the Essential building that buys the note for $420 a year has bought a document rather than a program. The quarterly review add-on is the more delicate of the two, because a Preventive customer who buys four reviews a year has bought the review that Portfolio includes; what Portfolio keeps is the mid-contract note, the tenant request line and the consolidated report, so the combination is close but not complete.

The two add-ons were priced with the same cost basis as the tiers. The capital planning note, sold to an Essential building at $420 a year, costs $162 to produce — three hours of work at the case's loaded rate — which leaves $258 and a margin of 61.43%. The quarterly review, sold to a Preventive customer at $140 a quarter, costs $81 and leaves $59, a margin of 42.14%. The two margins are far apart, and that is deliberate rather than accidental: the note is a document produced once and it is profitable to sell alone, while the review is a meeting that consumes the lead technician's time, and its price is set to make the meeting available rather than to make it a profit center. Neither add-on is sold below its cost, and a business that finds an add-on priced below its cost has found either a mis-priced add-on or an element that belongs in a tier.

Add-ons also do something for the business that is easy to overlook: they turn a request into a line. When a property manager asks for a review meeting, a business with an add-on answers with a price and a date; a business without one either gives the meeting away or refuses it, and both answers cost more than the meeting. The requests recorded at the refresh — five for a quarterly review across the whole base, four from owners for help with replacement projects — are small numbers that would have been given away free in a business whose offer had no place to put them. An add-on is that place, and the discipline it imposes is that the element has a cost, a price and a line in the invoice.

There is a limit to how many add-ons an offer can carry, and the limit is the buyer's patience rather than the business's capacity. The customer of the lowest tier at Nordec sees one add-on. The customer of the flagship sees one. The customer of the upper tier sees none, because the upper tier contains them. That is a small menu by the standards of a business that has decided to sell everything separately, and it is deliberate: an add-on is an option, and an option is a question the buyer has to answer before they can decide whether to buy, which brings the chapter to the subject of section 4.6.

The following table sets out what each of Nordec's tiers and the two add-ons leaves behind, using the case's cost basis. Read it with one question in mind: whether the margin of each line is consistent with the work that line requires, since a tier and an add-on that look similar on an invoice can differ widely in what they cost the business to deliver.

Table 4.2 — The cost and margin of each tier and add-on

DO THE MATH What each tier earns

Illustrative figures; the case is fictional. Step 1 — Essential is sold at $490 a month and costs $219.63 to serve, which leaves $490.00 − $219.63 = $270.37, a margin of 55.18%. Step 2 — Preventive is sold to a new customer from $725 a month and costs $336.73, which leaves $388.27, a margin of 53.56%; the same tier at the current customers' $712 leaves $375.27, a margin of 52.71%. Step 3 — Portfolio is sold at $900 a month and costs $404.23 — the Preventive cost plus the $67.50 of its four additions — which leaves $495.77, a margin of 55.09%. Step 4 — The increment itself is the useful figure: $175 of extra price against $67.50 of extra cost leaves $107.50, a margin of 61.43% on the step from Preventive to Portfolio. Step 5 — No tier and no add-on is sold below its cost, and the flagship is deliberately not the most profitable per dollar of price; the business wants to sell the tier that keeps a building under control, and the arithmetic above is a check on the contents rather than a ranking of the tiers by how much they earn.

WATCH OUT The tier nobody can tell apart

An option is undifferentiated when the buyer cannot name what the extra amount buys. The symptom is a sentence every business hears and few write down: "what do I get for the difference?" At Nordec the risk sat on the step from Preventive to Portfolio, where $175 a month buys four things. If those four are described as "a closer relationship" or "enhanced support", the step is undifferentiated and the buyer hears a price increase. If they are described as a mid-contract note for the buildings the first note flagged, a review meeting each quarter, a line for tenants' non-urgent requests, and one consolidated report across the manager's buildings, the step is a list. The remedy for an undifferentiated option is never in the price, because lowering the amount does not make the contents nameable; it is in the contents and in the words. Book A7 takes the same question from the pricing side in its Chapter 7, and its conclusion matches this one: a discount on a step nobody understands produces a cheaper step nobody understands.

4.6 Choice Architecture and Too Many Choices

The number of options a customer faces is itself a design decision, and it has a cost that a business rarely counts because it never appears on an invoice. The three studies reported together are the clearest evidence in this chapter. People were more likely to buy jams or chocolates, or to write an optional essay, when they were offered 6 options rather than 24 or 30, and they were more satisfied with the choice they made. The first study ran a tasting booth in an upscale grocery store in Menlo Park, California, showing 6 or 24 jams on alternate Saturdays: the larger display stopped 60% of passing customers against 40% at the small one, but about 30% of those who stopped at the small display bought a jar, against 3% of those who stopped at the large one (Iyengar & Lepper, 2000). The counter stopped more people and sold fewer jars.

A service business is not a jam display, and the finding does not mean that every offer must have three options. What it means is that the number of options has a cost that is paid in decisions not taken, and that a business should choose the number rather than inherit it. Nordec's offer asks a buyer to face three tiers and, depending on the tier, one add-on. A business that sold every element separately would ask a buyer to answer eight questions about their own building before they could reach a total, and each of those questions is a place where the buyer stops. The three tiers plus the two add-ons is not a small menu because small menus are fashionable; it is a small menu because the elements were grouped until each group could be described in a sentence.

The order matters as much as the number. A tier structure is read from somewhere, and the order in which the columns are presented changes what the reader holds on to. Nordec presents the three tiers in ascending order, with the flagship in the middle and marked as the recommended tier for an owner-operator, because that is the honest answer for the segment that most often reads a proposal, and because a business that buries its flagship in the middle of a page is making its reader look for it. The recommendation is not a device; it is an opinion the business is willing to defend, and the defense is written into the tier's contents. A business that recommends the middle tier to everybody has not made a recommendation, and a business that recommends the highest tier to everybody has made a sales pitch.

The labels are part of the design, and the test for a label is whether a customer could use it in a sentence to a colleague. The three names at Nordec were chosen in the buyers' words. Essential is what the owner of a small building said they needed: the essentials, no more. Preventive is what the offer is, a preventive program, and it is the word the business already used for its contract. Portfolio is what the property manager has: several buildings, reported on together. None of the three names contains a superlative, a precious metal, a number or a word invented for the occasion, and each of them can be said at a reception without embarrassment. A tier called Platinum would have told a buyer nothing about the contents, and a tier called Standard would have told them something false about the others.

There is a temptation to read this section as advice about steering buyers, and it should be closed rather than left open. The findings in this chapter describe how people choose under different arrangements of options, and none of them licenses a business to arrange its options so that a buyer takes something that is not good for them. What a business may honestly do with them is make each option clear, present the option it believes in, keep the number of options to what a buyer can compare, and then let the buyer choose. The limit is easy to state and hard to fake: if the buyer knew everything the business knows about the options, would the arrangement still be defensible? At Nordec the answer is yes, and the reason is that the recommended tier is the one the business is built to deliver, and the two other tiers are complete for the segments they serve.

4.7 Fences, Segments and Migration

A fence is the condition that decides who may buy a lower tier, and it is the part of a tier structure that a business is most likely to leave out. Without a fence, a lower tier is simply a cheaper version of the flagship, and every customer with a budget will find their way to it. The consequence is not only lost revenue; it is a business whose flagship stops being delivered, because the flagship was the version the operation was built around and the customers who remain on it are the ones who forgot to move down. A fence makes the lower tier a different offer for a different buyer, and Nordec's fence is a size: Essential is sold only for buildings under 3,500 square metres. The size was chosen because it is the line the interviews drew — the two buildings lost to the trades billed by the call were both under that size, and their owners both said the contract was more than the building needed — and it is checkable from the building record rather than from a judgment about the customer.

The fence and the price gap are two halves of one decision, and the half that belongs to this book is the fence. The gap is Book A7's work, in its Chapter 5, and it is the place where the difference between the tiers is turned into an amount. The two are read together, which is why this chapter states the ratios without building on them: at the flagship's $725, Essential stands at 67.6% of the Preventive price and Portfolio at 124.1%. What matters here is that the gap is large enough that the fence is doing work rather than decoration, and small enough that a customer who belongs in the middle tier is not pushed to the bottom one by the difference. A gap that is too wide makes the lower tier look like a different company; a gap that is too narrow makes the tiers interchangeable.

The migration rule at Nordec is the piece of the design that protects the existing base, and it has three parts. At renewal, every current customer is offered the three tiers, in writing, with the contents and the prices, so that the choice is real rather than notional. The default, if the customer says nothing, is Preventive at the customer's current price: no existing contract is moved, no price is changed, and no element is withdrawn. And the result is measured: downgrades are counted at every renewal window, because the number of downgrades is the honest test of whether the fence and the gap are in the right place. A business that offers its own customers a new and cheaper way to buy and then does not count how many take it has not run a migration; it has run a sale.

The default deserves its own paragraph, because it is the opposite of what a business usually wants to do with a new structure. The instinct is to move existing customers to the flagship's new price, or to a higher tier, at the first renewal after the redesign. The case does the opposite: the current customer keeps the current contract at the current price, and the new prices apply to new customers. There are two reasons. The first is that a customer who is doing nothing wrong should not be charged more for the business's own redesign, and a business that reprices its whole base at once has given every one of those customers a reason to open a conversation with a competitor. The second is arithmetic: the flagship's margin at the current customers' $712 is 52.71%, which is above the margin at which the contract was being delivered before the additions were identified, and it is the margin that pays for the redesign. A migration that keeps the base is a migration that can be paid for out of new signings; a migration that disturbs the base has to be paid for twice.

Two smaller decisions belong to the migration and are worth naming. The first is timing. The mid-contract capital planning note had been decided at the refresh and not announced, and it moved into Portfolio before any customer had received it, which is why the move cost nothing; an element that customers already receive cannot be moved up a tier without taking it away from someone. The second is the treatment of the upper tier. Portfolio is sold to a property manager, and a property manager usually holds several buildings; the tier is priced per building, so the manager buys it building by building, and the consolidated report makes the several buildings one relationship. That arithmetic is the reason the three property management firms mattered: they held 14 of the 46 buildings, or 30.4% of the base, and a tier that fits them changes the size of the accounts the business can hold rather than only the price of each one.

4.8 The Case's Three Tiers

The three tiers were built in a single working session that ran for an afternoon, and the reason it ran for only an afternoon is that the preparatory work had already been done. Sophie brought the twenty interview transcripts and the counts of the elements named among each person's top three. The lead technician brought the hours: what the monthly visit takes on a core checklist and in full, what can be finished inside two hours on a call, which parts habitually go missing, what the four additions of the upper tier would take in a technician's month. The administrator brought the off-site time: the one-page report, the record, the two notes, the review meeting, the tenant request line, the consolidated report, and how each of those interacts with a month that already has a scheduling job in it. The session's output was three columns and two add-ons, and the price came later, from Book A7's method.

The session began from the segments rather than from the elements, which is the method of section 4.3 applied rather than described. The small building owner who wants the essentials and nothing else became the first segment, and the elements they named became Essential: the monthly visit on a core checklist, urgent calls within ninety minutes day and night with up to two hours of work, the building health record, and a visit sheet emailed after each visit. The one-page monthly report was placed in the flagship rather than in Essential, because the report was named by 6 people of whom 5 were property managers, and a report exists to be forwarded to somebody who is not in the building. The capital planning note was placed in the flagship, because it was named by 12 people including 7 of the 8 owner-operator customers. The four property managers' requests — the mid-contract note, the quarterly review, the tenant request line and the consolidated report — became the step from the flagship to the upper tier, at a cost of $67.50 a month per building.

The fourth tier was drafted in the same session and removed in the same session, and its removal is the reason the structure has three steps rather than four. Portfolio Plus contained one thing, after-hours dispatch for non-urgent requests, and it failed on two counts. It undid a decision of Book A3 (Strategy and Competitive Positioning), where Nordec had drawn a reduce on after-hours coverage and made the night hours the property of genuine emergencies; a tier that sold nights for non-urgent work would have taken the business back to the position the strategy had left. And no interview had asked for it except the two owners who had been lost on after-hours coverage in the strategy year, who were two buyers rather than a segment. A tier with no segment is a list of costs looking for a reason, and the case is worth stating plainly because the temptation to add a tier above the flagship is strong: the space above the top column looks empty, and the business's own team will often fill it with the one thing they find hardest to deliver.

The names were chosen at the end and in the buyers' words, following the test of section 4.6. Essential came from the small building owners who said the contract was more than the building needed; it names what they want rather than what they are settling for. Preventive is the word the business already used for its contract, and it is the honest description of what the flagship is. Portfolio came from the language of the property managers, who talk about their portfolios of buildings and about reporting to owners. None of the names is a superlative, and none of them implies that the tier below is deficient. A customer who buys Essential has bought the essentials, not a lesser version of something else, and the difference matters to the way the tier is delivered as well as to the way it is sold.

The costs were checked before the tiers left the room, using the same basis throughout. Essential costs $219.63 a month to serve, which is the 2.5-hour core visit with its travel and consumables, the share of call work that falls to a smaller building, the upkeep of the record and the visit sheet. Preventive costs $336.73, which is the full monthly visit, the same call work, the record, the report and the capital planning note at renewal. Portfolio costs $404.23, which is Preventive plus $67.50 for the four additions. The margins that follow from the Book A7 prices are 55.18% for Essential, 53.56% for Preventive at $725 and 55.09% for Portfolio at $900. Read together, the three costs confirm the design's intention: the flagship is not the most profitable per dollar, and it is the one the business wants to sell, because it is the one that keeps a building under control and the one that nearly every current customer already holds.

The last decision of the session was the promise that binds the migration: nothing that a current Preventive customer receives is taken away. The monthly report left Essential, which is a tier that did not exist, and was replaced there by the visit sheet. The mid-contract note entered Portfolio before it was ever announced, so no customer lost it. The quarterly review and the tenant request line are offered to Preventive customers as add-ons rather than taken from them, and no customer of the flagship is asked to pay for a new tier to keep what they had. That promise is what allows the three tiers to be offered to the existing base at renewal with a default that changes nothing, and it is the reason the migration described in section 4.7 is a choice for the customer rather than a decision taken about them.

TOOLS OF THE TRADE What a package design needs

Five functions carry a package design, and none of them requires a product. The first is a spreadsheet holding the elements, the segments that named them, their counts and their costs, one row per element, because the whole method in section 4.3 is a sort of that sheet. The second is a one-page comparison of the tiers, in the customer's words, with the elements as rows and the tiers as columns, which is what becomes the proposal page and the first page a buyer reads. The third is a cost sheet per tier, built from the same loaded rates and hours as the operational costing, so that a tier's margin is visible before Book A7 sets its price. The fourth is a catalog or quoting tool that can hold more than one version of an offer and quote an add-on beside it, because a structure that exists only in the owner's head is renegotiated at every proposal; the function to look for is the ability to store several packages with their own contents, prices and descriptions, and to select one on a document. The fifth is a written exclusion list, kept with the tier descriptions, so that what a tier does not include is stated in the proposal rather than discovered at the first call. No brand name belongs on this list.

IN LA MACHINE Packages without tiers

Two pages of the application read for this book hold packages, and both of them hold each package as a single flat item. Business → Session Packages (productivity/packages, also reached at business/packages) defines packages of sessions or minutes, each with a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image; the page shows the profit and the margin per sale as the package is edited, and a package can be added, edited, deactivated or deleted. Business → Subscription Packages (business/subscription-packages) defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown, a description and an image; a package that is disabled lets existing subscribers keep their access until the end of their current billing period. Read together, the two pages can carry a three-tier structure honestly: three subscription packages named Essential, Preventive and Portfolio, each with its own price and its own cost, would show the margin of each one exactly as this chapter's arithmetic requires. What the platform does not have is a tier construct. There is no parent package with steps, no bundle that gathers items, and no add-on that attaches to a package, so the relationship between the three tiers lives in the descriptions and in the proposal rather than in the record, and an add-on is a fourth package of its own. The platform's own plans page (plans) presents its plans as tiers with a checklist of included features and an upgrade or downgrade path, which is what a tiered offer looks like; its prices are never to be quoted as a model for anyone else's.

Costly Mistakes

• Selling one package to every segment. A single contract sold to the small building and to the property manager is too much for one and too little for the other, and both of them will say so in the only language available to them, which is the price.
• Building tiers by distributing a list of services. Tiers made by cutting the existing list into three pieces produce three versions of the same thing, each missing a piece, instead of three offers that are each complete for a buyer.
• Leaving the exclusions to the customer's assumption. A tier whose limits are hidden misleads by its general impression, and an unwritten exclusion becomes an argument at the first call rather than a boundary in the contract.
• Adding a tier because the space above the flagship looks empty. Portfolio Plus was removed because it undid a strategic refusal and because no segment had asked for it; a tier with no segment is a column of costs waiting to be sold to the wrong buyer.
• Adding an add-on that empties the tier above it. An add-on protects the tier above when the customer would not otherwise move up and the combination stays clearly below the higher price; when the combination gives nearly everything for less, the higher tier will be chosen by nobody.
• Repricing the whole existing base at the first renewal after a redesign. The default that keeps the current contract at the current price is what makes a migration a choice for the customer, and it is what pays for the redesign out of new signings rather than out of the base.
• Leaving the lower tier unfenced. Without a condition on who may buy it, the lower tier becomes a cheaper route to the flagship, and the flagship stops being delivered to the customers who remain on it.

Key Takeaways

• Packaging arranges an offer's elements into named choices a buyer can compare, and it settles who each choice is for, what it contains, what it excludes and what it is called, leaving the price to Book A7.
• The three-step structure puts the segment that needs the essentials at the bottom, the operator's full program in the middle as the flagship, and the multi-building manager's reporting instrument at the top; each tier is complete for its own segment rather than a slice of a larger offer.
• Bundling is the sale of two or more separate products in one package; price bundling sells the parts at a discount without integrating them, product bundling integrates them, and Nordec's tiers are product bundles sold in a mixed structure, with two elements also sold alone as add-ons (Stremersch & Tellis, 2002).
• An add-on lets a lower-tier customer take one step in contents without taking the whole step in price; the capital planning note at $420 a year leaves a margin of 61.43% and the quarterly review at $140 a quarter leaves 42.14%.
• The research on choice describes what happens as options are added, read and multiplied — the attraction effect in 1982, tradeoff contrast and extremeness aversion in 1992, and the demotivation of 24 or 30 options in 2000 — and it is used here to design options that can be defended, never to steer a buyer (Huber et al., 1982; Simonson & Tversky, 1992; Iyengar & Lepper, 2000).
• A lower tier needs a fence so that it serves a new segment rather than cannibalizing the flagship; Nordec's is buildings under 3,500 square metres, and at renewal every current customer is offered the three tiers with the current contract at the current price as the default, while downgrades are counted.

Your Action Plan

1. Write the element list with the segment that named each element. Put every element of the offer on one row, with the segments and the counts that named it in the interviews, and mark the elements nobody named. Deliverable: one element table sourced from the interviews of Chapter 2. Verification: every element has at least one segment and a count beside it, and the elements with no segment are marked for removal.
2. Draft three columns, each complete for one segment, with its exclusions. Place in the lowest tier only what its segment cannot do without, add at each step what the next segment named, and write the common and particular exclusions under each column. Deliverable: three tier descriptions in the customer's words, with exclusions. Verification: a reader can say aloud what each tier adds and what each tier does not include without opening the contract.
3. Cost each tier and each add-on before any price is discussed. Build a cost sheet per tier from the same loaded rates and hours the operation uses, and show the margin at the price Book A7 will set. Deliverable: a cost sheet per tier and per add-on. Verification: no line is sold below its cost, and the increment between two tiers has its own cost and margin.
4. Set the fence and write the migration rule. Name the condition that limits the lowest tier and the segment it opens, and write the renewal rule that offers the tiers to current customers with the current contract and price as the default. Deliverable: one fence and one written migration rule. Verification: the fence is checkable from a record the business already keeps, and the migration rule names the default and the measure of downgrades.
5. Put the tiers and the add-ons into the documents and the system. Place the three columns and the add-ons in the proposal template and in the catalog or quoting tool, with the exclusions on the page, and check that the tool can quote an add-on beside a tier. Deliverable: a proposal page and a stored package structure. Verification: a person who has not read this chapter can quote the right tier and the right add-on from the tool alone, and the comparison page fits on one page.

In La Machine

Packaging is the one part of this book's subject the platform supports through two pages that hold a single flat item at a time (La Machine, 2026). Business → Session Packages (productivity/packages, also reached at business/packages) defines packages of sessions or minutes, each with a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image, and the page shows the profit and the margin per sale as the package is edited; a package can be added, edited, deactivated or deleted, and the same page lists each client's credit balance and history, with credits added by hand recorded as free gifts or adjustments and no payment. Business → Subscription Packages (business/subscription-packages) defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown, a description and an image; a package that is disabled lets existing subscribers keep their access until the end of their current billing period, which is the closest the platform comes to a migration rule. A three-tier structure can be carried honestly as three subscription packages with their own prices and costs, because the margin of each one is visible on the page exactly as section 4.5 requires.

What the platform does not have belongs in the same breath. There is no tier construct: a package is a single flat item, with no parent and no steps, no bundle that gathers items into one price and no add-on that attaches to a package, so the relationship between three tiers lives in their descriptions and in the proposal rather than in the record, and an add-on is a fourth package of its own. The platform has no guarantee or warranty register, no service-level tracking and no service credits, and no quote or proposal module, which means the document a customer reads is prepared outside it. Its own plans page (plans) presents the platform's plans as tiers, each with a checklist of included features and an upgrade or downgrade path, which is a useful illustration of what a tiered offer looks like; the prices shown there belong to the platform and are never to be quoted as a model for another business. Book N1 (Getting Started with La Machine) maps the modules the platform does cover, and a business should check a page against its own need before counting a function as covered.

Self-Check

1. What four questions does packaging settle, which question does it leave to Book A7, and why does the order matter?
2. What is the design rule for the contents of a tier, and what does the rule require at each step up?
3. What is the difference between price bundling and product bundling, and which one are Nordec's tiers, and why?
4. When does an add-on protect the tier above it, and when does it empty that tier?
5. What is a fence for, what is Nordec's fence, and what is the migration rule that goes with it?

Answers. 1. Packaging settles who each choice is for, what it contains, what it excludes and what it is called; it leaves the price to Book A7 (Pricing Strategy and Revenue Models), and the order matters because a package whose contents are undecided cannot be priced and cannot be rescued by a discount. 2. Each tier must be complete for the segment it serves, so that a customer in that segment can buy it and run the building without coming back for an element that should have been in it; each step up adds what the next segment named in the interviews, and only that, and each tier states its exclusions. 3. Price bundling sells separate products in one package at a discount without integrating them, so the bundle adds no value by itself and needs the discount; product bundling integrates the products, so the package is worth more than the parts (Stremersch & Tellis, 2002). Nordec's tiers are product bundles, because the building health record and the response commitment reinforce each other: the record is what allows the technician to bring the right part on the first visit, which is what makes the ninety-minute promise plausible. 4. An add-on protects the tier above when the customer who buys it would not otherwise have moved up and the combination stays clearly below the higher tier's price, and it empties the tier when the combination gives nearly everything the higher tier contains for less money; the test is that the higher tier must keep something the combination does not contain. 5. A fence decides who may buy the lower tier, so that the lower tier serves a new segment rather than giving the flagship's customers a cheaper route to the same thing; Nordec's fence limits Essential to buildings under 3,500 square metres, and the migration rule offers the three tiers at renewal with the current contract at the current price as the default for a customer who says nothing, while downgrades are counted.

Chapter References

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

Huber, J., Payne, J. W., & Puto, C. (1982). Adding asymmetrically dominated alternatives: Violations of regularity and the similarity hypothesis. Journal of Consumer Research, 9(1), 90–98. https://doi.org/10.1086/208899

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too much of a good thing? Journal of Personality and Social Psychology, 79(6), 995–1006. https://doi.org/10.1037/0022-3514.79.6.995

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages, business/subscription-packages and plans.

Simonson, I., & Tversky, A. (1992). Choice in context: Tradeoff contrast and extremeness aversion. Journal of Marketing Research, 29(3), 281–295. https://doi.org/10.1177/002224379202900301

Stremersch, S., & Tellis, G. J. (2002). Strategic bundling of products and prices: A new synthesis for marketing. Journal of Marketing, 66(1), 55–72. https://doi.org/10.1509/jmkg.66.1.55.18455

Chapter 5 — Guarantees, Service Levels and Risk Reversal

A buyer who cannot inspect a service before buying it carries a risk the seller could carry instead, and a guarantee, a service level with credits or a right to leave moves that risk to the business that controls the outcome — at a cost the business can calculate, and within limits the law sets.

What You'll Be Able to Do

• Distinguish the warranty the law imposes from the guarantee a business chooses, and say which of the two a given promise belongs to.
• Design a guarantee that is unconditional, meaningful to the customer, easy to understand, easy to invoke and easy to collect, in the terms the customer uses.
• Write a service level as a measurable promise with a threshold and a remedy, and choose the record that proves whether the level was met.
• Price a guarantee at the performance the business delivers today and at the performance it would deliver in a bad year, before the promise is published.
• Choose among the forms of risk reversal — a trial, a right to leave, a credit, a refund, a fixed price, payment for results — the one that reverses the risk the buyer actually feels.
• Recognize the legal limits on a guarantee, from the Competition Act to the Québec Consumer Protection Act and the Civil Code of Québec, and know when a lawyer must read the wording.

Why It Matters

A service is bought before it is seen. The customer signs, pays and waits, and what they receive is produced afterwards, by people they have not met, in conditions nobody has described to them. The seller knows what the service will be, because the seller has designed it and performed it many times. The buyer does not, and the gap between those two positions is where hesitation lives. It is also where a competitor wins a contract without being better: the buyer who cannot judge the work will judge the promise, and the offer that carries less risk is easier to sign even when it costs more.

The asymmetry is old enough to have been described plainly. Products such as cars, cameras and toaster ovens come with warranties, while services such as auto repair, wedding photography or catered dinners almost never do, although a service cannot be inspected before it is bought (Hart, 1988). The observation is not a complaint about sellers. It is a description of a gap in the market: the goods economy learned to carry the buyer's risk, and the service economy largely did not, even though services are the harder purchase. A small business that closes that gap is not being generous. It is answering a question the buyer is already asking, and answering it in a way the buyer can check.

The cost of carrying risk is real, and it is the reason this chapter spends most of its length on arithmetic rather than on enthusiasm. A promise that the business fails to keep costs money, and the amount is not a mystery: it is the number of failures multiplied by the remedy. What makes the arithmetic manageable is that the business already knows, or can find out in a week, how often it fails today. A guarantee priced against a measured failure rate is a decision. A guarantee written because it sounds confident is a bet, and the business that writes it will discover the odds at the worst possible moment, in a month when the failures arrive together.

In This Chapter

Section 5.1 explains why risk stands between the buyer and the offer, and what it means to move that risk to the party that controls the outcome. Section 5.2 separates the warranty the law imposes from the guarantee a business chooses, in Québec, in the United States and between businesses. Section 5.3 sets out how to design a guarantee that works, using the qualities the literature on service guarantees identifies. Section 5.4 turns a promise into a service level with a threshold, a record and a credit, and gives Nordec's levels by tier. Section 5.5 surveys the forms of risk reversal, from a trial to payment for results, and places them on a ladder. Section 5.6 prices a guarantee at the performance the business delivers today and at the performance of a bad year. Section 5.7 sets out the rules that apply, from the Competition Act to the Québec Consumer Protection Act and the Civil Code of Québec. Section 5.8 tells how Nordec chose its 90-minute guarantee, what it costs, and what it changed inside the business.

5.1 Why Risk Stands Between the Buyer and the Offer

A service cannot be inspected before it is bought. That sentence is the foundation of this chapter, and it is worth unpacking because every promise a business makes is an answer to it. When a customer buys a chair, they can sit in it, turn it over, and see the joint where the leg meets the seat. When a customer buys a year of preventive maintenance, there is nothing to sit in. The work will be performed by a technician who has not yet been assigned, on equipment the customer may not understand, at times that have not yet arrived. The customer is buying a description and a set of intentions, and they know it.

The literature on service design has treated this as the central practical problem of selling a service rather than as a philosophical one. The observation that products such as cars, cameras and toaster ovens come with warranties while services such as auto repair, wedding photography or catered dinners almost never do, although a service cannot be inspected before it is bought, is the starting point of the argument for unconditional service guarantees (Hart, 1988). The contrast is not accidental. A manufacturer can test a toaster before it leaves the plant and can repair or replace it afterwards, and the warranty is a cheap way to say so. A service is produced in the customer's presence, cannot be recalled, and cannot be inspected in advance, which is precisely the situation in which a promise would be worth the most.

The buyer's risk is not one risk but four, and a business that wants to reverse risk has to know which of the four its customer feels. The first is that the service does not work: the repair does not cure the problem, the report does not answer the question, the installation does not perform. The second is that it takes too long: the technician arrives on Thursday when the leak began on Monday, the proposal arrives after the decision has been taken elsewhere. The third is that it costs more than was said: the invoice carries hours nobody mentioned, parts nobody approved, a scope that grew quietly. The fourth is that it cannot be undone: the wrong part has been installed, the wrong advice has been followed, the customer has told their own client that the matter is settled.

Each of those four risks sits with somebody, and the question worth asking is who controls the outcome. The business controls whether the repair works, because it chooses the technician, the method and the part. The business controls how long it takes, because it sets the dispatch rules, the staffing and the response commitment. The business controls the cost, because it writes the scope and approves the extra hours. The business controls whether the work can be undone, because it decides what it will touch. In each of the four cases the party that carries the risk in the customer's mind is not the party that controls it, and that mismatch is the whole opportunity. Risk reversal is the practice of moving a risk to the party that controls the outcome.

The mismatch has a price for the seller, even when nothing goes wrong. A buyer who cannot judge the work discounts the offer to cover the uncertainty, and the discount is invisible because it never appears as a line on anything. It appears as a longer decision, a request for a lower price, a preference for the trades billed by the call, a contract that is not renewed. The buyer who is unsure does not say so; the buyer buys less, or buys later, or buys the smaller thing. Reversing the risk removes the discount, and the removal is the return on the promise.

Risk reversal takes three broad forms, and this chapter treats them in order of increasing commitment. The first is a promise with a remedy attached: if the service fails in a stated way, the customer receives a stated compensation without having to argue for it. The second is a measurable promise with a threshold and a credit: a service level, which says how fast, how often or how completely, and pays a fixed amount when the level is missed. The third is a right to leave: a period during which the customer may end the arrangement without penalty, which reverses the risk of having chosen wrongly. All three are promises, and all three cost money when they are broken.

The running case of this book sits squarely in this territory. Nordec Entretien inc. sells preventive maintenance and on-call technical service to commercial buildings, and its customers are businesses: owner-operators of mid-size buildings and property managers who answer to owners. Its contract promises a 90-minute response on urgent calls, day and night, and bills work beyond the scope at $96 an hour plus parts. Every one of the four buyer's risks is present in that arrangement. The work may not cure the fault, the technician may arrive late, the invoice may carry hours the owner did not expect, and a badly handled repair may leave a tenant without heat. The business controls all four, and the customer knows it does.

What this chapter does not do is set a price. The amounts that appear here are the case's own illustrative figures, and the method by which a tier or a guarantee is priced belongs to Book A7 (Pricing Strategy and Revenue Models). This chapter designs the promise, chooses the remedy, measures the failure rate and calculates what the promise costs. The price of the offer that carries it is a separate decision, taken with the method of Book A7, and the two decisions meet in the same sentence on the same page.

DEFINITION Legal warranty, conventional warranty, guarantee, service level and service credit

A legal warranty is the protection the law attaches to a sale whether or not anybody writes it down, and it cannot be removed by a clause in a contract. A conventional warranty is the warranty a merchant or a manufacturer gives in writing or in advertising, and it binds the person who gives it. An additional warranty is what the trade calls an extended warranty: a separate purchase by which a merchant undertakes to bear all or part of the cost of repairing or replacing goods, beyond the conventional warranty given free with them. A guarantee, in the sense this chapter uses, is a voluntary promise a business makes about its own service, with a remedy attached when the promise is not kept. A service level is a measurable promise with a threshold — a time, a rate, a completeness — and a service credit is the fixed amount the customer receives when the threshold is missed, usually applied to the next invoice without a claim.

5.2 Legal Warranties and Voluntary Guarantees

Two kinds of promise are easily confused, and the confusion costs businesses money. The first is the promise the law imposes on a seller, which exists whether or not anybody mentions it and which no clause can remove. The second is the promise the business chooses to make, which exists because the business decided to make it and which the business can shape, price and withdraw. A guarantee designed without an understanding of the first kind is a guarantee that may be redundant, or worse, may contradict something the law already requires.

In Québec, a merchant selling goods to a consumer is bound by a legal warranty that covers four things: the good serves its intended use, it lasts a reasonable time given its price, the contract and the conditions of use, it is free of significant hidden defects, and it matches its description in the contract and in advertising. The law sets no fixed number of years, and the legal warranty applies automatically, at no charge, even when a merchant says the good is sold without warranty (Office de la protection du consommateur, n.d.). The Office de la protection du consommateur distinguishes that legal warranty from the conventional warranty of the merchant or manufacturer and from an additional warranty bought separately, which is the three-way distinction a business should hold in mind before writing anything of its own.

The distinction matters because the three kinds of promise behave differently. The legal warranty cannot be excluded, so a business that tries to exclude it writes a clause that does not do what it says. The conventional warranty binds the merchant or the manufacturer who gives it, and in Québec goods and services must conform to the statements and advertising of the merchant or manufacturer, which bind them, and a warranty mentioned in a statement or an advertisement binds even if it is not written in the contract. An exclusion from a conventional warranty must be set out clearly, in separate successive clauses. The additional warranty is a product in its own right, sold to a consumer who wants coverage beyond what came free.

The same three-way structure exists in the United States in a different vocabulary. A written warranty on a consumer product costing more than $10 must be designated "full" or "limited"; for products over $15, the terms must be set out in a single, clear document and made available to the consumer before the sale; and requiring the consumer to pay for labor, or to return a registration card as a condition of coverage, makes a warranty "limited" (Federal Trade Commission, n.d.). The statute behind the guide defines a consumer product as tangible personal property normally used for personal, family or household purposes, and a written warranty as a written affirmation or promise, made in connection with the sale, on the nature of the material or workmanship, or that the product is defect-free or will meet a specified level of performance over a specified period. A "full" warranty must meet minimum standards: remedy within a reasonable time and without charge, no limit on the duration of implied warranties, an exclusion of consequential damages only if conspicuous, and a refund or replacement after a reasonable number of failed attempts to repair.

Between businesses, the picture is different again, and this is the part that governs Nordec. A contract between two businesses in Québec is governed by the Civil Code of Québec, not by the consumer protection statute. Under article 1726, a seller must warrant the property against latent defects that make it unfit for its intended use. Under article 2100, a contractor or provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice, and, where applicable, ensure that the work or service conforms to the contract; when bound to an obligation of result, they may be relieved of liability only by proving superior force. Under article 2102, before the contract, the provider must give the client, as far as circumstances permit, useful information on the nature of the task and the property and time it requires. Those articles are the floor beneath any guarantee a business-to-business seller writes, and they are the reason a guarantee is an addition to the legal position rather than a substitute for it.

The following table sets out the four kinds of promise side by side, with who gives each one, what it covers, whether the seller can exclude it and where it comes from. Read it as a map of the ground a guarantee occupies: the first three rows are promises the business inherits or sells, and only the fourth is the promise the business invents, which is the one this chapter is about.

Table 5.1 — Four kinds of promise

QUÉBEC & CANADA NOTE The legal warranty no contract removes

In Québec, a merchant who sells goods to a consumer is bound by a legal warranty that the law sets out in sections 37 and 38 of the Consumer Protection Act: the good must be fit for the purposes for which goods of that kind are ordinarily used, and it must be durable in normal use for a reasonable length of time, given its price, the terms of the contract and the conditions of use. The law sets no fixed number of years, and the warranty applies automatically, at no charge, even when a merchant says the good is sold without warranty (Office de la protection du consommateur, n.d.). Two consequences follow for a business designing a guarantee. First, the legal warranty protects consumers, not businesses buying from businesses, so a business-to-business seller such as Nordec is not bound by it and is governed instead by the Civil Code of Québec. Second, a business that sells to consumers cannot write its way out of the legal warranty, and a guarantee it adds sits on top of the legal warranty rather than replacing it. This book explains the law; it does not replace the advice of a lawyer.

5.3 Designing a Guarantee That Works

A guarantee is a design object, and like any design object it can be well made or badly made. The literature on service guarantees gives a set of qualities that separate the two, and they are worth taking one at a time because each of them corresponds to a way a guarantee fails in practice. The strongest guarantees are unconditional, meaningful to the customer, easy to understand and to invoke, and easy to collect (Hart, 1988). A guarantee that misses any one of those four is not a weaker version of a good guarantee; it is a different instrument, and it usually costs the business money while producing none of the benefit.

The first quality, that the guarantee be unconditional, is the one businesses resist most. A conditional guarantee — one that applies if the customer called before noon, if the equipment was under a maintenance contract, if the failure was not caused by wear — is easier to write and cheaper to honor, and it is nearly worthless to the buyer, because the buyer cannot tell in advance whether their situation falls inside the conditions. The unconditional guarantee is valuable precisely because the buyer does not have to work out whether it applies. The business pays for that clarity, and the payment is the point.

The second quality, that the guarantee be meaningful to the customer, is where the interviews of Chapter 2 do their work. A guarantee on a dimension the customer does not care about is a cost with no return. The customer's own words are the test: if the customer cannot finish the sentence "what I really need is..." with the thing being guaranteed, the guarantee is aimed at the wrong target. In the case followed through this book, the interviews put urgent response within 90 minutes in the top three for 18 of 20 people, which is as clear a signal as a small business is likely to get about which promise is worth making.

The third and fourth qualities, that the guarantee be easy to understand and easy to invoke, are about the customer's effort rather than the business's generosity. A guarantee that requires a claim form, a telephone call within a stated period, a reference number or a written explanation is a guarantee that most customers will not use, and the ones who do use it will be the ones with the most time and the least goodwill. The remedy should be automatic: the business knows whether it failed, because the business holds the record, and the customer should not have to prove it. The fourth quality, easy to collect, is the same principle applied to the money: a credit that appears on the next invoice without a request is collected; a credit that requires a conversation with an account manager is a negotiation.

A guarantee does three things inside the business, and the second and third are usually more valuable than the first. The first is that it compensates the customer when the service fails, which is its visible purpose. The second is that it forces the business to define good service in the customer's terms, because a guarantee cannot be written in the language of internal process. A business that promises "a technician on site within 90 minutes" has defined good service from outside; a business that promises "a timely response" has defined nothing. The third is that a guarantee produces failure data. Every credit paid is a recorded failure with a date, a customer and a cause, and a business that pays credits is a business that knows where its service breaks. The data is a by-product of the promise, and it is often worth more than the promise.

Choosing the promise is therefore a two-part decision: which promise, and which remedy. The promise should be the one the customer values most among the things the business controls, which is the intersection of the interviews of Chapter 2 and the analysis of control in section 5.1. The remedy should be large enough to be felt and small enough to be paid without argument. A remedy that is too small is an insult; a remedy that is too large turns every failure into a crisis and encourages the business to fight the claim. The case's own remedy, a credit of $150, is about a fifth of an average month's contract value, which is the kind of proportion that makes a customer notice without making the business argue.

The design also has to say what is not guaranteed, and to say it in the same document and the same voice. A guarantee that is silent about its limits will be read as covering everything, and the first customer who discovers otherwise will feel misled, which is a worse outcome than a narrower promise honestly stated. Nordec's contract excludes parts, work beyond two hours per incident, roof and envelope, elevators, and anything requiring a licensed trade the company does not hold, and the guarantee has to live inside those exclusions rather than beside them. The wording that does this well is short and specific, and it names the excluded things rather than gesturing at them.

METHOD Designing a guarantee in five steps

First, name the promise customers value most, in their words, taken from the interviews of Chapter 2 rather than from the business's own sense of what it does well. Second, measure how often the business fails that promise today, from the records it already keeps: arrival times, completion dates, invoice disputes, callbacks. Third, choose a remedy that is felt and make it automatic, so that the customer never has to claim it and the business never has to argue about it. Fourth, price the failures at the current rate and at a bad rate, and confirm that the business can pay the credits in a bad month without changing anything else. Fifth, write the promise in plain words, with its exclusions named in the same paragraph, and have a lawyer read the wording before it is published. Deliverable: one page of wording, a measured failure rate and a costed remedy. Pitfall: writing the promise before measuring the failure rate, which turns a decision into a bet.

WATCH OUT The guarantee nobody can collect

A guarantee with conditions, a claim form and a deadline is a guarantee that costs the customer more than it pays them, and it is worse than no guarantee at all, because it advertises a promise the customer cannot reach. The pattern is familiar: the promise is printed in the brochure, the conditions are in a document nobody reads, the claim must be made within seven days in writing, and the remedy is a discount on a future purchase. The customer who tries to collect spends an hour and receives an apology. The qualities that prevent this are the ones the literature on service guarantees sets out: the strongest guarantees are unconditional, meaningful to the customer, easy to understand and to invoke, and easy to collect (Hart, 1988). Each of those four is a test that can be applied to a draft in ten minutes. If the guarantee has conditions, it is not unconditional. If the customer cannot say what it covers, it is not understandable. If invoking it requires a form, it is not easy to invoke. If collecting it requires a conversation, it is not easy to collect.

5.4 Service Levels and Service Credits

A service level is a promise with a number in it, and the number is what makes it usable. "A fast response" is a sentiment; "a technician on site within 90 minutes of an urgent call, day or night" is a service level, because a customer can tell whether it happened and a business can count how often it did not. The number is not decoration. It is the thing that turns a promise into an operation, because a promise with a threshold can be measured, and a promise that can be measured can be managed, reported and priced.

The threshold needs a record behind it, and the record has to be produced by the ordinary course of the work rather than by a special effort. Nordec's record is the arrival time of the technician, which the dispatch system captures because the dispatcher logs the call and the technician logs the arrival. Nothing extra is required to produce it, which is the test of a usable record: if measuring the service level requires a new process, the service level will be measured for three months and then quietly abandoned. The record also has to be the customer's record as well as the business's, which means the customer should be able to see the same number the business sees, in the monthly report or on request.

The remedy attached to a missed service level is a service credit, and its design follows the same logic as the guarantee's remedy. The credit should be a fixed amount rather than a proportion, because a fixed amount is easy to understand and easy to apply, and because a proportion invites an argument about the base. The credit should be applied automatically, from the record, without a claim. And the credit should be large enough to be felt: in the case, a credit of $150 against an average contract of $712 a month is about a fifth of a month, which a customer notices and a business can pay. The reasoning behind the choice is worth stating plainly, because it is the reasoning a reader will reuse. The credit is large enough to be felt and small enough to be paid without argument, and it is automatic because the dispatch system records the arrival time, so the customer never has to claim it.

Service levels are also where the tiers of Chapter 4 become visible to the customer. A tier is not only a list of what is included; it is a set of promises about how the service behaves, and those promises differ by tier in ways a customer can compare. The following table sets out Nordec's service levels by tier, with the six dimensions on which the three tiers differ. Read it as the promise side of the packaging designed in Chapter 4: the same building, the same technicians and the same vehicles, described at three levels of commitment.

Table 5.2 — Nordec's service levels by tier

Two features of that table are deliberate and worth naming. The first is that the urgent-call level and the two hours of included work are identical across all three tiers, because they are the promises the interviews showed customers value most, and a business should not weaken its strongest promise to create a difference between tiers. The second is that the differences between tiers sit in the slower, more administrative promises — the non-urgent request, the report, the note and the review — which is where a property manager's needs genuinely differ from an owner-operator's. A tier structure that differentiates on the promise nobody cares about is a price list; a tier structure that differentiates on the promises customers actually weigh is an offer.

The service level also has a reporting life of its own, and the reporting is where a small business can distinguish itself at almost no cost. A customer who receives a monthly report stating that the level was met in every case has a fact to take to their own client or their own owner. A customer who receives no report has a feeling. The one-page monthly report of Chapter 4 is the vehicle for this, and the fifth working day is the deadline that makes it a promise rather than an intention. The report is also the place where the business can be honest about a miss before the customer notices it, which is a much better position than being asked about it.

WHO DOES THIS The dispatcher counts the minutes

In a business of Nordec's size, the service level is measured by the dispatcher, and the measurement is a by-product of the work rather than a task added to it. The dispatcher answers the urgent call, records the time it came in, assigns the technician, and records the time the technician reports arrival on site. The difference between the two times is the response time, and it is written into the building's record as part of the same entry. Once a month, the dispatcher counts the calls and the misses for the one-page report, and once a quarter the count goes into the review described in Chapter 9. The role matters because it is the point at which the customer's experience becomes a number the business can act on. Before the guarantee, the dispatcher's record was an internal measure that nobody outside the business saw. After it, the same record is the customer's evidence, and the dispatcher knows that a minute recorded carelessly is a credit paid or a credit wrongly withheld.

TOOLS OF THE TRADE What a guarantee needs

Five functions carry a guarantee, and a business should look for the functions rather than for a product. The first is a record of arrival times, produced by the dispatch or field service system as part of the ordinary handling of a call, so that the measurement costs nothing extra. The second is a credit rule in the billing system, so that a missed level produces a fixed credit on the next invoice without a manual calculation. The third is a register of credits, read every week, which is the failure log: it shows how many credits were paid, to whom, for what and when, and it is the raw material of the improvement work. The fourth is the one-page wording of the promise, in the customer's language, with the exclusions named in the same document. The fifth is the monthly report that states whether the level was met, so that the customer sees the same number the business sees. No brand names belong on this list, and no prices; the requirement is the function.

5.5 Risk Reversal: Trials, Exits and Paying for Results

A guarantee is one instrument among several, and choosing the right one is easier when the options are laid out in order of how much risk the seller takes on. The forms of risk reversal run from the promise the law already imposes, through the warranty a seller writes, to a service level with credits, to a right to leave, and finally to payment for results, in which the seller is paid when the outcome is achieved rather than when the work is done. Each rung moves more of the buyer's risk to the seller, and each rung costs the seller more when the outcome is not achieved.

The first two rungs are the promises a business inherits or writes. The legal warranty is the floor: it exists whether or not anybody mentions it, and it cannot be removed. The conventional warranty is the seller's own document, and it is a modest form of risk reversal because it usually covers defects in what was supplied rather than the outcome the buyer wanted. A business that stops at these two rungs has not reversed much risk at all, because the buyer's four risks — that it does not work, that it takes too long, that it costs more than said, that it cannot be undone — are largely untouched by a warranty on materials.

The third rung is the service level with credits, which is where a service business begins to carry real risk. The promise is measurable, the remedy is fixed, and the business pays when it fails. The fourth rung is the right to leave: a period during which the customer may end the arrangement without penalty, which reverses the risk of having chosen wrongly. The fifth rung is payment for results, in which the seller is paid on the outcome rather than on the effort, and it is the rung that carries the most risk because the seller's revenue depends on a result the seller does not entirely control.

The right to leave deserves more attention than it usually receives, because it is the cheapest form of risk reversal for a business that is confident in its work. A customer who may leave within a stated period is not a customer who will leave; the exit is valuable because it removes the fear of being trapped, and the fear is what stops the signature. The cost of the exit is the work already performed for a customer who uses it, and that cost is knowable in advance. In the case, the onboarding at risk is $324, which is six hours of a technician's time at a loaded cost of $54 an hour, and the arithmetic of the exit is the arithmetic of how often a customer uses it.

The case's first-90-days guarantee is exactly this instrument. A new customer may cancel within the first 90 days with no penalty and keeps the building record, which means the business gives away the record it created as well as the time it spent. The record is the right thing to give away, because it is worth more to the customer than to the business and because it is the artifact that makes the relationship feel like a purchase rather than a trap. The cost of the exit is small if few customers use it: if one new customer in fifteen used it, the cost would be about $21.60 per signing, or about $238 a year at eleven signings. A business that can state that number has made the exit a decision rather than a gesture.

Chapter 7 adds a third instrument to the case, and it is worth naming here because it belongs to the same family. The Building Check, a productized entry offer at a fixed price of $950, credits the full amount against the first three months if a contract is signed within 60 days. That is a form of risk reversal aimed at the decision to try, rather than at the decision to stay: the buyer's risk is that the assessment will be money spent with nothing to show, and the credit removes it by making the assessment free to a buyer who continues. The instrument is a credit rather than a refund, which keeps the cash in the business while removing the buyer's fear.

The following figure places the five forms on a single ladder. Figure 5.1 shows a ladder of five rungs rising from left to right, beginning with the legal warranty at the bottom and ending with payment for results at the top, with the share of the risk carried by the seller growing at each rung. Nordec's two guarantees are marked on the ladder: the 90-minute credit sits on the third rung, beside the service level with credits, and the first-90-days exit sits on the fourth rung, beside the right to leave.

Figure 5.1 — The risk-reversal ladder. Each rung moves more of the buyer's risk to the seller; Nordec's 90-minute credit sits on the third rung and its first-90-days exit on the fourth.

Choosing a rung is a matter of matching the instrument to the risk the buyer feels, and the matching is not obvious. A buyer whose fear is that the work will not be done at all is not helped by a discount; a buyer whose fear is that the price will grow is not helped by a right to leave. The interviews of Chapter 2 are the place where the fear is named, and the instrument follows from the naming. A business that offers a trial to a buyer who fears the cost, or a refund to a buyer who fears the delay, has spent money on a remedy that does not answer the question being asked.

WHAT CHANGED A warranty of good working order from October 5, 2026

In Québec, sections 38.1, 38.2 and 38.8 of the Consumer Protection Act, added by S.Q. 2023, c. 21, s. 3, came into force on October 5, 2026. New goods of listed kinds sold or leased to consumers carry a warranty of good working order whose duration is set by regulation, covering parts and labor, and the merchant must state that duration prominently near the advertised price. Among the listed kinds are ranges, refrigerators, washing machines, television sets, computers, cellular telephones, air conditioners and heat pumps, and other goods set by regulation. The change matters to any reader who sells those goods to consumers, because a warranty of good working order with a stated duration is a promise the seller must both give and advertise, and the advertising requirement is the part most easily missed. It does not bind Nordec, which sells services to businesses and sells no goods, and it does not change the contract between two businesses. A reader who sells appliances, air conditioners or heat pumps to consumers should read the sections and take advice before the date.

5.6 The Cost of a Guarantee

A guarantee costs what the business fails to deliver, and that sentence is the whole of the arithmetic. The cost is not the value of the promise; it is the number of times the promise is broken, multiplied by the remedy. A business that knows its failure rate can therefore know the cost of its guarantee before it publishes it, and a business that does not know its failure rate is writing a promise whose price it will discover later. The measurement is usually available already, in the records the work produces, and the work of pricing a guarantee is mostly the work of counting.

The counting has two parts, and both are necessary. The first is the current rate: how often does the business fail the promise today, with the people, the vehicles and the systems it has? The second is the bad rate: how often would it fail in a month when two technicians are sick, a vehicle is in the shop and the weather produces a week of urgent calls? A guarantee priced only at the current rate is priced at the average, and averages are what a business experiences over a year rather than in a month. The bad rate is the number that decides whether the guarantee is safe, because the guarantee is paid in the bad month.

The case gives both numbers, and they are worth following through because the method transfers to any promise. Nordec's contract carries a 90-minute response commitment on urgent calls, day and night, and the last twelve months produced 253 urgent contract calls with 243 on site within 90 minutes, which is 96%. At that rate, about 10 calls a year would miss the level and the credits would cost $1,500, or 0.38% of the contract revenue of $393,024. At the 88% the redesign of Chapter 3 targets for its first year, about 30 calls would miss and the credits would cost $4,500, or 1.15% of contract revenue. The two figures bracket the cost of the promise, and the difference between them is the price of a bad year.

The following table sets out those two cases side by side, with the share of calls on site within 90 minutes, the misses a year, the credits a year and the credits as a share of contract revenue. Read it as the pricing of the promise: the first row is what the guarantee costs at the performance the business delivers today, and the second is what it would cost at the performance the business expects while a redesign is bedding in.

Table 5.3 — What the 90-minute guarantee costs

The value side of the ledger is harder to count and should not be ignored. The first item is the failure data: every credit paid is a recorded failure with a date and a cause, and the register of credits is the most honest list of the business's weaknesses it will ever have. The second is the renewal protected: a customer who received a credit without asking for it has a reason to renew that has nothing to do with price, and in a business where the renewal rate is the whole of the growth, that reason is worth more than the credit. The third is the decision it forces: a business that has promised a response time cannot leave a dispatch rule unexamined, because the rule now has a price attached to it.

The danger is guaranteeing what the business does not control, and it is the most common way a guarantee becomes expensive. A business controls its own arrival, its own workmanship and its own scope, and it does not control the customer's equipment, the availability of a part from a supplier, the weather, or the customer's decision to approve an extra hour of work. A guarantee that reaches into the uncontrolled territory will be invoked for reasons the business cannot fix, and the credits will be paid without producing any improvement. The remedy is to draw the boundary explicitly and to say in the wording what is outside it, which is what Nordec did when it declined to guarantee the outcome of a repair beyond its scope, anything involving parts, and non-urgent requests after hours.

The boundary is also where a guarantee meets the exclusions of the contract designed in Chapter 4. Nordec's contract excludes parts, work beyond two hours per incident, roof and envelope, elevators, and anything requiring a licensed trade the company does not hold. A guarantee that promised a working repair would collide with those exclusions, because the repair may need a part the contract does not cover. A guarantee that promises arrival within 90 minutes does not collide with them, because arrival is entirely within the business's control. The choice of promise is therefore also a choice about how much of the contract's exclusions the guarantee has to fight.

DO THE MATH What the 90-minute guarantee costs

Illustrative figures; the case is fictional. Step 1 — The last twelve months produced 253 urgent contract calls. Step 2 — At 96% on site within 90 minutes, 4% are missed: 253 × 0.04 = 10.12, or about 10 misses a year. Step 3 — At a credit of $150, the year costs 10 × $150 = $1,500, which is $1,500 ÷ $393,024 = 0.38% of contract revenue. Step 4 — At the 88% the redesign targets, 12% are missed: 253 × 0.12 = 30.36, or about 30 misses, costing 30 × $150 = $4,500, which is 1.15% of contract revenue. Step 5 — The first-90-days exit costs the onboarding at risk, $324 of technician time, divided by the fifteen new customers it would take for one to use it: $324 ÷ 15 = $21.60 per signing, or about $238 a year at eleven signings.

KEY CONCEPT A guarantee costs what the business fails to deliver

The cost of a guarantee is not the size of the promise; it is the failure rate multiplied by the remedy. A business that promises a 90-minute response and misses it ten times a year pays ten credits, and a business that misses it thirty times pays thirty. The arithmetic is that simple, and it is the reason a guarantee should never be written before the failure rate is measured. The measurement is usually already available, because the work produces the record: arrival times, completion dates, invoice disputes, callbacks. Once the rate is known, the guarantee becomes a decision with a price, and the price can be compared with the value of the renewals it protects and the failure data it produces. The second half of the concept is that the rate is not fixed. A business that pays credits and reads the register is a business that can improve the rate, which lowers the cost of the promise over time. The guarantee is therefore both a cost and an instrument of improvement, and the register of credits is where the two meet.

5.7 The Rules That Apply to Guarantees

A guarantee is a public statement, and public statements about what a business will do are regulated. The rules that apply to a guarantee in Canada are not a single statute but a small set of them, and a business that writes a promise without knowing which ones apply is exposed in a way that has nothing to do with whether the promise is honest. The exposure is greatest for a business that sells to consumers, and it is real, though smaller, for a business that sells to other businesses.

The first rule is federal and applies to everyone. Under section 74.01(1)(b) of the Competition Act, a representation to the public of the performance, efficacy or length of life of a product that is not based on an adequate and proper test is reviewable conduct, and the burden of proving the test rests on the person who makes the representation. Under section 74.01(1)(c), a representation in the form of a warranty or guarantee that is materially misleading, or that has no reasonable prospect of being carried out, is reviewable conduct. Under section 52, it is a criminal offense to knowingly or recklessly make a representation to the public that is false or misleading in a material respect, by any means, and the Crown does not have to prove that anyone was deceived. The provisions apply in every province and to sales to businesses as well as to consumers, which means a business-to-business seller such as Nordec is inside their reach.

The practical consequence of section 74.01(1)(b) is that a performance claim needs a test behind it. A business that says its service reduces unplanned interventions by half should be able to show the measurement that supports the statement, and the burden of showing it is on the business. The practical consequence of section 74.01(1)(c) is that a guarantee must be one the business can actually keep. A guarantee that is written for its marketing effect, with no plan for paying the remedy, is not merely optimistic; it is the kind of representation the provision addresses. The two provisions together set the standard this chapter has been applying throughout: measure the failure rate, and promise only what the business can deliver.

The second set of rules is provincial and applies to consumer offers. In Québec, the Consumer Protection Act governs contracts between a merchant and a consumer, and it contains several provisions a guarantee must respect. An exclusion from a conventional warranty must be set out clearly, in separate successive clauses, under section 44. A consumer may sue the merchant or the manufacturer for a latent defect under section 53, and neither may plead ignorance of it. Under section 54, a consumer may sue the merchant or the manufacturer directly on the obligations of sections 37 and 38, which are the legal warranty provisions described in section 5.2. And sections 38.1, 38.2 and 38.8, in force on October 5, 2026, add a warranty of good working order for new goods of listed kinds sold or leased to consumers, covering parts and labor, with the duration set by regulation and stated prominently near the advertised price.

The third set of rules governs contracts between businesses, and it is the one that applies to Nordec. In Québec, the Civil Code of Québec governs the contract. Under article 2100, a contractor or provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice, and, where applicable, ensure that the work or service conforms to the contract; when bound to an obligation of result, they may be relieved of liability only by proving superior force. Under article 2102, before the contract, the provider must give the client, as far as circumstances permit, useful information on the nature of the task and the property and time it requires. A guarantee written into such a contract is an addition to those obligations, and it should be drafted so that it does not contradict them.

The fourth set of rules is American and applies to a business that sells consumer products in the United States. The Magnuson-Moss Warranty—Federal Trade Commission Improvement Act defines a consumer product and a written warranty in section 2301, and sets the minimum standards of a "full" warranty in section 2304: remedy within a reasonable time and without charge, no limit on the duration of implied warranties, an exclusion of consequential damages only if conspicuous, and a refund or replacement after a reasonable number of failed attempts to repair. A warranty that does not meet those standards must be called "limited". The Commission's guide adds the practical requirements: a written warranty on a consumer product costing more than $10 must be designated "full" or "limited", and for products over $15 the terms must be set out in a single, clear document and made available to the consumer before the sale (Federal Trade Commission, n.d.).

The rules share a single idea, which is that a promise must be one the business can keep and must be stated in a way the customer can understand. That idea is the same one the design section of this chapter arrived at from the customer's side, and the convergence is not a coincidence: a guarantee that is easy to understand is also a guarantee that is easy to comply with, and a guarantee the business can keep is also a guarantee the customer can rely on. The rules do not require a business to make a guarantee. They require that a guarantee, once made, be true.

THE RULE A guarantee must be one the business can keep

Under section 74.01(1)(c) of the Competition Act, a representation to the public in the form of a warranty or guarantee that is materially misleading, or that has no reasonable prospect of being carried out, is reviewable conduct. The provision applies in every province and to sales to businesses as well as to consumers, so it reaches a business-to-business seller as well as a retailer. Read together with section 74.01(1)(b), which requires an adequate and proper test behind a claim about performance and places the burden of proving the test on the person who makes the claim, the rule sets a standard that is entirely practical: measure the failure rate before promising a remedy, and promise only what the business can deliver. A guarantee written for its effect in a proposal, with no plan for paying the credits, is the kind of representation the provision addresses. This book explains the law; it does not replace the advice of a lawyer.

5.8 The Case's 90-Minute Guarantee

Nordec's guarantee was not chosen because it sounded strong. It was chosen because the interviews of Chapter 2 put urgent response within 90 minutes in the top three for 18 of the 20 people interviewed, which made it the promise customers valued most among the things the business controls. The choice followed the method of section 5.3 in order: the promise came from the customers' words, the failure rate came from the dispatcher's records, the remedy was chosen to be felt and automatic, the cost was calculated at two rates, and the wording was written in plain words and read by a lawyer.

The record that made the guarantee possible already existed. The dispatcher logs every urgent call and the time it arrives, and the technician reports arrival on site, so the response time is a by-product of the ordinary handling of the call rather than a new measurement. The last twelve months produced 253 urgent contract calls, 243 of them on site within 90 minutes, which is 96%, and an average response of 64.4 minutes. The average is the number the business is proud of; the 96% is the number the guarantee is built on, because a guarantee is about the tail rather than the average, and the tail is where the customer's bad day lives.

The remedy is a credit of $150, applied automatically to the next invoice, and the wording is short enough to be read aloud. The wording says: "If our technician is not on site within 90 minutes of your urgent call, day or night, we credit $150 on your next invoice. You do not need to ask." The choice of a credit rather than a refund of the month is deliberate. The credit is large enough to be felt, at about a fifth of an average month's contract value, and small enough to be paid without argument. It is automatic because the dispatch system records the arrival time, so the customer never has to claim it, which is the quality the literature on service guarantees calls easy to invoke and easy to collect (Hart, 1988).

The guarantee also says what it does not cover, in the same document and the same voice. Nordec did not guarantee the outcome of a repair beyond its scope, because the outcome depends on the equipment and on the customer's decisions about replacement. It did not guarantee anything involving parts, because the contract excludes parts. And it did not guarantee non-urgent requests after hours, because the contract books those for the next working morning. The exclusions are not a retreat from the promise; they are what makes the promise keepable, and a customer who reads them knows exactly what they have bought.

The lawyer's reading was a small, fixed piece of work. A lawyer read the wording for three hours at a fixed fee, which is the arrangement described in Chapter 10, and the reading produced two changes: a sentence naming the exclusions in the same paragraph as the promise, and a sentence confirming that the credit is applied without any requirement on the customer. The cost of the reading is trivial against the cost of a wording that promises more than the business intends, and the reading is the point at which a marketing sentence becomes a contractual one.

The second guarantee is aimed at a different risk. A new customer's fear is not that the response will be slow; it is that they have chosen wrongly and will be trapped for a year. The first-90-days guarantee answers that fear directly: a new customer may cancel within the first 90 days with no penalty and keeps the building record. The onboarding at risk is $324, which is six hours of a technician's time at a loaded cost of $54 an hour, and the record is given away with it because the record is worth more to the customer than to the business. If one new customer in fifteen used the exit, it would cost about $21.60 per signing, or about $238 a year at eleven signings, which is a price a business can state out loud.

The cost of the 90-minute guarantee is known at both rates. At 96%, about 10 misses a year cost $1,500, which is 0.38% of contract revenue. At the 88% the redesign of Chapter 3 targets for its first year, about 30 misses would cost $4,500, which is 1.15% of contract revenue. The second figure is the one that matters, because it is the figure the business would pay in the year the promise is hardest to keep, and a business that can pay 1.15% of contract revenue in credits without changing anything else has priced its promise correctly.

What the guarantee changed inside Nordec is worth more than the credits. Before it, the dispatcher's record of arrival times was an internal measure that nobody outside the business saw, and the 96% was a number the business used to describe itself. After it, the same record is the customer's evidence, and the dispatcher knows that a minute recorded carelessly is a credit paid or a credit wrongly withheld. The measure became the customer's measure, which is the change a guarantee is supposed to produce, and the register of credits became the most honest list of the business's weaknesses it has.

CASE IN POINT Nordec Entretien inc.: a promise with a price on it

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. Sophie Nadeau chose the 90-minute guarantee because 18 of the 20 people interviewed named urgent response in their top three, and because the dispatcher's records already measured it: 253 urgent calls in twelve months, 243 on site within 90 minutes, an average of 64.4 minutes. The remedy is a $150 credit applied automatically to the next invoice, and the wording is one sentence long. At 96%, the guarantee costs about $1,500 a year, or 0.38% of contract revenue; at the 88% the redesign targets, it would cost about $4,500, or 1.15%. The first-90-days exit for new customers risks $324 of onboarding, about $21.60 per signing if one customer in fifteen used it. A lawyer read the wording for three hours at a fixed fee. The guarantee did not make Nordec's technicians faster. It made the dispatcher's record the customer's record, which is what made the promise real.

IN LA MACHINE Guarantees live outside the platform

The application copy read for this book has no guarantee or warranty register, no service-level tracking and no service credits, and no refund workflow (La Machine, 2026). A guarantee therefore lives outside the platform, in the contract, in the dispatch records and in the billing system, and the platform's role is narrower than a reader might expect. Two pages are still useful. At Business → Subscription Packages (business/subscription-packages), a package that is disabled lets existing subscribers keep their access until the end of their current billing period, which is the platform's own version of a right to leave: a customer who stops paying is not cut off mid-period. And the wording of a guarantee can be kept in a package description at Business → Session Packages (productivity/packages, also reached at business/packages), where each package carries a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image. That is a place to store the wording, not a way to enforce it. The counting of misses, the register of credits and the monthly statement that the level was met are all done outside the platform.

Costly Mistakes

• Writing the guarantee before measuring the failure rate. A promise priced at nothing is a bet, and the business discovers the odds in the month when the failures arrive together rather than one at a time.
• Attaching conditions the customer cannot see in advance. A guarantee that applies only if the call came before noon, or only if the equipment was under contract, is worth nothing to a buyer who cannot tell whether their situation qualifies.
• Making the customer claim the remedy. A credit that requires a form, a deadline or a conversation is collected by the customers with the most time and the least goodwill, and it advertises a promise most buyers never reach.
• Guaranteeing what the business does not control. A promise about the outcome of a repair, the availability of a part or the customer's own decisions will be invoked for reasons the business cannot fix, and the credits will be paid without producing any improvement.
• Setting a remedy too small to be felt. A discount on a future purchase, or a credit of a few dollars, costs the business the price of the promise and returns nothing, because the customer notices the failure and not the remedy.
• Publishing a performance claim with no test behind it. Under section 74.01(1)(b) of the Competition Act, a representation about performance that is not based on an adequate and proper test is reviewable conduct, and the burden of proving the test rests on the business that made the claim.
• Leaving the exclusions out of the wording. A guarantee that is silent about its limits will be read as covering everything, and the first customer who discovers otherwise feels misled, which is a worse outcome than a narrower promise honestly stated.

Key Takeaways

• A service cannot be inspected before it is bought, and the buyer's four risks — that it does not work, that it takes too long, that it costs more than said, that it cannot be undone — sit with the party that controls the outcome only if the seller moves them there.
• Products such as cars, cameras and toaster ovens come with warranties while services such as auto repair, wedding photography or catered dinners almost never do, although a service cannot be inspected before it is bought (Hart, 1988), and the gap is the opportunity a guarantee fills.
• The strongest guarantees are unconditional, meaningful to the customer, easy to understand and to invoke, and easy to collect (Hart, 1988). A guarantee that misses any of the four, in this book's reading, costs the business money while producing little of the benefit.
• A service level is a promise with a threshold and a record behind it, and a service credit is the fixed remedy applied automatically when the threshold is missed; Nordec's urgent-call level is identical across all three tiers because it is the promise customers value most.
• A guarantee costs what the business fails to deliver: at 96% of 253 urgent calls, about 10 misses cost $1,500, or 0.38% of contract revenue, and at 88%, about 30 misses cost $4,500, or 1.15%.
• Under section 74.01(1)(c) of the Competition Act, a guarantee that is materially misleading or that has no reasonable prospect of being carried out is reviewable conduct, and the rule applies to sales to businesses as well as to consumers.

Your Action Plan

1. Name the promise your customers value most, in their words. Take the interviews or the conversations you already have and write down the one promise that appears most often among the things you control. Deliverable: one sentence naming the promise. Verification: the sentence uses the customer's words rather than an internal process description.
2. Measure how often you fail that promise today. Count the failures from the records the work already produces — arrival times, completion dates, invoice disputes, callbacks — over the last twelve months. Deliverable: a failure rate and the number of failures a year. Verification: the count comes from a record, not from an estimate, and the record is one the customer could see.
3. Choose a remedy that is felt and make it automatic. Set a fixed amount that is large enough to be noticed and small enough to be paid without argument, and arrange for it to appear on the next invoice without a claim. Deliverable: a remedy and the rule that applies it. Verification: no customer action is required to receive it.
4. Price the guarantee at the current rate and at a bad rate. Multiply the failures at today's performance and at the performance of a bad year by the remedy, and express both as a share of revenue. Deliverable: two costed scenarios. Verification: the bad-year figure can be paid without changing anything else in the business.
5. Write the wording in plain words and have a lawyer read it. Write the promise and its exclusions in one short paragraph, in the customer's language, and have a lawyer read it before it is published. Deliverable: one page of wording and a lawyer's reading. Verification: the wording names the exclusions in the same paragraph as the promise, and the reading is recorded with its date.

In La Machine

Two routes in the application copy read for this book touch this chapter, and both touch it at the edges (La Machine, 2026). At Business → Subscription Packages (business/subscription-packages), packages are billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown, a description and an image; a package that is disabled lets existing subscribers keep their access until the end of their current billing period. That behavior is the platform's own version of a right to leave, and it is worth understanding as an example rather than as a feature to configure: a customer who stops paying is not cut off in the middle of a period they have already paid for. At Business → Session Packages (productivity/packages, also reached at business/packages), each package carries a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image, and the page shows the profit and the margin per sale as the package is edited. A guarantee's wording can be kept in a package description there, which is a place to store the sentence and not a way to enforce it.

What the platform does not do belongs in the same breath, and it is most of this chapter's subject. La Machine has no guarantee or warranty register, no service-level tracking and no service credits, and no refund workflow. It has no free trial. The counting of misses, the register of credits read every week, the automatic credit on the next invoice and the monthly statement that the level was met are all done outside the platform, in the dispatch records, the billing system and the contract. A reader who expects to configure a guarantee inside the platform will not find one, and a reader who keeps the wording in a package description and the counting in a spreadsheet has the whole instrument. Book N1 maps the modules the platform does cover, and Book K2 (CRM and ERP Systems) treats the systems that hold the customer record a guarantee is measured against.

Self-Check

1. What are the buyer's four risks, and why does it matter which party controls each of them?
2. What is the difference between a legal warranty, a conventional warranty, an additional warranty and a voluntary service guarantee?
3. What four qualities make a guarantee work, and what does a guarantee do inside the business beyond compensating the customer?
4. What is a service level, what is a service credit, and why is Nordec's urgent-call level the same across all three tiers?
5. What does the 90-minute guarantee cost at 96% and at 88%, and what does the first-90-days exit cost per signing?

Answers. 1. The four risks are that the service does not work, that it takes too long, that it costs more than was said, and that it cannot be undone; the party that controls the outcome is the party that should carry the risk, and in each of the four cases the business controls the outcome while the customer carries the fear, which is the mismatch risk reversal corrects. 2. A legal warranty is imposed by law, applies automatically and cannot be excluded; a conventional warranty is given by a merchant or manufacturer in a document or in advertising and binds the person who gives it; an additional warranty is a separate purchase by which a merchant undertakes to bear all or part of the cost of repairing or replacing goods beyond the conventional warranty; a voluntary service guarantee is a promise a business makes about its own service, with a remedy attached, which the business designs and can withdraw. 3. The four qualities are that the guarantee be unconditional, meaningful to the customer, easy to understand and to invoke, and easy to collect (Hart, 1988); inside the business a guarantee also forces the business to define good service in the customer's terms and produces failure data, because every credit paid is a recorded failure with a date and a cause. 4. A service level is a measurable promise with a threshold, such as a technician on site within 90 minutes of an urgent call, and a service credit is the fixed amount the customer receives when the threshold is missed, applied automatically to the next invoice; the urgent-call level is the same across all three tiers because the interviews showed it is the promise customers value most, and a business should not weaken its strongest promise to create a difference between tiers. 5. At 96% of 253 urgent calls, about 10 misses cost $1,500, or 0.38% of contract revenue; at 88%, about 30 misses cost $4,500, or 1.15%; the first-90-days exit risks $324 of onboarding, which is about $21.60 per signing if one new customer in fifteen used it, or about $238 a year at eleven signings.

Chapter References

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

Federal Trade Commission. (n.d.). Businessperson's guide to federal warranty law. Retrieved September 22, 2026, from https://www.ftc.gov/business-guidance/resources/businesspersons-guide-federal-warranty-law

Hart, C. W. L. (1988, July). The power of unconditional service guarantees. Harvard Business Review. https://hbr.org/1988/07/the-power-of-unconditional-service-guarantees

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages and business/subscription-packages.

Magnuson-Moss Warranty—Federal Trade Commission Improvement Act, 15 U.S.C. §§ 2301 and 2304 (United States). Legal Information Institute, Cornell Law School — read September 22, 2026.

Office de la protection du consommateur. (n.d.). Legal warranties. Gouvernement du Québec. Retrieved September 22, 2026, from https://www.opc.gouv.qc.ca/en/consumer/good-service/goods/electronic/warranty/legal-warranties

Chapter 6 — Value-Added Services and Complementary Offers

A service earns its place in an offer only when the target customer values it more than it costs to deliver and it fits the strategy, and every service that passes that test still has to be placed — included in a tier, sold on its own, or declined out loud.

What You'll Be Able to Do

• Distinguish a value-added service from an addition that merely lengthens the list, by asking whether it changes the outcome or the effort for the customer.
• Separate the core service, the supplementary services built around it and the complementary offers another firm provides, and say plainly which of the three the business is promising.
• Apply the three tests — value to the target segment, cost to deliver per customer per year, and fit with the strategy — to any service a customer, a technician or a colleague proposes.
• Decide whether a service is included in a tier, sold separately as an add-on, or declined, and record the decision with its date so that the next request meets an answer.
• Design a complementary offer with a partner, including a referral list given without a fee, and state what the business does not promise on the partner's behalf.
• Make an added service visible through a delivery ritual, a dated and signed document and a summary at renewal, so that the customer can see what the business actually did.

Why It Matters

Every business that has been running for a few years has a set of services it delivers and does not charge for. They arrived one at a time, each of them reasonable on the day it was granted, and none of them was ever placed anywhere: not in a tier, not on a price list, not in a refusal. They simply became part of what the business does, and the business absorbed the cost of them in the margin, quietly, month after month. The running case of this book reached that point at its first annual refresh, with four such services and a figure attached to them: $58.50 a month per building, $32,292 a year across forty-six buildings, against a price increase of $21 a month that brought in $11,592. The difference, $20,700 a year, was delivered to customers and never priced.

The reason this matters is not that the business was generous. It is that the generosity was invisible to the customer and expensive to the business at the same time, which is the worst combination available. A customer who receives a capital planning note, a monthly report and an up-to-date building record does not experience them as gifts; the customer experiences them as the contract, and the contract is what the customer compares with the alternative when the renewal arrives. A business that gives a service away therefore buys nothing with it. It does not win the renewal more cheaply, because the customer never knew the service was optional; it does not earn the right to raise the price, because the price was never attached to the service; and it does not learn whether the service was worth delivering, because nobody ever asked whether it was worth paying for.

The remedy is a discipline rather than a budget cut, and it is the subject of this chapter. A service is added deliberately, tested against three questions, and then placed in one of three ways: included in a tier because it belongs to the promise the customer is buying, sold separately because some customers want it and others do not, or declined out loud because it fails a test. The third outcome is the one small businesses reach least often and need most, because a refusal that is written down stops the same request from being reconsidered from the beginning every time it arrives. The chapter closes on the case's own service menu, where eight candidates were judged in a single session and five were placed, and on the two things that made the placement real: a referral list given to owners without a fee, and a capital planning note that is delivered in a meeting rather than sent as a file.

In This Chapter

Section 6.1 sets out what makes a service value-added, and why a longer list is not a stronger offer. Section 6.2 separates the core service, the supplementary services around it and the complementary offers another firm provides. Section 6.3 gives the three tests — value to the target segment, cost to deliver per customer per year, and fit with the strategy — and places the case's eight candidates on a four-area chart. Section 6.4 turns the tests into three decisions: include, sell separately or decline. Section 6.5 covers complementary offers and partners, from the referral list given without a fee to the promises a business must never make on a partner's behalf. Section 6.6 examines the hidden cost of free services through the case's four unpriced additions. Section 6.7 shows how to make an added service visible to the customer who receives it. Section 6.8 describes the session in which the case's service menu was decided.

6.1 What Makes a Service Value-Added

The phrase "value-added" is used so loosely in business conversation that it has almost stopped meaning anything. In this book it means one specific thing, and the specificity is what makes it usable. A service is value-added when it changes the outcome the customer receives or the effort the customer has to spend, in a way the customer recognizes. A service that does neither is not value-added, however pleasant it is, however much work it took, and however good it makes the business feel about itself. The test is not whether the business did something. The test is whether something changed on the customer's side.

That distinction is easy to state and hard to apply, because the two kinds of addition look identical from inside the business. A technician who spends twenty minutes writing a clear summary of a visit has done work that a customer can read, file and use when a tenant asks what happened. A technician who spends twenty minutes filling in an internal form has done work that no customer will ever see. Both take twenty minutes and both cost the same. Only one of them changes anything for the customer, and the business that cannot tell them apart will eventually pay for both and be thanked for neither.

Theodore Levitt made the underlying argument in a way that has worn well. There is no such thing as a commodity, Levitt argued; all goods and services can be differentiated (Levitt, 1980). The article goes further in a direction that matters for a business selling to other businesses: industrial products and services offer more room for differentiation than consumer goods do, because the buyer is an organization with a process, a budget and a set of people who have to live with the purchase. That is precisely the terrain of the case followed through this book. A monthly preventive visit is a service any mechanical contractor can perform. What surrounds it — a record of the building, a note that tells an owner what will need replacing and roughly what it will cost, a report that a property manager can forward to the owners they answer to — is where the offer becomes difficult to compare and therefore difficult to replace.

The second half of the argument belongs to James Anderson and James Narus, who looked at the services around a manufactured product and found them neglected. Manufacturers had invested in flexible production and modular products, the authors observed, and had overlooked the services that surround the core product, even though those services shape differentiation, costs and profitability. Their recommendation is the sentence this chapter is built on: supplementary services should be designed as deliberately as the product, and together with it, so that they meet the needs of different customer segments while cost and price stay under control (Anderson & Narus, 1995). Designed deliberately means three things at once. It means the service has a defined content, so that two customers receive the same thing. It means the service has a cost the business has counted. And it means the service has a place in the offer, which is either inside a tier or beside it as something a customer can buy.

The case shows what happens when the third element is missing. Over two years, Nordec added four services to its contract without placing any of them: the upkeep of the building health record, the one-page monthly report, the capital planning note delivered at renewal, and the mid-contract note for buildings whose equipment the first note had flagged. Each addition was small when it was made, and each was defensible on its own. The record's upkeep cost $13.50 a month per building. The report cost $27.00. The note cost $13.50. The mid-contract note cost $4.50. Together they came to $58.50 a month per building, which is $32,292 a year across forty-six buildings, against a price that had risen by $21 a month and brought in $11,592. The services were real, the customers valued them, and the business paid for them out of its own margin without ever deciding to.

DEFINITION Core, supplementary and complementary services

The core service is what the customer is buying: the thing the contract is about, the reason the customer called in the first place. For a maintenance contractor, the core is the scheduled preventive visit and the response commitment that goes with it. A supplementary service is a service the business itself delivers around the core, designed and priced together with it: a record of the building's equipment, a monthly report, a planning note, a review meeting. Supplementary services are the business's own work, and the business answers for them. A complementary offer is something the customer needs that another firm provides better, and that the business points to rather than performs: a licensed trade it does not employ, a replacement project, an engineering assessment. A value-added service is any of the three that changes the outcome or the effort for the customer in a way the customer recognizes. The distinction between supplementary and complementary is not about importance; it is about who performs the work and who carries the promise.

The practical consequence of the definition is that a business should be able to say, for every service it delivers, which of the three it is and where it sits in the offer. A service that cannot be classified is a service that has not been designed, and a service that has not been designed cannot be costed, cannot be promised consistently and cannot be defended at a renewal. The next section takes the three categories apart and shows why the boundary between the second and the third is the one small businesses cross most often without noticing.

6.2 Core, Supplementary and Complementary

The core service is the part of the offer that a customer could describe to somebody else in one sentence. For the case, that sentence is a monthly preventive visit to a commercial building's mechanical, plumbing and door systems, with a ninety-minute response commitment on urgent calls, day and night. Everything the customer is buying is in that sentence, and everything else in the contract is either a supplementary service the business delivers or a complementary offer it points to. The distinction sounds academic and is not, because the three categories carry different costs, different promises and different risks, and a business that mixes them will eventually promise something it cannot deliver.

Supplementary services are the business's own work, performed by its own people, under its own name. They are the reason two contractors quoting the same core service are not interchangeable. Anderson and Narus describe the design problem precisely: supplementary services should be designed as deliberately as the product and together with it, so that they serve different customer segments while cost and price remain controlled (Anderson & Narus, 1995). The phrase "different customer segments" is doing real work in that sentence. A property manager who answers to a set of building owners wants a consolidated report across the buildings in their portfolio; an owner-operator who occupies a single building wants a note telling them what will need replacing and roughly when. Both are supplementary services, both are delivered by the same business, and they are not the same service. Designing them together with the core means deciding, segment by segment, which of them belongs inside the promise and which is offered beside it.

The manufacturer's version of the same question was set out by Richard Wise and Peter Baumgartner, who observed that many large manufacturers saw little profit growth through the 1990s despite real productivity gains, and argued that the profit lies downstream, in the services around the product (Wise & Baumgartner, 1999). The article is about manufacturers and the logic transfers without modification to a service business. A contractor that performs the same visit for the same price as its competitors has commoditized its own core, and the room left for it to be chosen lies in what it does around the visit. The case's capital planning note is exactly a downstream service in that sense: it is not the maintenance work, it is the information the maintenance work produces, organized so that an owner can budget. It costs the business three hours a year per building and it is the single service that seven of eight owner customers named among their top three.

At the far end of the same road lies the experience, which B. Joseph Pine II and James Gilmore described as a fourth kind of economic offering after commodities, goods and services. An experience occurs, they wrote, when a company uses services as the stage and goods as props to engage a person in a memorable event, and they traced the idea through a birthday cake: the ingredients bought as commodities, the cake mix as a good, the bakery cake as a service, and the party staged by a business as an experience (Pine & Gilmore, 1998). A maintenance contractor does not stage parties, and the article is not an instruction to try. It is a way of seeing where the addition of services leads when it is taken seriously: the customer is not only buying a working building, but a relationship with the people who keep it working, and the moments in that relationship — the call answered at night, the note handed over at renewal, the technician who knows the building — are what the customer remembers and describes to others.

Complementary offers are the third category and the one most often confused with the second. A complementary offer is something the customer needs that another firm provides better, and the business's role is to point to it rather than to perform it. For the case, replacement projects are the clearest example: when a building's equipment reaches the end of its life, the owner needs a licensed contractor to replace it, and Nordec does not do projects. The strategy signed a year earlier had closed that wider path unless at least five owners asked for replacement projects after reading their planning note; four had asked, and the path stayed closed. What the business does instead is described in section 6.5: it keeps a list of three licensed contractors and gives the owner the list together with the note, without a referral fee.

THE RULE Supplementary services are designed as deliberately as the core

The rule this chapter applies comes from the study of supplementary services in business markets (Anderson & Narus, 1995): supplementary services should be designed as deliberately as the product, and together with it, so that they meet the needs of different customer segments while cost and price remain under control. The rule has four parts, and a business that drops any one of them loses the benefit. Designed means the service has a defined content that two customers receive in the same way. As deliberately as the core means it receives the same attention as the thing the contract is about, rather than being improvised by whoever is on site. Together with it means the service and the core are designed as one offer, so that the service is not an afterthought bolted on at a renewal. With cost and price controlled means the business has counted what the service costs to deliver and has decided where that cost is recovered — inside a tier, in an add-on, or not at all, as a deliberate choice rather than a drift.

The boundary between supplementary and complementary is where small businesses get into trouble, and the trouble has a recognizable shape. A customer asks whether the business can also do something adjacent — replace a unit, handle a trade the business does not employ, take on a building outside its territory. The business says yes because the customer is standing there and the answer is easy. The work is then subcontracted, and the business is now answerable for a promise it does not control, delivered by a firm it does not manage, under a contract it did not write. The rule that prevents this is simple and belongs in the offer itself: the business performs what it has designed, and it points to what it has not. A pointed-to service is still a service to the customer, and it costs the business nothing but the honesty of saying who does it.

6.3 Which Services to Add

Every business receives more ideas for new services than it can deliver, and the ideas arrive from three directions at once. Customers ask for things. Technicians propose things, because they see the same problem three times in a month and know what would fix it. Owners and managers read about something and wonder whether it applies. The question this section answers is how to sort those ideas without either refusing everything or accepting everything, and the answer is three tests applied in a fixed order.

The first test is value to the target segment, and it is answered with a count rather than an opinion. The interviews described in Chapter 2 are the source: twenty people were asked what mattered to them, and the elements they named among their top three were recorded. Urgent response within ninety minutes was named by eighteen of twenty. The monthly preventive visit was named by fourteen. The capital planning note was named by twelve — seven of the eight owner customers, two of the six property managers, and three of the six prospects who had declined a proposal after seeing a sample. The monthly report was named by six: one owner and five managers, and no prospects. The building health record was named by four. Those counts are the first test, and their shape is the lesson. The same service can be highly valued by one segment and nearly invisible to another, and a business that adds a service because "customers want it" without knowing which customers has not passed the first test; it has skipped it.

The second test is cost to deliver per customer per year, counted in hours as well as in money. The case's figures are the model. The building health record's upkeep costs a quarter-hour a month, which at a loaded technician cost of $54 an hour is $13.50 a month, or $162 a year. The one-page monthly report costs twenty minutes of a technician and fifteen minutes of the administrator, which is $27.00 a month, or $324 a year. The capital planning note costs three hours a year, or $162. The mid-contract note costs two hours a year for about half the buildings, or $162 for the buildings that receive it. The quarterly review costs $324 a year. The tenant request line costs $216. The consolidated report costs $108. Counting the cost is not an accounting exercise performed for its own sake; it is what makes the third decision possible, because a service whose cost is unknown cannot be placed anywhere.

The third test is fit with the strategy, and it is the test that does the most work in a small business because it is the one that produces refusals. The strategy signed a year before this book begins made three refusals: one-off emergency calls for buildings without a contract, new buildings outside the two territories, and bundled tenders that combine cleaning, security and maintenance. It also created three things and raised them: the capital planning note, the building health record and the one-page monthly report. A candidate service fits when it strengthens one of those choices and does not undo one of those refusals. An energy-efficiency check fails the third test not because it is a bad service but because it belongs to the wider path of retrofit projects that the strategy closed. Non-urgent after-hours dispatch fails it because it undoes the reduce the strategy deliberately made: the business chose to book non-urgent requests for the next working morning, and a service that sends a technician at night for a non-urgent request takes that choice back. A filter and consumables subscription fails it because parts are excluded from the contract by design.

Figure 6.1 places the case's eight candidates on a single chart, and the chart is the fastest way to see how the three tests work together. The figure is a square divided into four areas by two axes. The horizontal axis runs from left to right and shows how much the target segment values the service, measured by the interview counts of Chapter 2. The vertical axis runs from bottom to top and shows how well the service fits the strategy at an acceptable cost. The upper-right area is labeled "include", the upper-left area is labeled "sell separately", and the lower half is labeled "decline". The eight services appear as dots with their names beside them: the capital planning note, the mid-contract note, the quarterly review, the tenant request line and the consolidated report sit in the upper-right area; the capital planning note as an add-on for the smallest tier sits toward the upper-left; and the energy-efficiency check, the non-urgent after-hours dispatch and the filter and consumables subscription sit in the lower half.

Figure 6.1 — Include, sell separately or decline. Nordec's eight candidate services placed by how much the target segment values them and how well they fit the strategy at an acceptable cost.

METHOD The three tests for a new service

First, name the segment that asked and count how many asked. "A customer asked" is not a count; "three of six property managers asked" is. If the request came from one person, say so, and treat it as one person's request until a second and third arrive. Second, cost the service per customer per year, hours included, using the loaded cost of the person who performs it and the number of times a year it is delivered. Third, check the service against the strategy's choices and refusals: does it strengthen something the strategy raised, and does it undo something the strategy refused? Fourth, decide: include it in a tier, sell it separately as an add-on, or decline it. Fifth, write the decision down with its date and the reason, so that the next request for the same service meets an answer rather than a fresh conversation. Deliverable: a dated decision on a counted request. Pitfall: deciding on the strength of the request rather than on the count.

The three tests are applied in order for a reason. The first test is cheap and eliminates most ideas, because most ideas come from one person. The second test is the one that changes minds, because a service that a segment values can still be unaffordable at the price the segment will pay for it. The third test is the one that protects the business from itself, because the most attractive new service is often the one that quietly reverses a choice the strategy made. A service that passes all three is a service the business can add with confidence, and a service that fails any one of them is not added, however good the argument for it sounds on the day.

6.4 Include, Sell Separately or Decline

The three tests produce a decision, and the decision has three possible answers. The service is included in a tier, which means it becomes part of the promise the customer buys and its cost is recovered inside the price of that tier. The service is sold separately, which means it is offered beside the tiers as an add-on with its own price, and customers who want it buy it. Or the service is declined, which means the business does not offer it, and the refusal is written down with its reason. Each answer is correct in different circumstances, and the most common mistake in small businesses is not choosing the wrong one; it is failing to choose at all, so that the service is delivered free and the decision is made by default.

Inclusion is the right answer when the service belongs to the promise and the segment that values it is the segment the tier is for. The capital planning note is the clearest case. Seven of the eight owner customers named it among their top three, and the note is the thing that makes the offer different from a competitor's visit. Putting it inside the Preventive tier means every customer on that tier receives it, the cost is recovered in the price, and the promise is consistent: a customer who buys Preventive knows a note is coming at renewal. Inclusion has a second benefit that is easy to overlook. A service inside a tier cannot be forgotten at renewal, because it is part of what the customer is buying and part of what the business is measured on.

Selling separately is the right answer when the service is valuable to some customers and irrelevant to others, or when it is valuable to a segment that buys a smaller tier. The capital planning note is again the example, in the opposite direction. The Essential tier exists for buildings under 3,500 square metres, and the note is not included in it, because the tier is built to a lower price and the note would not fit inside it. It is offered as an add-on at $420 a year, with a cost of $162 and a margin of 61.43%. A customer on Essential who wants the note buys it; a customer who does not, does not pay for it. The quarterly review works the same way: it is included in the Portfolio tier, where the segment that wants it buys, and it is offered as an add-on to Preventive at $140 a quarter, with a cost of $81 and a margin of 42.14%.

Declining is the right answer when the service fails one of the three tests, and it is the answer that requires the most discipline because it is the only one that produces no revenue. Three of the case's eight candidates were declined. The energy-efficiency check was asked for by two owners and declined because it belongs to the wider path of retrofit projects that the strategy closed; offering it would have reopened a path the business had decided not to walk. Non-urgent after-hours dispatch was declined because it undoes the reduce the strategy made, and because the two owners who wanted it had already been lost on that point — offering it now would have meant rebuilding a service the business had deliberately removed. The filter and consumables subscription was declined because nobody asked for it and because parts are excluded from the contract by design. Each refusal was written down with its reason, which is what makes it a decision rather than a mood.

The following table sets out the case's eight candidate services with the evidence behind each one, the cost per building per year, whether it fits the strategy, and the decision taken. Read it as the record of a single session rather than as a menu: the columns are the three tests of section 6.3 in the order they were applied, and the last column is what the business committed to. The two services that appear twice in the decision column are the ones that were included in one tier and sold separately beside another.

Table 6.1 — Nordec's eight candidate services

WATCH OUT The request that becomes a promise

The dangerous request is not the one that arrives from a customer the business has never served. It is the one that arrives from a good customer, in a difficult week, and is granted as a favor. A property manager asks whether the business can send somebody at night for a request that is not urgent; the answer is yes, once, because the manager is a good customer and the request is small. Nothing is written down. Six months later the same manager asks again, and the answer is yes again, because the first yes is now the precedent. By the second year, three managers are asking, the business is staffing nights for non-urgent work, and the reduce the strategy made has been reversed without a decision. The remedy is not to refuse the first request. It is to record every exception in a register, with the date and the customer, and to decide within the quarter whether the exception becomes part of a tier, an add-on with a price, or a refusal that is communicated. An exception that is recorded is a candidate service; an exception that is not is a promise nobody chose to make.

The three decisions are not permanent, and the chapter would be misleading if it suggested otherwise. A service declined because nobody asked can be reconsidered when three customers ask. A service included in a tier can be moved to an add-on if the segment that values it turns out to be narrower than the interviews suggested. What makes the decisions useful is not their finality but their visibility: a business that has written down what it includes, what it sells and what it refuses can change any of the three deliberately, and a business that has written down nothing changes all three by accident.

6.5 Complementary Offers and Partners

A small business cannot perform everything its customers need, and the customers know it. The question is not whether to point them somewhere else; the question is how to do it in a way that serves the customer, protects the business and stays honest about who is doing the work. There are three shapes this takes, and they differ in what the business promises and what it receives.

The first shape is a referral without a fee. The business keeps a short list of firms it has seen do good work, gives the customer the list, and takes nothing for the introduction. The case does exactly this for replacement projects. When a building's equipment reaches the end of its life, the owner needs a licensed contractor to replace it, and Nordec does not do projects. The business keeps a list of three licensed contractors and gives the owner the list together with the capital planning note, without a referral fee. The list is short on purpose. A list of ten is a directory and helps nobody; a list of three is a recommendation, and a recommendation is what the owner actually needs when a note tells them a rooftop unit has two summers left. The absence of a fee is what makes the recommendation credible, and it is also what keeps the business out of the promise: Nordec is saying who it has seen do good work, not who will do this job well.

The second shape is a referral with a disclosed fee. The business introduces a customer to a partner, the partner pays for the introduction, and the customer is told that this is the arrangement. Disclosure is not a courtesy; it is what separates a recommendation from an advertisement. In Canada, the Competition Act makes a representation to the public that is false or misleading in a material respect a reviewable practice under section 74.01(1)(a), and the same conduct can be a criminal offense under section 52 when it is made knowingly or recklessly, with no need for the Crown to prove that anyone was deceived. A referral presented to a customer as an independent recommendation, while a fee is being paid for it, is a representation about the nature of the recommendation, and a customer who would have weighed the advice differently had they known is a customer who was misled in a material respect. The book explains the law and does not replace the advice of a lawyer; a business that intends to charge for referrals should have the arrangement reviewed and should disclose it in plain words.

The third shape is a partner bundle, where two firms sell a combined offer under an arrangement they have agreed. This is the most demanding of the three, because the customer experiences one offer and the two firms experience two contracts. A bundle works when the two services are genuinely complementary, when the customer can tell which firm is responsible for which part, and when each firm can deliver its half without depending on the other's schedule. It fails when the customer cannot tell who to call, which is the failure mode that makes bundles unpopular with customers even when they are cheaper. A small business considering a bundle should write down, before agreeing to it, what happens when one half is late, who the customer calls, and what the customer is told about the arrangement.

What a business must never do is promise on a partner's behalf. The promise the customer hears is the promise the business made, whatever the contract says, and a business that says "they will be there Tuesday" has made Tuesday its own commitment. The rule that follows is narrow and easy to apply: the business describes what the partner does, names the partner, and lets the partner make the partner's promises. A maintenance contractor can say that a licensed electrical contractor on its list handles service electrical work; it cannot say when that contractor will arrive, what the work will cost, or that the work will be done to a standard the contractor has not agreed to. The same rule applies to the documents the business hands over. The capital planning note lists equipment approaching the end of its life with a replacement window and a budget range; it does not quote a replacement price, because the business does not sell replacements and cannot know what a contractor will charge.

The strategic value of a referral list is worth stating plainly, because it is easy to read as a courtesy and it is not. A business that hands an owner a list of three contractors at the moment the owner needs one has done something no competitor has done: it has been useful at a moment when it had nothing to sell. That is the moment a customer remembers, and it is the moment that makes the next renewal conversation easier. It also keeps the business's own refusals intact. The wider path of replacement projects stays closed, the business does not take on work it cannot control, and the customer still gets what they need. Book F1 (Go-to-Market and Channel Strategy) treats the design of these arrangements from the channel side, and Book E7 (Account Management, Customer Success and Expansion) treats the account relationship in which a referral list becomes part of what a customer expects from a supplier.

QUÉBEC & CANADA NOTE Referral fees and what must be disclosed

In Canada, the Competition Act makes a representation to the public that is false or misleading in a material respect a reviewable practice under section 74.01(1)(a), and the same conduct can be a criminal offense under section 52 when it is made knowingly or recklessly; the Crown does not have to prove that anyone was deceived. A referral presented to a customer as an independent recommendation, while a fee is being paid for it, is a representation about the nature of that recommendation, and a customer who would have weighed the advice differently had they known the arrangement is a customer who was misled in a material respect. The provision applies in every province and to sales to businesses as well as to consumers, which matters here because the case's contracts are between two businesses. Nordec's referral list for replacement projects carries no fee, and the business says so when it hands the list over. This book explains the law and does not replace the advice of a lawyer; a business that intends to charge for referrals should have the arrangement reviewed and should disclose it plainly.

The three shapes are not ranked, and a business can use all three at once for different things. What they share is a requirement that the customer can tell who is doing what. A referral list with no fee, a disclosed paid introduction and a partner bundle are all honest arrangements; an undisclosed fee and a bundle whose responsibilities are unclear are not. The test a small business can apply without a lawyer is simple: if the customer knew exactly how this arrangement works, would they still take the recommendation the same way? If the answer is no, the arrangement needs to be disclosed or redesigned.

6.6 The Hidden Cost of Free Services

The four services the case added without pricing them are worth examining one at a time, because the pattern only becomes visible when the additions are laid side by side. The building health record's upkeep was added in the third quarter of the strategy year, at a cost of a quarter-hour a month, or $13.50 a month per building. The one-page monthly report was added in the same year, at twenty minutes of a technician and fifteen minutes of the administrator, or $27.00 a month. The capital planning note at renewal was added in the same year, at three hours a year, or $13.50 a month. The mid-contract note for buildings whose equipment the first note had flagged was decided at the refresh, at two hours a year for about half the buildings, or $4.50 a month. Each of the four was small. Each was defensible. None of them was placed.

The arithmetic is what makes the pattern undeniable. The four additions cost $58.50 a month per building, which is $32,292 a year across forty-six buildings. Over the same period the contract price rose from $691 to $712 a month, an increase of $21, which brought in $11,592 a year. The difference is $20,700 a year: value that was delivered to customers and never priced. The margin tells the same story from the other side. The contract margin of a building in a territory fell from 59.08% at $691, before the additions, to 52.08% at $712, after them. The business did not decide to accept a lower margin. It decided four times to add something small, and the margin moved on its own.

The effect on customers is the part that is harder to see and more expensive in the long run. A free service trains the customer to expect it, and an expectation is a price. Once a customer has received a monthly report for a year, the report is not an extra; it is part of what the contract is, and the customer will compare a renewal that removes it with a competitor who never offered it and conclude that the business has cut something. The same is true of the capital planning note. A service given away cannot later be sold to the customer who received it, and it cannot easily be sold to a new customer either, because the existing customers are receiving it free and the price list has to be consistent. Free services therefore do two things at once: they cost money now, and they remove a source of revenue later. This is the book's own observation about how an offer behaves, not a finding from a source, and it is the reason this chapter treats placement as part of the design rather than as a pricing question left to Book A7 (Pricing Strategy and Revenue Models).

The document that comes out of this examination is the one Book A7, Chapter 8, needs when the offer has improved. It is a list of what was added, when, what it cost to add, and where it went in the new offer. The list is not a complaint about the past; it is the evidence that the offer is better than it was, and it is what allows a price to be attached to the improvement without guesswork. A business that can say "we added four services over two years, they cost $58.50 a month per building, and here is where each of them now sits" is a business that can have a rational conversation about price. A business that cannot say that is a business that will either raise the price without justification or fail to raise it at all.

The following table sets out the four additions with the date each was made, its monthly cost per building, its annual cost across forty-six buildings, and where it was placed in the new offer. Read it as the bridge between this chapter and the pricing work: the first three columns are what the business spent, and the last column is what it decided to do about it. The total row is the figure that made the case for a service menu rather than a fifth free addition.

Table 6.2 — What the four additions cost before they were placed

KEY CONCEPT A free service is a price decision

A service delivered without a price is not outside the pricing system; it is inside it, with the price set at zero. That is why the phrase "we will give this one away" is misleading. Nothing is given away. The cost of the service is absorbed by the margin, which means it is paid for by the business, and the customer receives it at no charge, which means the customer learns that it is free. Both halves of that transaction have consequences. The business's margin falls without a decision being taken, and the customer's expectation rises without a promise being made. The two effects compound: the more the business gives, the more the customer expects, and the harder it becomes to attach a price to the service later, because the customer who has received it free for a year will not accept a new charge for it. The remedy is not to charge for everything. It is to place everything: a service inside a tier is priced, a service sold as an add-on is priced, and a service declined is not delivered. There is no fourth option in which the service is delivered and the price is not decided.

The lesson generalizes beyond the case, and it applies to businesses of every size. A solo consultant who answers questions between engagements is delivering a service without placing it. A small firm that includes a site visit in every quote is doing the same. The question to ask is not whether the service is worth delivering; it is where the service sits. If it belongs to the promise, it goes inside the price and the price reflects it. If it does not, it is sold beside the offer or it is declined. What a business cannot do is deliver it and leave the decision to the margin, because the margin will make the decision, and the margin will make it badly.

DO THE MATH The services that were given away

Illustrative figures; the case is fictional. Step 1 — The four additions cost, per building per month: the building health record's upkeep $13.50, the one-page monthly report $27.00, the capital planning note at renewal $13.50, and the mid-contract note $4.50. Step 2 — Added together, they come to $58.50 a month per building. Step 3 — Across forty-six buildings and twelve months, $58.50 × 12 × 46 = $32,292 a year of service delivered. Step 4 — Over the same period the contract price rose from $691 to $712 a month, an increase of $21, which across forty-six buildings and twelve months is $21 × 12 × 46 = $11,592 a year of additional revenue. Step 5 — The difference, $32,292 − $11,592 = $20,700 a year, is the value that was delivered and never priced, and it is the figure that moved the contract margin of a territory building from 59.08% to 52.08%.

6.7 Making Added Value Visible

A service the customer does not notice has no value to the customer, whatever it cost to deliver. This is not a marketing observation; it is a design requirement, and it follows directly from the definition in section 6.1. A value-added service changes the outcome or the effort for the customer in a way the customer recognizes. If the customer does not recognize it, the change did not happen on the customer's side, and the business has spent money on an internal process. The remedy is not advertising. It is a delivery ritual: a fixed way the service is handed over, repeated every time, so that the customer experiences it as part of the offer rather than as an occasional kindness.

The case's capital planning note is the model, and its design is worth setting out in full because every element of it is doing work. The note is delivered in a thirty-minute meeting at renewal, never simply emailed. The meeting is the ritual: it puts a person in front of the customer, it gives the note a moment, and it makes the note something that happened rather than something that arrived. The first page of the note lists the three pieces of equipment with the shortest replacement window and a budget range for each, so that the customer's eye lands on the three things that matter before anything else. Each note is dated and signed by the lead technician, which does two things at once: it makes the note a document rather than a file, and it puts a named person's judgment behind it. And a sample note, with the owner's name removed, goes into every proposal, so that a prospect can see what they will receive before they buy it.

Each of those four choices answers a specific failure. Emailing the note fails because a file in an inbox is not a conversation, and the customer who does not open it has received nothing. A note without a first page of three items fails because a list of twenty pieces of equipment is a document the customer will not read, and an unread note changes nothing. A note without a date and a signature fails because it looks like a printout, and a printout carries no one's judgment. A note that is never shown to prospects fails because the value of the note is invisible until it is received, and a prospect who cannot see it will compare the offer on price alone. The four choices together turn three hours of work a year into something the customer can describe to somebody else, which is the only test of visibility that matters.

The same reasoning applies to the other services in the menu. The one-page monthly report is one page because a longer report is not read, and it is sent after each visit rather than at the end of the month because the visit is what the customer remembers. The building health record is updated on site, in the closing note the technician writes before leaving, because a record updated a week later is a record that has lost the detail. The quarterly review meeting in the Portfolio tier exists partly to deliver the consolidated report in person, for the same reason the note is delivered in a meeting. The tenant request line is visible because the tenant receives an acknowledgment, which is what tells the property manager that the service is working.

The danger in all of this is the opposite failure, which is hiding services in a long list. A business that has built a genuinely strong offer can bury it under a description that nobody reads, and the customer then compares on the only thing that is easy to compare, which is price. The test is whether the customer can name, without looking, the two or three things the business does that a competitor does not. If the customer cannot, the services exist and the value does not. A short list, delivered through a ritual, with a document that carries a date and a name, is worth more than a long list delivered silently.

WHO DOES THIS The owner rules, the technicians propose

In a small service business, the two roles in this chapter are distinct and both are necessary. The owner rules: the owner applies the three tests, decides whether a service is included, sold separately or declined, and writes the decision down with its date. The owner is the only person who can weigh a service against the strategy, because the owner is the person who made the strategy's choices and refusals. The technicians propose: they are the people who see the same problem three times in a month, who know which part is always missing, and who hear what a customer actually complains about. A business that does not collect their proposals loses its best source of candidate services, and a business that lets them grant requests on site loses control of the offer. The division that works is that technicians bring the observation and the owner brings the decision, and the register of requests and exceptions is where the two meet. In the case, the lead technician proposed the van stock that cut second visits, and the owner decided the tier contents.

TOOLS OF THE TRADE What a service menu needs

Four functions carry a service menu, and none of them requires a specialist product. The first is a register of requests and exceptions, one place where every request from a customer and every exception granted on site is written down with its date and the person who asked, so that a pattern can be seen before it becomes a promise. The second is a costing sheet that turns hours into money at the loaded cost of the person who performs the work, per customer per year, so that the second test of section 6.3 can be applied in minutes rather than in an afternoon. The third is the strategy page itself, the one-page record of the choices and refusals the business has committed to, kept where the person applying the tests can see it. The fourth is a catalog that marks what is included and what is sold separately, so that a customer, a technician and a salesperson all read the same offer. No brand names belong on this list; the requirement is the function.

Visibility has a cost, and it should be counted like any other. The thirty-minute meeting at renewal is half an hour of a technician's time, and the quarterly review is three hours a year. Those hours are part of the cost of the service, not a separate expense, and they belong in the costing sheet beside the work itself. A business that counts the work and forgets the delivery will find that its margins are lower than its arithmetic predicted, and the reason will be the meetings it did not count.

6.8 The Case's Service Menu

The service menu was decided in a single session at the annual refresh, and the session had a shape worth copying. The eight candidates were listed first, each with the evidence behind it: who asked and how many, the cost per building per year, and the fit with the strategy. The three tests were applied in order, and the decisions were recorded as they were taken. The session took about two hours, and it produced five placements, three refusals, a referral list and a decision about how the capital planning note would be delivered. The reason it worked is that the evidence had been gathered beforehand, in the interviews of Chapter 2 and in the costing work of Chapter 1, so the session was a set of decisions rather than a research project.

The five placements were the capital planning note, the mid-contract note, the quarterly review, the tenant request line and the consolidated report. The capital planning note was included in Preventive and Portfolio and offered as an add-on for Essential, because seven of eight owner customers named it and because it is the service that most clearly distinguishes the offer. The mid-contract note was included in Portfolio, because it was decided at the refresh for the buildings whose equipment the first note had flagged and because it belongs to the segment that wants the closest attention. The quarterly review was included in Portfolio and offered as an add-on to Preventive, because three of six managers asked for it and five requests had been recorded. The tenant request line was included in Portfolio, because three of six managers asked for it and because it answers a problem the managers described in their own words: tenants calling them directly about things that are not urgent. The consolidated report was included in Portfolio, because four of six managers asked for it and because a manager who answers to several owners needs one document rather than one per building.

The three refusals were the energy-efficiency check, non-urgent after-hours dispatch and the filter and consumables subscription. The energy-efficiency check was asked for by two owners and declined because it belongs to the wider path of retrofit projects that the strategy closed; four owners had asked for replacement projects after reading their notes, against the five the strategy required to reopen the path, and the path stayed closed. Non-urgent after-hours dispatch was declined because it undoes the reduce the strategy made, and because the two owners who had been lost on after-hours coverage wanted a technician at night for requests that were not urgent, which is precisely the service the business had decided not to offer. The filter and consumables subscription was declined because nobody asked for it and because parts are excluded from the contract by design. Each refusal was written down with its reason and its date.

Two things came out of the session that were not on the list of eight. The first was the referral list for replacement projects: three licensed contractors, given to the owner together with the capital planning note, without a referral fee, while the wider path stays closed. The second was the decision to make the note visible, which meant delivering it in a thirty-minute meeting at renewal rather than emailing it, putting the three pieces of equipment with the shortest replacement window and a budget range on the first page, dating and signing each note, and putting a sample note with the owner's name removed into every proposal. Neither of those was a new service. Both of them changed what an existing service was worth to the customer.

The menu changed the tiers of Chapter 4 in a specific way, and the change is worth stating because it shows how a service decision becomes a packaging decision. Before the session, the offer was one contract with everything in it, and the four additions of the previous two years were inside that contract without being named. After the session, the offer had three tiers with different contents and two add-ons. The building health record is in all three tiers, because every customer needs it and it is what makes the record useful. The monthly report and the capital planning note are in Preventive and Portfolio, and the note is an add-on for Essential. The mid-contract note, the quarterly review, the tenant request line and the consolidated report are in Portfolio, and the quarterly review is an add-on for Preventive. The result is that a customer on the smallest tier can buy the note they want, a customer on the middle tier receives the note and can buy the review, and a customer on the largest tier receives everything the business has designed.

CASE IN POINT Nordec Entretien inc.: eight candidates, five placed

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The session took two hours and produced a menu that had not existed before. Five services were placed: the capital planning note in Preventive and Portfolio with an add-on for Essential, the mid-contract note in Portfolio, the quarterly review in Portfolio with an add-on for Preventive, the tenant request line in Portfolio, and the consolidated report in Portfolio. Three were declined with their reasons written down: the energy-efficiency check, because the wider path of retrofit projects stays closed; non-urgent after-hours dispatch, because it undoes the reduce the strategy made; and the filter and consumables subscription, because nobody asked and parts are excluded. The referral list of three licensed contractors was drawn up and given with the note, without a fee. The note itself was redesigned to be delivered in a thirty-minute meeting at renewal, with the three shortest replacement windows and their budget ranges on the first page, dated and signed by the lead technician, and a sample placed in every proposal.

WHAT CHANGED From free extras to a priced menu

Two dated changes meet in this chapter. At Nordec, over the strategy year and the refresh that followed it, four services were added to the contract without being placed — the building health record's upkeep, the one-page monthly report, the capital planning note at renewal and the mid-contract note — at a combined cost of $58.50 a month per building, $32,292 a year across forty-six buildings, against a price increase of $21 a month worth $11,592 a year. At the refresh, all four were placed: three inside tiers and one inside the largest tier, with the note also sold as an add-on for the smallest. The second change is older and comes from the sources. In 1995, the case for designing supplementary services deliberately, with their cost and price controlled, was made from the study of manufacturing firms (Anderson & Narus, 1995). In 1999, the case for moving downstream, into the services around a product, was made from the same direction (Wise & Baumgartner, 1999). The case's refresh is the small-business version of both arguments arriving at once.

IN LA MACHINE A service sold separately

An add-on can be recorded in the application as its own item, and the two pages that do this are the session packages page and the subscription packages page (La Machine, 2026). Business → Session Packages (productivity/packages, also reached at business/packages) defines packages of sessions or minutes: for each package a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost — the page shows the profit and the margin per sale as the package is edited —, a description and a promotional image. Business → Subscription Packages (business/subscription-packages) defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown. Business → Promotions (business/promotions) defines a promotion with a name, a slogan, a description, a discount percentage, a status, start and end dates and an image, and applies it to subscription packages, session packages or store products. What the platform does not have is the link between an add-on and a tier: each package is a single, flat item, and nothing records that the capital planning note is included in one tier and sold beside another. That relationship lives in the catalog the business keeps, not in the application.

Costly Mistakes

• Adding a service because one customer asked. A request from one person is a request from one person; the first test of section 6.3 is a count, and a service added without one is a service the business will deliver to everybody and be paid for by nobody.
• Delivering a service without placing it. A service that is neither inside a tier nor sold beside it is priced at zero, and the margin makes the decision that nobody took.
• Counting the work and forgetting the delivery. The thirty-minute meeting at renewal and the quarterly review are hours, and a costing sheet that omits them will predict a margin the business does not earn.
• Granting an exception on site and writing nothing down. An exception that is not recorded becomes a precedent, and a precedent granted three times is a service the business is delivering without having decided to.
• Presenting a paid referral as an independent recommendation. A representation to the public that is false or misleading in a material respect is a reviewable practice under the Competition Act, and a customer who would have weighed the advice differently had they known is a customer who was misled.
• Promising on a partner's behalf. The promise the customer hears is the promise the business made, and a business that says when a partner will arrive has taken on a commitment it does not control.
• Hiding a strong service in a long list. A customer who cannot name the two or three things the business does that a competitor does not will compare the offer on price alone, and the services will have cost the business money without changing the comparison.

Key Takeaways

• A service is value-added when it changes the outcome or the effort for the customer in a way the customer recognizes; a longer list is not a stronger offer, and the test is what changed on the customer's side.
• The core service is what the customer is buying, supplementary services are the business's own work designed and priced with the core, and complementary offers are what another firm provides better and the business points to rather than performs.
• Three tests sort every candidate: value to the target segment, counted rather than asserted; cost to deliver per customer per year, hours included; and fit with the strategy's choices and refusals, where the refusals do most of the work.
• Every service that passes the tests is placed in one of three ways — included in a tier, sold separately as an add-on, or declined with its reason written down — and there is no fourth option in which it is delivered and the price is not decided.
• The case's four unpriced additions cost $58.50 a month per building, $32,292 a year across forty-six buildings, against a price increase worth $11,592, leaving $20,700 a year of value delivered and never priced, and moving the contract margin from 59.08% to 52.08%.
• A service the customer does not notice has no value to the customer, which is why the capital planning note is delivered in a thirty-minute meeting at renewal, with three replacement windows and their budget ranges on the first page, dated and signed by the lead technician, and a sample placed in every proposal.

Your Action Plan

1. Open a register of requests and exceptions. Write down every request a customer makes and every exception granted on site, with the date and the person who asked, and read the register once a quarter. Deliverable: one register with at least the last quarter's entries. Verification: every entry names a customer, a date and what was asked for, and the entries that repeat are marked.
2. Cost the services the business already delivers. Take each service the business delivers today, count the hours it takes per customer per year at the loaded cost of the person who performs it, and write the annual figure beside it. Deliverable: a costing sheet covering every service currently delivered. Verification: the hours include the delivery ritual — the meeting, the call, the report — and not only the work itself.
3. Apply the three tests to the candidates on the register. For each candidate, count how many customers asked, cost it per customer per year, and check it against the strategy's choices and refusals. Deliverable: a one-page assessment per candidate. Verification: every candidate has a count, a cost and a fit judgment, and no candidate is assessed on the strength of a single request.
4. Decide and record each candidate. Place each one as included in a tier, sold separately as an add-on, or declined, and write the decision down with its date and its reason. Deliverable: a dated service menu. Verification: every service the business delivers appears on the menu, and every declined candidate has its reason recorded.
5. Make the two most valuable services visible. Choose the two services customers value most, give each a delivery ritual, a document that carries a date and a signature, and a sample that goes into every proposal. Deliverable: two redesigned delivery rituals. Verification: a customer who received each service can name it, and a prospect who read a proposal can describe it.

In La Machine

An add-on is recorded in the application as its own item, and two pages do that work (La Machine, 2026). Business → Session Packages (productivity/packages, also reached at business/packages) defines packages of sessions or minutes, each with a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost — the page shows the profit and the margin per sale as the package is edited —, a description and a promotional image, and a package can be added, edited, deactivated or deleted. Business → Subscription Packages (business/subscription-packages) defines packages billed monthly or yearly to the user's clients, each with a name, a billing period, a price and a cost in Canadian dollars, with the profit per period shown, and a package that is disabled lets existing subscribers keep their access until the end of their current billing period. Business → Promotions (business/promotions) defines a promotion with a name, a slogan, a description, a discount percentage, a status, start and end dates and an image, applies it to subscription packages, session packages or store products, and sends it to chosen clients in the application and optionally by email; the client sees active promotions at Offers → Promotions (client/promotions), with the original price struck through beside the discounted price.

What the platform does not do belongs in the same breath, because it is the part that decides where the service menu lives. La Machine has no tiers, bundles or add-ons for the user's own packages: each package is a single, flat item, and nothing in the application records that a service is included in one tier and sold beside another. There is no guarantee or warranty register, no service-level tracking and no service credits, and no refund workflow and no free trial. The relationship between a service and the tier it belongs to therefore lives in the catalog the business keeps outside the platform, and the application holds the items a customer can buy. The revenues and expenses page at business/subscriptions shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet, which is the place to count which add-on sold and how often. Book N1 (Getting Started with La Machine) maps the modules a small business otherwise buys separately, and Book N2 (Piloting Tasks, Executions and Projects) describes the tasks pages that can carry the redesign work.

Self-Check

1. What makes a service value-added, and why is a longer list not a stronger offer?
2. What is the difference between a supplementary service and a complementary offer, and why does the difference matter when a customer asks for something adjacent?
3. What are the three tests for a new service, in what order are they applied, and which one produces the refusals?
4. What are the three possible decisions for a service that passes the tests, and when is each the right one?
5. Why is a free service a price decision, and what did the case's four unpriced additions cost?

Answers. 1. A service is value-added when it changes the outcome the customer receives or the effort the customer has to spend, in a way the customer recognizes; a longer list is not a stronger offer because a service that changes nothing on the customer's side costs the business money and buys nothing with it, and the customer compares the offer on the things they can see. 2. A supplementary service is the business's own work, designed and priced together with the core, and the business answers for it; a complementary offer is something the customer needs that another firm provides better, and the business points to it rather than performing it. The difference matters because a business that treats a complementary offer as supplementary promises work it does not control, delivered by a firm it does not manage. 3. The three tests are value to the target segment, counted from the interviews rather than asserted; cost to deliver per customer per year, hours included; and fit with the strategy's choices and refusals. They are applied in that order because the first is cheap and eliminates most ideas, the second changes minds, and the third protects the business from reversing a choice it has made; the third test produces the refusals. 4. The service is included in a tier when it belongs to the promise and the segment that values it is the segment the tier is for; it is sold separately as an add-on when it is valuable to some customers and irrelevant to others, or valuable to a segment that buys a smaller tier; and it is declined when it fails one of the three tests, with the reason written down. 5. A free service is a price decision because the price is set at zero: the cost is absorbed by the margin and the customer learns that the service is free, so the margin falls without a decision and the expectation rises without a promise. The case's four additions cost $13.50, $27.00, $13.50 and $4.50 a month per building, which is $58.50 a month, or $32,292 a year across forty-six buildings, against a price increase worth $11,592 a year, leaving $20,700 a year of value delivered and never priced.

Chapter References

Anderson, J. C., & Narus, J. A. (1995, January–February). Capturing the value of supplementary services. Harvard Business Review. https://hbr.org/1995/01/capturing-the-value-of-supplementary-services

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages, business/subscription-packages and business/promotions.

Levitt, T. (1980, January). Marketing success through differentiation—of anything. Harvard Business Review. https://hbr.org/1980/01/marketing-success-through-differentiation-of-anything

Pine, B. J., II, & Gilmore, J. H. (1998, July–August). Welcome to the experience economy. Harvard Business Review. https://hbr.org/1998/07/welcome-to-the-experience-economy

Wise, R., & Baumgartner, P. (1999, September–October). Go downstream: The new profit imperative in manufacturing. Harvard Business Review. https://hbr.org/1999/09/go-downstream-the-new-profit-imperative-in-manufacturing

Chapter 7 — Productizing Services and Servitizing Products

Service firms and product firms are moving toward the same middle from opposite ends — the service firm by turning expertise into a named, fixed, repeatable package, the manufacturer by wrapping its products in services and sometimes selling the result instead of the object — and each movement has its own gains and its own traps.

What You'll Be Able to Do

• Explain the product–service continuum and say why manufacturers add services while service firms package their expertise, and why both movements are happening at the same time.
• Define a productized service as a named package with a fixed scope, fixed deliverables, a fixed duration and a fixed price, sold the same way each time before it is performed.
• Write the one-page statement of a productized service, including its exclusions, the customer's obligations and the rule that governs a request for more.
• Design a productized entry offer that a buyer can purchase at low risk and that produces the facts for a larger proposal, with a credit against the main contract as its risk reversal.
• Describe servitization as a firm changing its capabilities and processes to sell integrated combinations of products and services that deliver value in use, and distinguish product-oriented, use-oriented and result-oriented arrangements.
• Recognize when productizing or servitizing does not pay, from the financial evidence across 10,028 firms and from the three ways a service package fails.

Why It Matters

A service business sells something the buyer cannot see before buying it, and that single fact shapes every difficulty the business meets. The buyer cannot inspect the work, cannot compare two quotes that describe different amounts of it, and cannot know whether the price asked is fair for the effort involved. The seller, on the other side, cannot delegate work that lives in one person's head, cannot train a new employee against a description that does not exist, and cannot sell the same thing twice without rebuilding the scope from the beginning. Both difficulties have the same root, and both are reduced by the same move: naming the work, fixing its edges, and performing it the same way each time.

The move is not a matter of taste, and it is not a marketing trick. It is the difference between an offer a buyer can decide on and an offer a buyer must negotiate. A package with a name, a scope, a list of deliverables, a duration and a price can be read in five minutes, compared with an alternative, approved by a second person and bought. A custom engagement with the same content, described in a proposal written from scratch, takes weeks to assemble and is decided by a conversation. The content may be identical; the decision is not. This is why the entry offer designed in this chapter is a scheduled assessment with a fixed price rather than a conversation about what an assessment might cost.

The same movement runs in the other direction, and a small service firm meets it whether or not it is looking. The manufacturers whose equipment a maintenance firm services sell maintenance plans and extended coverage alongside the equipment itself, and for a building owner those plans are a servitized product competing with part of the firm's preventive scope. The firm cannot stop the manufacturers from selling them. What it can do is answer with something the manufacturers cannot offer, which in this case is a record of every system in the building whoever made it, and a package that a buyer can purchase, read and compare. Understanding both movements is what allows a small firm to see where its offer sits, what it is competing against, and which of the two directions it should be moving in.

In This Chapter

Section 7.1 describes the product–service continuum and the two movements toward its middle. Section 7.2 defines productizing and sets out what the customer gains, what the firm gains and what it gives up. Section 7.3 covers the four fixed parts of a productized service and the one page that states them. Section 7.4 designs the productized entry offer and its credit against the main contract. Section 7.5 defines servitization and what the evidence says about it. Section 7.6 runs from product-oriented arrangements to result-oriented ones. Section 7.7 examines when services do not pay, on both the manufacturing side and the service side. Section 7.8 follows the Building Check from a checklist to a product with a price.

7.1 Two Movements Toward the Middle

A manufacturer and a service firm used to be two different kinds of business, and the difference was easy to state. The manufacturer made an object, sold it, and the relationship largely ended at the loading dock. The service firm performed work, sold hours or a promise of availability, and the relationship lasted as long as the work did. A buyer knew which kind of business they were dealing with, and each kind kept to its own side of the market. That division has been dissolving for decades, and it has been dissolving from both ends at once.

The best way to see the dissolution is to stop thinking of products and services as two categories and to think of a single band instead. At the left end of the band sits a pure product: an object, made in advance, identical in every copy, bought and taken away. At the right end sits a pure service: work performed on the customer's premises, by a person, at the moment it is needed, different every time. Almost nothing a small business sells sits at either end. A restaurant sells a good and a service in the same transaction. A software subscription is a product that is delivered continuously. A maintenance contract is a service that has been given a shape. The interesting question is never which end an offer belongs to, but where on the band it sits and which way it is moving.

The movement from the left toward the middle is called servitization, and it is the manufacturer's direction. A firm that makes equipment begins to sell installation, training, spare parts, remote monitoring, maintenance plans and extended coverage alongside the object. Some manufacturers go further and stop selling the object at all, selling instead the use of it, or the result the customer wanted from it. The reason is not sentimentality about customers. An object is easy to compare with another object, and comparison pushes the price down. A service wrapped around the object is harder to compare, and it can carry a margin the object no longer carries.

The movement from the right toward the middle is productizing, and it is the service firm's direction. A firm that has been selling its expertise by the hour, or by a proposal written for each customer, begins to name what it does, fix its edges, and sell it the same way each time. The assessment becomes a package with a price. The training becomes a course with a syllabus. The audit becomes a report with a fixed list of what it covers. The reason here is not only margin. A service that is different every time cannot be delegated, cannot be taught, cannot be priced with confidence and cannot be compared by the buyer, and each of those limits the size the firm can reach.

Both movements are happening now for reasons that have little to do with fashion. Buyers have more information and less patience, and an offer that cannot be described in a page is an offer that is hard to approve. Purchasing in larger organizations is done by people who must justify a decision to somebody else, and a package with a fixed price and a fixed scope is easier to justify than an open-ended engagement. On the manufacturing side, the object itself has become harder to differentiate, which pushes the difference into everything around it. On the service side, the expertise has become harder to sell on its own, because a buyer who cannot see the work wants something they can hold.

Figure 7.1 draws the band and the two movements on it. The figure shows a horizontal band running from pure products on the left to pure services on the right. An arrow rises from the left end toward the middle, labeled "servitizing: adding services, then selling use or results". An arrow rises from the right end toward the middle, labeled "productizing: naming, fixing and repeating a service". The two arrows meet in a middle area, and in that area sit two examples side by side: a productized assessment of the kind this chapter designs, and a manufacturer's maintenance plan sold with a piece of equipment. The point the figure makes is that the two examples are different in origin and similar in shape.

Figure 7.1 — Two movements toward the middle. Manufacturers servitize by wrapping services around their products; service firms productize by naming, fixing and repeating their expertise, as Nordec does with the Building Check.

DEFINITION Productized service, servitization and product-service system

A productized service is a service that has been given the shape of a product: it has a name, a fixed scope with stated exclusions, a fixed list of deliverables, a fixed duration and a fixed price, and it is sold the same way each time before it is performed. Servitization is a firm changing its capabilities and processes to move from selling products to selling integrated combinations of products and services that deliver value in use (Baines et al., 2009). A product-service system is the arrangement that results, and the literature distinguishes several kinds of it: a product-oriented arrangement, in which the product is sold and services are added to it; a use-oriented arrangement, in which the customer pays for access to the product rather than for the product; and a result-oriented arrangement, in which the customer pays for the outcome and the provider decides how to deliver it (Neely, 2008). The three terms describe positions on one band, not three separate industries.

What the two movements share is more important than what separates them. In both, the offer becomes something a buyer can describe to a colleague, compare with an alternative and decide on without a negotiation. In both, the firm must decide in advance what is inside the offer and what is outside it, and must write the boundary down. In both, the boundary is where most of the trouble lives, because a boundary that is not written is a boundary that will be argued about after the work is done. And in both, the movement is reversible: a firm that productizes too far can add a custom option back, and a manufacturer that servitizes too far can retreat to the object.

For a small service firm, the practical consequence is a question about direction rather than about labels. Nordec Entretien inc., the maintenance firm followed through this book, is a service business moving from the right toward the middle, and it meets manufacturers moving from the left toward the middle on the same buildings. The manufacturers sell maintenance plans and extended coverage with the equipment they make. Nordec answers with a record of every system in the building, whoever made it, and with a package a building owner can buy. Neither movement is right in general. Each is right or wrong for a particular firm, and the rest of this chapter is about telling the two apart.

7.2 What Productizing Means

Productizing means taking a service the firm already performs and giving it the four things a product has. The first is a name, because a thing without a name cannot be asked for, recommended or compared. The second is a scope, stated as what the service covers and, with equal care, what it does not. The third is a list of deliverables, so that the buyer knows what will exist at the end that did not exist at the beginning. The fourth is a duration, so that the buyer knows when the work starts and when it finishes. Around those four sits the price, which Book A7 (Pricing Strategy and Revenue Models) sets, and which this book treats as a given once the contents are designed.

The four parts are not a formality, and the order in which they are written matters. A firm that starts with the price will produce a number that has no relation to the work. A firm that starts with the scope will produce a boundary that can be priced. The scope is the hard part, because it requires the firm to say out loud what it will not do, and most service firms have spent years avoiding that sentence. The exclusions are not a limitation on the offer; they are what makes the offer possible, because an offer that covers everything cannot be priced, cannot be delegated and cannot be finished.

What the customer gains from the shape is worth stating plainly, because it is the reason a productized service sells. The first gain is clarity. The buyer knows what will happen, when it will happen, what it will cost and what will be handed over at the end. A custom engagement leaves all four of those questions open until the proposal is written, and often after. The second gain is lower risk. A fixed price with a fixed scope transfers the uncertainty of the effort from the buyer to the seller, which is exactly the uncertainty a buyer is least equipped to carry. The third gain is comparison. A package with a name and a price can be placed beside another package with a name and a price, and the buyer can decide. A custom engagement can only be compared with another custom engagement, which means it is usually compared with nothing.

What the firm gains is equally concrete, and it is usually larger than the firm expects. The first gain is repeatability. When the same service is performed the same way, the second performance is faster than the first, the tenth is faster than the second, and the firm learns where the time actually goes. A custom engagement is performed once and the learning is lost with it. The second gain is delegation. Work that has a written scope, a checklist and a report template can be handed to somebody who did not design it, which is the only way a service firm grows past the number of people who can hold the whole method in their heads. The third gain is a lower cost to sell. A package that can be described in a page is sold in a conversation; a custom engagement is sold through a proposal that takes days to write, and the days are spent whether or not the proposal is won.

The gains come with a cost, and the cost is real. The first thing given up is flexibility. A productized service cannot be reshaped for each customer without becoming a custom job again, and the firm that reshapes it has paid the price of productizing without collecting the benefit. The second thing given up is some of the high-value custom work. There are customers whose needs genuinely do not fit the package, and a firm that has productized everything will either refuse them or sell them the package and disappoint them. The answer is not to refuse productizing; it is to keep one door open for the work that does not fit, and to know which door it is.

The distinction between a productized service and a custom job with a name is the distinction this chapter turns on, and it is worth stating as a test. If every sale in the last ten required a new scope, the offer is not productized. If the last ten sales were performed from the same checklist, with the same deliverables and the same duration, and the price did not move, the offer is productized. The test is not about the words on the website. It is about what happened in the last ten sales, and it can be answered from the firm's own records in an afternoon.

KEY CONCEPT A productized service is sold before it is performed

The defining feature of a productized service is not that it is simple, and it is not that it is cheap. It is that the service is fully described and priced before anybody performs it, so that the buyer is deciding on a known thing rather than on a promise. Everything else follows from that one feature. Because it is described in advance, it can be sold by somebody who is not the person who will perform it. Because it is priced in advance, the firm carries the risk of the effort and must therefore know what the effort is. Because it is the same each time, it can be improved, delegated and counted. A service sold before it is performed is a product in the only sense that matters to a small business: it can be put in a catalog, bought by a stranger and delivered by a team.

The last point deserves a sentence of its own, because it is where productizing meets the rest of this book. A productized service is not the opposite of a value-added offer; it is the container for one. The elements that make an offer valuable — the response commitment, the record, the report, the note that helps an owner plan — are easier to deliver consistently when they sit inside a package with a fixed shape than when they are improvised for each customer. Chapter 4 builds the tiers, Chapter 5 builds the guarantees that sit on top of them, and Chapter 6 decides which additional services belong inside which tier. This chapter supplies the shape that all of them are placed into.

7.3 Scope, Deliverables, Price and Duration

The four fixed parts of a productized service are scope, deliverables, duration and price, and each of them answers a question the buyer would otherwise have to ask. The scope answers what is covered. The deliverables answer what will exist at the end. The duration answers when. The price answers how much, and it is the only one of the four that this book does not design, because a price set without a scope is a guess. The four parts are written on one page, and the page is the offer. A firm that needs three pages to describe a productized service has not finished productizing it.

The scope is written as a list of what the service covers, in the language the customer uses rather than the language of the trade. For an assessment of a building, the scope names the systems that will be inspected, and it names them the way the owner thinks about them: the mechanical systems, the plumbing, the doors. The scope also states how the inspection will be performed, because the method is part of what the buyer is paying for, and it states what the buyer must provide for the work to happen. A scope that does not say what the buyer must provide is a scope that will be renegotiated on the first day.

The exclusions are written with the same care as the scope, and they are the part most firms leave out. An exclusion is not an apology. It is a statement that the offer has edges, and edges are what make a fixed price possible. For a building assessment, the exclusions name the parts of the building that will not be inspected, the trades that the firm does not hold, and the fact that no repair is included. Each exclusion protects both sides: the buyer is not surprised by a bill, and the firm is not asked to perform work it cannot perform or did not price.

The deliverables are the objects the buyer receives, and they are the part of the offer a buyer can hold. They should be listed as things rather than as activities, because a buyer cannot hold an activity. A report, a record, a note, a proposal: each is a deliverable, and each can be described in a sentence. The deliverables are also where the sample belongs. A firm that can show a sample of its report, with the previous customer's name removed, has answered most of the buyer's remaining doubt, because the buyer can see the thing they are being asked to pay for.

The duration is stated twice, because it has two parts. The first is the time the work takes, which for an assessment is one working day on site. The second is the time until the results arrive, which for the same assessment is five working days. Both matter to the buyer, and the second matters more, because a buyer who has paid for an assessment is waiting for the answer. A duration that is stated and met is the cheapest form of credibility a service firm can build.

The customer's obligations belong on the same page, and they are not a formality either. For a building assessment, the owner must provide access to every technical room and the last two years of invoices, and the assessment cannot be performed without them. Writing the obligations down does two things. It makes the offer performable, because the firm knows in advance what it needs. And it makes the offer fair, because a buyer who has been told what is required cannot later say that they were not.

The change rule is the last element of the page, and it is the element that keeps a productized service from becoming a custom job by accident. The rule is short: anything outside the scope is quoted separately before it is performed. A customer who wants more is not refused; the customer is given a second, separate offer for the additional work, with its own scope and its own price. The rule protects the package, because the package stays the same size, and it protects the relationship, because the customer is not told no.

The following table sets out the Building Check on one page, in the form the offer takes when it is written down. The columns are the part of the offer, what the firm fixes about it, and why it is fixed rather than left open. Read it as a template rather than as a description of one firm's assessment: the rows are the rows any productized service needs, and the reason column is the argument for each one.

Table 7.1 — The Building Check on one page

WATCH OUT The package that is really a custom job

The failure is quiet and it looks like good service. A firm writes a package, sells it once, and the customer asks for a small change. The change is granted, because refusing feels rigid. The next customer asks for a different small change, and it is granted for the same reason. A year later the package has been sold twelve times and performed twelve different ways, the price no longer relates to the effort, and nobody can say what the package contains. The test is arithmetic rather than philosophical: count how many of the last ten sales needed a change to the scope. If the answer is more than two, the offer is a custom job wearing a name, and the fix is not to raise the price. The fix is to write the exclusions that were missing, add the change rule, and sell the package ten times unchanged before changing anything about it.

METHOD Productizing a service in five steps

First, find the service the firm already performs the same way most of the time. The candidate is not the most profitable work and it is not the most interesting work; it is the work that already has a routine, because a routine is a scope that has not been written down yet. Second, write its scope and its exclusions on one page, in the customer's language, and include what the customer must provide. Third, fix the deliverables and produce a sample of each one, with the previous customer's name removed, so that a buyer can see what they will receive. Fourth, set the duration, state the customer's obligations and have the price set with the method of Book A7 (Pricing Strategy and Revenue Models), because a price set before the scope exists is a guess. Fifth, sell it ten times unchanged before changing anything, and count the changes that were requested in those ten sales; the count is the measure of whether the package holds.

The page is the offer, and it is worth saying what the page is not. It is not a contract, although its contents become part of the agreement between the firm and the customer, and the Québec note below says more about that. It is not a marketing document, although it can be used as one. It is the firm's own statement of what it sells, written so that the person who performs the work, the person who sells it and the person who buys it are all reading the same thing. A firm that has that page has a product. A firm that does not has a habit, and habits are hard to delegate.

7.4 The Productized Entry Offer

A productized entry offer is a small, paid, fixed-scope first purchase that lets a buyer try the firm and produces the facts for a larger proposal. It is small enough that the decision does not need a committee, paid so that the buyer is committed rather than merely interested, and fixed in scope so that the firm can perform it without a negotiation. Its purpose is not to make money on its own, although it should not lose money. Its purpose is to turn a stranger into a customer with a file, and to give the firm enough information to make a serious proposal.

The first quality of a good entry offer is that it is small. A buyer who has never worked with the firm is being asked to take a risk, and the risk should be small enough to take on a Tuesday afternoon. The second quality is that it is paid. A free assessment attracts people who are not buyers, consumes the firm's best technician's time, and produces nothing that can be followed up. A paid assessment attracts people who have a reason to want the answer, and the payment itself is evidence that the reason is real. The third quality is that the scope is fixed, so that the firm can perform the work with the same checklist every time and learn from the repetition.

The fourth quality, and the one that makes the entry offer worth designing carefully, is that it produces the facts for a larger proposal. An entry offer that ends in a report the buyer reads alone is a small sale. An entry offer that ends in a report the firm can use to write a proposal is a pipeline. For a building assessment, the facts are the state of the systems, the equipment approaching the end of its life, the five most urgent findings and the amount of unplanned work the building has been generating. Those facts are exactly what a proposal for a maintenance contract needs, and they were produced by the buyer's own building rather than by a sales conversation.

The credit against the main contract is what turns the entry offer into a risk reversal, and Chapter 5 supplies the reasoning. The buyer is told that the price of the assessment is credited in full against the first three months if a contract is signed within sixty days. The effect is that the assessment costs the buyer nothing if they go on to buy the contract, and costs them the stated price if they do not. The firm has not given anything away in the first case, because the assessment work is the onboarding the contract would have required anyway. It has been paid for the second case, which is the case where the buyer took the firm's time and did not buy.

The entry offer also serves the customer journey that Book B1 (Marketing Fundamentals for Business Owners) describes, in which a prospect is met with an offer at every stage rather than with a single request for a large commitment. A stranger is not ready to buy a maintenance contract, and asking them to do so wastes the first conversation. The same stranger may be ready to buy an assessment of their building, because the assessment answers a question they already have. The entry offer is the stage of the journey at which the buyer can act, and it moves them to the stage where the larger offer makes sense.

QUÉBEC & CANADA NOTE A fixed price is a promise under the Civil Code

In Québec, a contract between two businesses for services is governed by the Civil Code of Québec, and two of its articles bear directly on a productized service. Article 2100 provides that a contractor or provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice, and, where applicable, ensure that the work or service conforms to the contract; when bound to an obligation of result, they may be relieved of liability only by proving superior force. Article 2102 provides that, before the contract, the provider must give the client, as far as circumstances permit, useful information on the nature of the task and the property and time it requires. The fixed scope and the fixed price of a productized service are precisely what the client relies on, and the one page that states them is part of what the contract contains. A firm that performs something other than what its page describes has changed the agreement, whatever the page says. This book explains the law and does not replace the advice of a lawyer.

DO THE MATH The Building Check's margin

Illustrative figures; the case is fictional. Step 1 — The Building Check takes 6 hours on site by the lead technician, 2 hours of writing and a 1-hour meeting with the owner, plus 0.4 hour of travel, which is 6 + 2 + 1 + 0.4 = 9.4 hours. Step 2 — At the loaded technician cost of $54 an hour, the cost is 9.4 × $54 = $507.60. Step 3 — The price is $950, fixed, and it is the same for every building in the two territories. Step 4 — The margin is $950.00 − $507.60 = $442.40, which is 442.40 ÷ 950 = 46.57% of the price. Step 5 — If the owner signs a contract within 60 days, the $950 is credited in full against the first three months, and the check becomes the onboarding the contract would have paid for anyway: the firm has spent 9.4 hours of a technician's time and has acquired a contract building with a record already built.

The following table sets out what the Building Check costs and what it earns, item by item. The columns are the item, the hours it consumes and the amount it carries, so that the cost, the price and the margin can be read in one place. Read it as the arithmetic behind a productized entry offer: the hours are the reason the price is what it is, and the margin is the reason the offer can be performed without losing money while it builds the pipeline.

Table 7.2 — What the Building Check costs and earns

An entry offer is not a loss leader, and the distinction matters. A loss leader is priced below its cost in the hope of a later sale, and it attracts buyers who want the cheap thing. An entry offer is priced above its cost and below the main contract, and it attracts buyers who have the problem the main contract solves. The Building Check earns $442.40 whether or not the buyer ever signs a contract, and it earns the contract's onboarding when they do. That is the shape to aim for: an offer that pays for itself on the day it is sold, and pays twice when it converts.

7.5 What Servitizing Means

Servitization is the movement from the other end of the band, and it is worth defining precisely because the word is used loosely. One review of the literature defines it as a firm innovating its capabilities and processes to move from selling products to selling integrated combinations of products and services that deliver value in use (Baines et al., 2009). Three parts of that definition carry weight. The first is that the firm changes its capabilities and processes, not merely its price list; a manufacturer that adds a service desk without changing how it builds, sells or supports its products has added a service, not servitized. The second is that what is sold is an integrated combination, so that the product and the service are designed together rather than bolted together. The third is value in use, which means the customer is buying what the combination does rather than what it contains.

The same review finds that servitization has the potential to sustain revenue and improve profitability, and it is equally clear about the limits (Baines et al., 2009). It warns that servitization is not a solution for every firm, and it notes how little practical guidance existed for the firms attempting it. Both halves of that finding are useful to a small business reading this chapter. The first half says that the movement is not a fashion and can pay. The second half says that the evidence for it is thinner than the enthusiasm around it, and that a firm should move carefully rather than because a competitor has.

The older argument for the same movement was made about manufacturing at the end of the 1990s, when many large manufacturers saw little profit growth through the decade despite real productivity gains; the argument, as its own title says, is that the profit lies downstream, in the services around the product (Wise & Baumgartner, 1999). The reasoning is simple enough to apply to a small firm. If the object is easy to copy and easy to compare, the margin on the object falls, and the margin moves to whatever is harder to copy: the installation, the maintenance, the monitoring, the guarantee, the knowledge of how the customer uses the thing. The article offers no figure for how much profit sits downstream, and none should be inferred from it.

For a small service firm, servitization is mostly something to recognize rather than something to do. The firm does not make equipment, so it cannot wrap services around its own product. What it can do is notice that the equipment it maintains arrives with services attached. The manufacturers of the equipment Nordec services sell maintenance plans and extended coverage with their equipment, and for a building owner those plans are a servitized product competing with part of Nordec's preventive scope. The competition is real and it is narrow: a manufacturer's plan covers the equipment that manufacturer made, and it covers it well, because the manufacturer knows the machine.

The narrowness of that competition is where the answer lies, and it is the answer this chapter's case turns on. A building does not contain one manufacturer's equipment. It contains a mechanical system, a plumbing system, doors, and equipment from several makers installed over twenty years. A manufacturer's plan covers a slice of that, and the building owner must still find somebody for the rest. Nordec's building record covers every system in the building, whoever made it, and the record is the thing a manufacturer cannot produce. The firm's answer to servitization on its market is not to servitize in turn; it is to sell the one thing that the servitized products cannot cover.

CASE IN POINT Nordec Entretien inc.: the assessment that became a product

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The Building Check did not begin as a product. It began as the first visit a technician makes to a building that has just signed a contract, performed informally, without a price and without a page. What turned it into a product was a decision at the annual refresh: the firm needed a way to reach buildings in its two territories that were not under contract, and it needed a first purchase small enough for an owner to approve alone. The scope was written from two things the firm already had, the preventive visit checklist and the capital planning note, and the price was set with the method of Book A7. The offer respects the three refusals of Book A3 (Strategy and Competitive Positioning) because it is a scheduled assessment sold only in the two territories, never an emergency call, and it leads to a contract or to nothing. The pilot of Chapter 8 offered it by letter and call to forty buildings; five bought it, and two signed a contract within the pilot.

THE RULE Name it, fix it, repeat it

A productized service satisfies three conditions, and a package that fails any one of them is a custom job with a name. Name it: the service has a name a customer can use in a sentence, and the name refers to the same thing every time it is spoken. Fix it: the scope, the exclusions, the deliverables, the duration and the price are written on one page and do not change from sale to sale; a change requested by a customer is quoted separately rather than absorbed. Repeat it: the service is performed from the same checklist each time, so that the second performance is faster than the first and the tenth is faster than the second, and so that somebody other than its designer can perform it. The three conditions are cumulative. A named service that is not fixed is a slogan, and a fixed service that is not repeated is a template nobody follows.

The last point about servitization is the one a small firm should carry away. The movement is not available to every business, and the evidence for its financial benefits is mixed, as the next two sections show. What is available to every service firm is the observation that its customers are being sold servitized products by somebody, and that those products cover a part of what the customer needs. The firm's opportunity is in the part they do not cover, and finding it requires knowing what the customer's whole problem is rather than only the slice the firm has been selling.

7.6 From Products to Outcomes

The literature on product-service systems sorts the arrangements into kinds, and the sorting is useful because each kind changes who carries the risk. One study of servitization works with the three classic kinds and extends them with two more (Neely, 2008). The three classic kinds are product-oriented, use-oriented and result-oriented, and the two additions are integration-oriented and service-oriented. The five are not a ladder of sophistication; they are five ways of drawing the boundary between what the customer buys and what the provider does, and a firm can sit in any of them profitably.

A product-oriented arrangement is the closest to a plain sale. The customer buys the product, and the provider adds services to it: installation, training, spare parts, a maintenance plan, an extended coverage agreement. The customer owns the object and carries the risk of it failing. This is the arrangement a building owner meets when they buy a piece of equipment and a manufacturer's maintenance plan with it, and it is the arrangement most manufacturers actually operate, whatever language they use.

A use-oriented arrangement changes what the customer pays for. The customer pays for access to the product rather than for the product, so the provider keeps ownership and the customer pays by the period or by the use. The risk of the object sitting idle moves to the provider, and the provider has a reason to keep it working. The customer's question changes from "what does it cost to buy" to "what does it cost to have", which is a different calculation and often a simpler one to approve.

A result-oriented arrangement goes further and sells the outcome. The customer pays for the result and the provider decides how to produce it. The provider carries the risk of the effort, and the customer stops caring about the equipment entirely. The arrangement is attractive in theory and difficult in practice, because the provider must be able to control the variables that produce the result, and a provider who cannot control them has taken on a risk it cannot manage. A firm that promises a building will never have an unplanned failure has promised something no maintenance firm can deliver.

The two further kinds are worth naming because they describe what many small firms actually do (Neely, 2008). An integration-oriented arrangement sells the integration of products and services, so that the customer buys one coordinated answer rather than several separate ones. A service-oriented arrangement sells the service itself, with the product as a means to it. Both are positions on the band rather than departures from it, and both are closer to what a maintenance firm sells than the three classic kinds are.

What changes when a customer pays for availability or for results rather than for ownership is worth stating in this book's words, because it is the heart of the matter. The customer stops buying an object and starts buying a state of affairs: the building works, the equipment is available, the problem is somebody else's. The provider stops selling a thing and starts carrying a risk. The price is no longer a comparison of objects; it is a comparison of promises, and the promise must be written precisely enough that both sides know when it has been kept. That is the same discipline as the one page of a productized service, applied to a longer period.

For a small service firm, the practical lesson is about the size of the promise. A maintenance contract with a ninety-minute response commitment and a credit when the commitment is missed is a modest result-oriented element inside a product-oriented contract, and it is a promise the firm can keep and measure. A contract that promised the building would never fail would be a result-oriented arrangement the firm could not control. The line between the two is drawn by asking what the firm can actually influence, and the answer is usually narrower than the sales conversation suggests.

WHO DOES THIS The lead technician owns the Building Check

In a small service firm, a productized service needs an owner who is not the person who sells it, and the owner should be the person whose craft the service contains. At Nordec, the Building Check belongs to the lead technician. The lead technician performs the six hours on site, writes the report and the capital planning note, and attends the one-hour meeting with the owner. The lead technician also maintains the checklist the assessment follows, which is the same checklist as the preventive visit, and keeps the sample report current by removing the customer's name from the most recent one. The dispatcher books the assessment into the calendar as a scheduled day rather than an emergency slot, and the administrator prepares the page, the credit and the proposal that follows. The owner of the firm does not perform the work and does not own the package; the owner decides whether the package stays in the catalog. The division matters because a package owned by the person who sells it drifts toward whatever the last customer asked for, and a package owned by the person who performs it stays where the checklist is.

The distinction between the five kinds is also a warning about language. A firm that describes a product-oriented arrangement as a partnership, or a maintenance plan as a guarantee of uptime, has moved the customer's expectation without moving the arrangement, and the gap between the two is where disputes live. The words a firm uses about its offer are part of the offer, and a promise of performance or of length of life should be something the firm can test and show. A firm that sells a promise should be able to produce the evidence behind it.

7.7 When Services Do Not Pay

The evidence on servitization is not uniformly encouraging, and a chapter that presented only the opportunity would be misleading a reader. One study examined data on 10,028 firms incorporated in 25 countries, identified twelve servitization approaches, and compared the financial consequences of servitized and non-servitized manufacturers (Neely, 2008). The findings are worth stating exactly, because they are more interesting than either the enthusiasts or the skeptics allow.

The first finding is that servitized manufacturers were larger in revenue than the firms that were not servitized. The second is that, in aggregate, they earned lower profits as a percentage of sales. The two findings sit together and describe a pattern rather than a contradiction: the firms that moved into services grew their top line and did not grow their margin with it. The third finding is that the effect depended on size. Servitization appeared to pay off in smaller firms and proved more problematic in larger ones. The fourth is that the sample held a higher proportion of bankrupt servitized firms than expected. None of these findings says that servitization is a mistake. Together they say that it is a movement with a cost, that the cost falls unevenly, and that a firm should enter it with its eyes open.

The explanation for the pattern is not mysterious, and it applies to a small service firm as much as to a manufacturer. Services are harder to scale than products, because they consume people's time, and a firm that adds services adds people. Services are also harder to price, because the customer has no object to compare them against, and a firm that cannot price a service tends to price it low. And services carry a risk the product did not: the firm is now responsible for an outcome it does not fully control. A manufacturer that sells a maintenance plan has taken on the customer's downtime, and a maintenance firm that sells a response commitment has taken on its own travel time. Both have moved risk onto their own books, and both must be paid for carrying it.

The service side of the same problem has three shapes, and a firm can check itself against each one. The first is a scope too rigid. A package whose exclusions are drawn so tightly that the work cannot be performed without stepping outside them produces a change order on every sale, and a customer who receives a change order on every sale concludes that the fixed price was not fixed. The remedy is not to widen the scope but to redraw it around what the work actually requires, which is a job for the person who performs it rather than the person who sells it.

The second shape is a price too low for the effort. A firm that productizes a service it has never measured will price it from the hours it remembers rather than the hours it spends, and the remembered hours are always fewer. The remedy is to perform the package a few times with a time record before the price is fixed, which is why the method in this chapter sells the package ten times before changing anything. The third shape is a package nobody asked for. A firm can productize a service that its customers do not want to buy in that form, and the package then sits in the catalog while the custom work continues beside it. The remedy is the interview work of Chapter 2, which asks customers what they need before the package is written rather than after it fails to sell.

TOOLS OF THE TRADE What a productized service needs

Six functions carry a productized service, and none of them is a specialist product. The first is a one-page scope, in a form that can be sent to a customer and read in five minutes, with the exclusions and the customer's obligations on the same page. The second is a checklist, the same one every time, that the person performing the work follows and that the firm can revise when the work changes. The third is a report template with a current sample, so that a buyer can see the deliverable and so that the person writing it does not start from a blank page. The fourth is a booking calendar that holds the assessment as a scheduled day rather than an emergency slot, with the duration blocked out. The fifth is a catalog entry with a fixed price, in whatever place the firm lists what it sells, so that the offer can be bought rather than negotiated. The sixth is a way to apply the credit against the main contract, with a rule for when it expires and a record of which customers used it. No brand name belongs on this list; the requirement is the function.

WHAT CHANGED From going downstream in 1999 to the 2009 review

The argument that profit lies downstream, in the services around a product, was published in 1999, at a moment when many large manufacturers were seeing little profit growth through the decade despite real productivity gains (Wise & Baumgartner, 1999). Nine years later, data on 10,028 firms in 25 countries gave a more complicated picture: servitized manufacturers were larger in revenue and, in aggregate, earned lower profits as a percentage of sales, with the effect depending on size and a higher proportion of bankrupt servitized firms than expected (Neely, 2008). One year after that, a review of the literature defined servitization as a firm innovating its capabilities and processes to sell integrated combinations that deliver value in use, found potential to sustain revenue and improve profitability, warned that it is not a solution for every firm, and noted how little practical guidance existed (Baines et al., 2009). The three dates mark a movement from an argument, to evidence, to a cautious synthesis.

The three shapes of failure have a common root, and it is the same root as the entry offer's. A productized service is a promise about what will happen, and a promise is only as good as the firm's knowledge of what the work costs and what the customer wants. A firm that writes the page without measuring the work will price it wrong. A firm that writes the page without asking the customer will sell it to nobody. A firm that writes the page and then changes it for every customer has written a slogan. The remedy in all three cases is the same: measure the work, ask the customer, and hold the page still.

7.8 The Case's Building Check

Nordec productized an assessment rather than creating a new service, and the choice is the first thing worth understanding about the case. The firm already performed an assessment of every building it took under contract: the first visit, the walk-through, the notes taken in the technical rooms. What it did not have was a version of that assessment that could be sold to a building it had not yet signed. Creating a new service would have meant designing something the firm did not know how to perform. Productizing the existing one meant writing down what the technicians already did, which is a smaller and safer piece of work.

The scope was written from two things the firm already had. The first is the preventive visit checklist, which lists the mechanical, plumbing and door systems the firm inspects every month. The second is the capital planning note, which the firm had begun producing at renewal and which lists the equipment approaching the end of its life with a replacement window. The Building Check uses the first to decide what is inspected and the second to decide what is delivered, and the result is an assessment whose contents are the firm's own routine rather than a new invention. The exclusions were written from the contract's existing exclusions: roof and envelope, elevators, anything requiring a licensed trade the firm does not hold, and any repair.

The price is $950, fixed, and it is the same for every building in the two territories. The cost is $507.60, which is 9.4 hours at the loaded technician cost of $54 an hour, travel included, and the margin is $442.40, or 46.57% of the price. The credit is the part of the offer that makes it an entry offer rather than a small job: the $950 is credited in full against the first three months if a contract is signed within sixty days. The credit is the risk reversal of Chapter 5 applied to a first purchase, and its arithmetic is the arithmetic of the onboarding the contract would have required anyway.

The sample deliverable is what makes the offer sellable by somebody who is not the lead technician. The firm keeps a Building Check report with the owner's name removed, and uses it in proposals and on the website, so that a buyer can see the record, the note, the five most urgent findings and the proposal before deciding. A sample is not a marketing flourish; it is the answer to the buyer's central doubt, which is whether the report will be worth the price. A firm that cannot show a sample has not yet produced a deliverable it is willing to show.

The Building Check also respects the strategy that Book A3 (Strategy and Competitive Positioning) set. That strategy refused three kinds of work: one-off emergency calls for buildings without a contract, new buildings outside the two territories, and bundled tenders. The Building Check is a scheduled assessment sold only in the two territories, never an emergency call, and it leads to a contract or to nothing, so it breaks none of the three refusals. It also serves the value curve the strategy created, because the record and the note are the elements the firm chose to raise, and the assessment is the cheapest way to put them in front of a building owner who has not yet bought anything.

On the other side of the same market sit the servitized products, and the case is explicit about them. The manufacturers of the equipment Nordec maintains sell maintenance plans and extended coverage with their equipment, and for a building owner those plans compete with part of Nordec's preventive scope. The competition is narrow, because a manufacturer's plan covers the equipment that manufacturer made. Nordec answers with the building record, which covers every system in the building, whoever made it, and which no manufacturer can produce. The firm's position is not that the manufacturers' plans are poor; it is that a building is more than the equipment of one maker.

IN LA MACHINE A fixed package in the catalog

The application copy read for this book can hold a productized service as a package with a price and a cost (La Machine, 2026). At Business → Session Packages (productivity/packages, also reached at business/packages), a package is defined with a name, a type (sessions or minutes), a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image, and the page shows the profit and the margin per sale as the package is edited; the same page lists each client's credit balance and its history, and credits added by hand are recorded as free gifts or adjustments, with no payment. On the client's side, Offers → Session Packages (client/packages) lets a client buy a package through a checkout that is still a demonstration (/checkout-demo) and charges no card, or approve an order the user placed on the client's behalf. What the platform does not do is stated plainly: there are no tiers, bundles or add-ons for the user's own packages, each package being a single flat item, and there is no guarantee register, no service-level tracking and no service credits, so the credit against a later contract is applied outside the platform.

What the pilot of Chapter 8 will test is whether the Building Check does the work it was designed to do. The decision rules were written before the pilot started, and two of them concern this chapter's offer. The first is that the Building Check converts at least one buyer in four into a contract within sixty days. The second is that the tier mix of new signings stays within the bounds the firm set, with Essential at most 40% of new signings. The pilot offered the Building Check by letter and call to forty buildings, five bought it, and two signed a contract within the pilot, which passes the first rule. The honest reading of the pilot is in Chapter 8, and it includes the limits of what ten proposals can show. What the pilot does establish is that a productized entry offer can be sold by a firm that had never sold one, that it can be performed from a checklist, and that it can produce a contract.

The Building Check is a small offer, and its importance is out of proportion to its price. It is the first thing Nordec sells that a buyer can read in five minutes, compare with an alternative and approve alone. It is the first thing the firm sells that somebody other than Sophie can perform from a written scope. And it is the first thing the firm sells that produces, on the day it is delivered, the facts a proposal needs. Those three qualities are what productizing a service is for, and they are available to any service firm that is willing to write down what it already does.

Costly Mistakes

• Writing a package without writing its exclusions. A scope that says what is covered and not what is not covered produces a change order on the first sale, and a customer who receives change orders concludes that the fixed price was never fixed.
• Pricing the package from remembered hours. A firm that has never measured the work prices it from the hours it recalls, and the recalled hours are always fewer than the hours spent; the remedy is to perform the package with a time record before the price is set.
• Changing the package for every customer. A package reshaped for each buyer has paid the cost of productizing and collected none of the benefit, because it can no longer be delegated, compared or improved.
• Selling a free assessment instead of a paid one. A free assessment attracts people who are not buyers, consumes the best technician's time and produces nothing that can be followed up, while a paid one is evidence that the buyer's reason is real.
• Promising an outcome the firm does not control. A result-oriented promise such as guaranteed uptime or a building that never fails transfers a risk the firm cannot manage, and the gap between the promise and the arrangement is where disputes live.
• Assuming servitization pays because the argument is attractive. The evidence across 10,028 firms shows larger revenue and, in aggregate, lower profit as a share of sales, with the effect depending on size and a higher proportion of bankrupt servitized firms than expected (Neely, 2008).
• Productizing a service nobody asked for. A package written without the customer interviews of Chapter 2 sits in the catalog while the custom work continues beside it, and the firm concludes that productizing does not work.

Key Takeaways

• Manufacturers and service firms are moving toward the same middle from opposite ends: the manufacturer by wrapping services around its products and sometimes selling use or results, the service firm by naming, fixing and repeating its expertise.
• A productized service has a name, a fixed scope with stated exclusions, a fixed list of deliverables, a fixed duration and a fixed price, and it is sold the same way each time before it is performed.
• The customer gains clarity, lower risk and the ability to compare; the firm gains repeatability, delegation and a lower cost to sell; what the firm gives up is flexibility and some of the high-value custom work.
• A productized entry offer is small, paid and fixed in scope, and it produces the facts for a larger proposal; the credit against the main contract is its risk reversal, and it costs the firm nothing when the buyer converts because the work is the onboarding the contract required anyway.
• Servitization is a firm changing its capabilities and processes to sell integrated combinations of products and services that deliver value in use (Baines et al., 2009), and the evidence on its financial consequences is mixed: servitized manufacturers were larger in revenue and, in aggregate, earned lower profits as a percentage of sales, with the effect depending on size (Neely, 2008).
• Nordec's Building Check is a scheduled assessment at $950, sold only in the two territories, costing $507.60 and earning $442.40 at a margin of 46.57%, credited in full against the first three months when a contract is signed within sixty days.

Your Action Plan

1. Find the service the firm already performs the same way most of the time. Look for the work that already has a routine, because a routine is a scope that has not been written down, and ignore the most profitable and the most interesting work as candidates. Deliverable: one named candidate service. Verification: the last five times the work was performed can be described from memory in the same order.
2. Write the one page: scope, exclusions, deliverables, duration, the customer's obligations and the change rule. Write it in the customer's language, put the exclusions on the same page as the scope, and state what the customer must provide for the work to happen. Deliverable: a one-page statement of the offer. Verification: a person who has never performed the work can read the page and say what is inside and outside the offer.
3. Produce a sample of every deliverable and set the price with Book A7. Take the most recent piece of work, remove the customer's name, and use it as the sample; then have the price set with the method of Book A7 (Pricing Strategy and Revenue Models) rather than from remembered hours. Deliverable: a sample for each deliverable and a fixed price. Verification: the price is above the measured cost of the work, and the sample can be shown to a prospect without editing.
4. Sell the package ten times unchanged and count the changes requested. Perform each sale from the same checklist, quote anything outside the scope separately, and record every request for a change. Deliverable: ten sales and a count of the changes requested. Verification: no more than two of the ten sales required a change to the scope, and the time each sale consumed was recorded.
5. Design the entry offer and its credit. Choose a small, paid, fixed-scope first purchase that produces the facts for the main proposal, and write the credit against the main contract with its expiry rule. Deliverable: an entry offer with a price, a scope and a credit rule. Verification: the entry offer earns a positive margin on the day it is sold, and the credit is recorded when it is used.

In La Machine

Two routes in the application copy read for this book carry a productized service (La Machine, 2026). At Business → Session Packages (productivity/packages, also reached at business/packages), a package is defined with a name, a type (sessions or minutes), a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost, a description and a promotional image, and the page shows the profit and the margin per sale as the package is edited; the same page lists each client's credit balance and its history, and credits added by hand are recorded as free gifts or adjustments, with no payment. On the client's side, Offers → Session Packages (client/packages) lets a client buy a package through a checkout that is still a demonstration (/checkout-demo) and charges no card, or approve an order the user placed on the client's behalf. A fixed package with a price and a cost, bought by a client without a negotiation, is the closest thing the platform has to the offer this chapter designs.

What the platform does not do belongs in the same breath, because a firm that expects it to do more will build the wrong process. La Machine has no tiers, bundles or add-ons for the user's own packages: each package is a single, flat item, so the three tiers of Chapter 4 and the add-ons that sit beside them are described in the firm's own documents rather than assembled in the platform. It has no guarantee or warranty register, no service-level tracking and no service credits, so the ninety-minute guarantee of Chapter 5 and the credit against a later contract are applied outside it, with the credit recorded as a free gift or an adjustment on the client's balance. It has no refund workflow and no free trial, and it has no quote or proposal module, so the proposal that follows a Building Check is written elsewhere. The promotions page at business/promotions can carry a discount on a package to chosen clients, and the client sees it at Offers → Promotions with the original price struck through beside the discounted price; Book N1 maps the modules the platform does cover.

Self-Check

1. What are the two movements toward the middle of the product–service band, and which direction does a service firm travel?
2. What four things does productizing give a service, and what does the firm give up in exchange?
3. What belongs on the one page that states a productized service, and why are the exclusions as important as the scope?
4. What makes a productized entry offer different from a loss leader, and how does the credit against the main contract work?
5. What did the study of 10,028 firms find about the financial consequences of servitization, and what are the three ways a service package fails?

Answers. 1. The two movements are servitization, in which a manufacturer adds services around its products and sometimes sells use or results instead of the object, and productizing, in which a service firm names, fixes and repeats its expertise; a service firm travels the second direction, from the service end of the band toward the middle. 2. Productizing gives the service a name, a fixed scope with exclusions, a fixed list of deliverables and a fixed duration, with a price set around them; what the firm gives up is flexibility, because the package cannot be reshaped for each customer without becoming a custom job, and some of the high-value custom work that genuinely does not fit the package. 3. The page states the scope, the exclusions, the deliverables, the duration, the price, the customer's obligations and the change rule; the exclusions matter as much as the scope because edges are what make a fixed price possible, and a scope without exclusions produces a change order on the first sale. 4. A loss leader is priced below its cost in the hope of a later sale and attracts buyers who want the cheap thing, while an entry offer is priced above its cost and below the main contract and attracts buyers who have the problem the contract solves; the credit returns the price of the entry offer against the first three months when a contract is signed within the stated window, so the offer costs the buyer nothing when they convert and costs the firm nothing either, because the assessment work is the onboarding the contract would have required. 5. The study found that servitized manufacturers were larger in revenue and, in aggregate, earned lower profits as a percentage of sales, that the effect depended on size with servitization appearing to pay off in smaller firms and proving more problematic in larger ones, and that the sample held a higher proportion of bankrupt servitized firms than expected (Neely, 2008); the three ways a service package fails are a scope too rigid, a price too low for the effort, and a package nobody asked for.

Chapter References

Baines, T. S., Lightfoot, H. W., Benedettini, O., & Kay, J. M. (2009). The servitization of manufacturing: A review of literature and reflection on future challenges. Journal of Manufacturing Technology Management, 20(5), 547–567. https://doi.org/10.1108/17410380910960984

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages, client/packages and /checkout-demo.

Neely, A. (2008). Exploring the financial consequences of the servitization of manufacturing. Operations Management Research, 1(2), 103–118. https://doi.org/10.1007/s12063-009-0015-5

Wise, R., & Baumgartner, P. (1999, September–October). Go downstream: The new profit imperative in manufacturing. Harvard Business Review. https://hbr.org/1999/09/go-downstream-the-new-profit-imperative-in-manufacturing

Chapter 8 — Testing and Iterating the Offer

An offer is a set of hypotheses about what customers will choose and what the business can deliver, and the cheapest moment to find out which of those hypotheses are wrong is before the whole customer base has been moved onto it.

What You'll Be Able to Do

• Write each assumption behind an offer as a hypothesis, with the measure that would confirm or refute it, the threshold and the decision the threshold triggers.
• Choose the cheapest test that can settle a question, from an interview to a controlled experiment, and state plainly what each rung of the ladder can and cannot prove.
• Run a sixty-day pilot of a real offer with real customers in a limited area, on decision rules written before the first proposal leaves the office.
• Read the results of a small sample honestly, using the margin of error of a proportion, and say what ten proposals can and cannot show about a close rate.
• Judge when a conjoint study is worth its cost and when interviews and a pilot are enough for the decision at hand.
• Iterate an offer after a test — one change at a time, each with a recorded reason — and retire the tier or the offer that nobody chooses.

Why It Matters

Everything this book has designed so far is a belief about other people. Chapter 4 assembled three tiers with their contents, their exclusions and their prices. Chapter 5 wrote two guarantees. Chapter 6 sorted eight candidate services into those that fit and those that do not. Chapter 7 built an entry offer that carries a fixed fee and a credit against the first three months of a contract. Each of those designs rests on statements about what customers will do: that owners of smaller buildings will prefer a smaller contract to leaving for the trades billed by the call, that property managers will pay an increment for a quarterly review and a consolidated report, that a capital planning note is the element that makes a proposal persuasive, that a 90-minute guarantee will cost a handful of credits a year rather than a dozen, that a technician can complete the shorter preventive visit in the time the price assumes. None of those statements has yet met a customer. They are hypotheses, and a hypothesis is a sentence that has not been tested.

The cost of discovering a wrong hypothesis is not symmetrical, and that asymmetry is the whole argument of this chapter. A test that fails costs a few weeks, some carefully kept notes and a handful of awkward conversations with customers who were told about an offer that will now change. A roll-out that fails costs the customer base. Nordec Entretien inc. holds 46 buildings under contract, 38 of them inside its two territories, on an average of $712 a month and a run-rate of $393,024 a year. Its renewal rate is 95.5%: in the strategy year, 22 buildings renewed and one was lost. Those figures are the measure of what a wrong offer puts at risk, because a customer who is offered a badly designed tier at renewal is a customer who is being asked to reconsider a relationship, and a customer who is asked to reconsider a relationship sometimes decides not to continue it. The business that tests its offer on ten prospects and eight renewals learns the same lesson at a fraction of the exposure.

The discipline this chapter asks for is not statistical sophistication, because a business of seven people cannot run experiments on thousands of cases and does not need to. The discipline is smaller and harder: decide in advance what result would change the decision, write that down, run the test on real customers in a limited area, record the numbers every week, and then read the results against the rules rather than against the story the team hoped to tell. A business that does those five things ends a sixty-day pilot knowing which of its beliefs survived contact with customers, which one broke, and what it will change before the offer reaches the rest of the customer base. A business that skips them does not end the pilot with less knowledge; it ends the pilot believing that it already knew the answers, which is a more expensive outcome.

In This Chapter

Section 8.1 explains why an offer is a set of hypotheses and why the cost of discovering a wrong one after the roll-out is so much larger than the cost of discovering it before. Section 8.2 shows how to write a hypothesis with its measure, its threshold and the decision it triggers, and why the rules must exist before the first result is seen. Section 8.3 walks up the ladder of tests that can be run before anything is built, from interviews to a letter that measures interest. Section 8.4 distinguishes a pilot from a controlled experiment and says when a small business can randomize and when it cannot. Section 8.5 explains the margin of error of a proportion in ordinary words and uses the case's own numbers to show what ten proposals can and cannot show. Section 8.6 describes conjoint analysis and the situations in which it earns its cost. Section 8.7 covers iteration and retirement, the changes that follow a test and the measures that follow a roll-out. Section 8.8 tells the case's sixty-day pilot from the rules written the week before to the decision taken at day 60.

8.1 Why Test an Offer Before Launching It

An offer is not a description of a service. It is a wager about behavior, placed by a business that will pay for it if the wager is wrong. The wager has three parts, and they fail in different ways. The first part concerns who buys what: which customer will choose which tier, which add-on will attach, and which prospect will decide that the whole offer is more than the building needs. The second part concerns what customers value: which element of the offer is the reason a proposal gets signed, and which element is a cost the business carries without being thanked for it. The third part concerns what the business can deliver: how long the work actually takes, how often the promise is broken, how many credits a guarantee costs in a year. The first two parts are answered by customers, and the third is answered by operations. None of the three can be settled at a desk.

The offer designed in the earlier chapters of this book carries a set of beliefs that can be written down as sentences. The first is that an owner of a building under 3,500 square metres will buy a smaller contract at the Essential level rather than continue with a patchwork of trades; this belief was formed because two proposals were lost in the strategy year to the trades billed by the call, and both came from small buildings whose owners said the contract was more than the building needed. The second is that a property manager will pay an increment per building per month for a quarterly review meeting, a tenant request line and a consolidated report across the buildings in the portfolio, on the reasoning that a manager's job is to demonstrate control to the owners they report to. The third is that the capital planning note is the element that carries a proposal: 12 of 20 people interviewed named it among their top three elements, but 12 of 20 is not the same as a signature. The fourth is that a guarantee of a 90-minute response costs about ten credits a year at the current rate of 96% of urgent calls on site within the commitment, and about thirty credits at the 88% the plan targets for the first year. The fifth is that the shorter preventive visit of the Essential tier can be completed in 2.5 hours by a technician without the visit turning into a rushed inspection that produces a callback.

Each of those sentences is a hypothesis, and each has a price attached to being wrong. If the first is wrong, the business has built a tier for a customer who does not exist and has spent design and administration time on it. If the second is wrong, the Portfolio tier sells to nobody and the property managers who hold 14 buildings stay where they are. If the third is wrong, the capital planning note is a cost of $13.50 a building a month that no customer counts, and the persuasive element of the proposal is something else entirely. If the fourth is wrong, the guarantee that is meant to reassure owners is a line of credits that grows faster than the revenue it protects. If the fifth is wrong, the cost to serve of the cheapest tier is understated, and the tier loses money every month on every building it holds.

The temptation to skip the test is strongest in a business that has just finished designing something. A design is a satisfying object: it hangs together, it answers the objections the team raised in the meeting, and it makes the owner feel that the work of the last weeks has produced a thing that can be sold. Testing it means holding that object up to a customer who may shrug. That is why the argument for testing has to be made on grounds the business already accepts, and the lean start-up literature makes it in exactly those terms. The lean method favors experimentation over elaborate planning, customer feedback over intuition and iterative design over big design up front, and it treats an early venture as a search for a repeatable, scalable business model whose hypotheses are tested and revised with customers (Blank, 2013). The offer of an established small business is not a start-up, but the logic holds at its scale: the tiers, the guarantees and the entry offer are hypotheses about a repeatable way of selling, and the cheapest place to test a hypothesis is a small corner of the market.

The alternative — launching everything at once and correcting whatever breaks — has a specific cost in a business whose customers renew. Nordec's roll-out plan calls for every current customer, at renewal, to be offered the three tiers, with Preventive at the customer's current price as the default. That is 22 renewal conversations in a year, and every one of them is a customer who is being asked to think about the relationship rather than simply to continue it. If the tiers are mispriced, mispackaged or misdescribed, the correction arrives after the conversations have happened, and the customers who found the offer confusing have already formed a view. The sixty-day pilot in section 8.8 exists to move that correction earlier: it runs on new proposals and on the renewals falling inside a sixty-day window, in the two territories, while every other customer continues to receive exactly what they receive today.

There is a second, quieter benefit to testing that has nothing to do with the offer. A test is a rehearsal of the delivery. The Building Check of Chapter 7 has never been performed for a paying customer; its cost of $507.60 rests on 6 hours on site, 2 hours of writing and a 1-hour meeting. The Essential tier's shorter visit has never been performed on a real building either. Running the offer for sixty days therefore tests the operations at the same time as it tests the market, and it does so at a volume where a mistake is a lesson rather than an incident. A business that has rehearsed its offer on a handful of customers walks into the roll-out knowing what the visit sheet looks like, what the owner says when the technician calls to ask for approval to exceed the included hours, and how long the closing note takes to write.

WHAT CHANGED Online experiments moved from large platforms to small firms

The change is one of access rather than of method. In 2012 an idea for displaying advertising headlines on a search engine waited more than six months in a backlog; once it was tested in a controlled experiment it raised revenue by 12%, worth more than $100 million a year in the United States, and several large technology companies each ran more than 10,000 controlled experiments a year (Kohavi & Thomke, 2017). Those numbers belong to firms with millions of users, and the same article makes the point that follows from them: any company with at least a few thousand daily active users can run such tests, and the tests make a decision evidence-driven rather than a matter of opinion. The practice was then written down for ordinary organizations, as a discipline of defining key metrics and, ideally, a single overall evaluation criterion, of checking whether results can be trusted and of iterating on what the results show (Kohavi et al., 2020). The lean start-up argument of 2013 pushed in the same direction from the other side, asking a business to treat its plan as a set of hypotheses to be revised with customers (Blank, 2013). What changed, in short, is that the discipline of testing an offer stopped belonging to firms with data scientists.

8.2 Hypotheses and Decision Rules

A hypothesis that cannot fail is not a hypothesis, and most of the sentences a business writes about its offer cannot fail. "Customers want a fast response" cannot be refuted by anything a pilot will show, because customers do want a fast response and the sentence has no threshold in it. "Owners of small buildings will buy the Essential tier" is closer, but it still does not say how many owners, over what period, or what the business will do if the answer is two out of ten. The work of this section is to turn the beliefs of the previous one into sentences with a measure, a threshold and a consequence.

The measure is the number that will be counted, and it must be a number the business already records or can record in a minute. Proposals sent, proposals signed, renewals offered, renewals downgraded, letters sent, letters answered, urgent calls received, urgent calls missed, credits issued. Every one of those counts is a natural by-product of doing the work, which is why an offer test does not require a new system; it requires a place to write the counts down at the end of each week. The measure must also be counted over the same period for every row of the test, so that a rule about close rates is applied to proposals decided in the window and not to proposals that were still open when the window closed.

The threshold is the number that separates a pass from a failure, and it is chosen before the test so that the test has a chance to fail. A threshold is useful when it is set slightly above what the business expects, on the logic that a result below it means the hypothesis needs another look. It is useless when it is set at the level the business hopes to achieve, because then a result just below the threshold is treated as an approximation of a pass and the rule has no work to do. The pilot in section 8.8 set its close-rate threshold at 50% on at least 10 proposals, against the 57.9% of the strategy year, and it set the downgrade threshold at one renewal in four, which is a number the team regarded as the maximum it could absorb without rethinking the tier structure.

The decision is what happens when the threshold is met or missed, and it is the part that businesses omit. A test whose result triggers no decision is a report. The pilot's rules named three outcomes in advance: meet the close-rate rule and roll the tiers out to new proposals and renewals; miss it and return to the design of the entry tier before any roll-out; and breach a kill criterion — a rate of downgrades higher than one in three, or more than four credits a month — and stop the pilot and restore the previous offer for the remaining customers in the window. Kill criteria matter more than pass criteria in a small business, because the pass criteria protect the offer and the kill criteria protect the customer base.

The following table sets out the five decision rules written before the case's sixty-day pilot began, with the result each rule produced and the verdict the team recorded at day 60. Read the table as a whole rather than row by row: its purpose is to show that every rule was written with a threshold and a consequence attached, and that the verdicts were recorded against the thresholds rather than against the team's impression of a good two months.

Table 8.1 — The pilot's decision rules and results

The table has a weakness that is worth naming before the reservations of section 8.5, and it is the weakness of every table of this kind: the rules are reasonable, the results are above the thresholds, and the reader's eye still has to be told to look at the right-hand column first. The verdict "pass" appears five times, and a pass is not a proof. It is a decision that the business is willing to proceed on the evidence it has, and the willingness is a judgment rather than a computation. A business that reads a table like this one and concludes that its tiers are correct has misunderstood what the table shows; a business that reads it and concludes that nothing in it blocks the roll-out has read it correctly.

Two habits make the rules hold. The first is that the rules are written by the person who will read the results, in the company of the person who will collect them, on a single page, before the test starts. The second is that the page is dated and kept, so that a rule cannot quietly become a different rule when the results arrive. The temptation to move a threshold is strongest when a pilot is going well in every respect except one, and that is precisely the moment when the one exception is the information the business paid for.

DEFINITION Hypothesis, decision rule, pilot, controlled experiment and margin of error

A hypothesis is a sentence about what customers will do or what the business can deliver, written so that a count could contradict it. A decision rule is a hypothesis paired with a measure, a threshold and the action the threshold triggers, written before the test begins; kill criteria are the rules whose breach ends the test rather than shaping the roll-out. A pilot runs the real offer, at a real price, with real customers, in a limited place or period, and reads the results against the rules. A controlled experiment assigns customers at random to the current offer and to the new one and compares the two groups on the same measures, which is what makes the comparison attributable to the change rather than to the period. The margin of error of a proportion is the range around an observed share within which the true share is likely to sit, given the number of cases observed; it grows as the sample shrinks, and it is the reason a result read from ten cases is a direction rather than a number.

KEY CONCEPT Set the threshold before you see the result

A threshold is worth nothing if it is chosen after the result. A decision rule written in advance does two jobs at once, and the second is the one businesses underestimate. The first job is to tell the team what to count, which keeps the pilot focused on a handful of numbers instead of a dashboard. The second job is to make the result capable of being disappointing, and a result that cannot disappoint carries no information. A threshold chosen after the fact is not a threshold but a description of what happened, and a description of what happened always sounds like a reason to continue. The practice that protects a business is simple and slightly uncomfortable: the page of rules is dated and signed by the people who will read the results, it names the decision each threshold triggers, and when the results arrive it is read aloud before anyone offers an opinion about how the two months felt.

8.3 Tests Before Building

The cheapest test of an offer is the one that does not require the offer to exist. A business can learn a great deal about a tier before a single technician performs the shorter visit, and the ladder in Figure 8.1 arranges the available tests in the order in which their cost rises. Each rung buys more certainty than the one below it, and each one costs more time, more money or more of the business's credibility with its customers, so the rule is to climb only as high as the decision requires.

Figure 8.1 shows a ladder of five rungs rising from left to right. The first rung is interviews with customers, the second is a one-page description of the offer shown to customers, the third is a letter or a landing page that measures interest in the offer before it exists, the fourth is a pilot with real customers at a real price, and the fifth is a controlled experiment that compares the new offer with the current one across randomly assigned groups. Along the left side of the ladder, the cost of each rung rises as the ladder rises; along the right side, the strength of the evidence each rung produces rises with it, and the top rung stands above the others on both sides.

Figure 8.1 — The ladder of offer tests. Each rung costs more and proves more; a small business climbs only as high as the decision requires.

The first rung is the interview, and Chapter 2 treated it in full. Twenty conversations produced the elements that customers name when they are asked what matters, and those elements gave the tiers their contents. What an interview cannot do is tell the business what a customer will pay. A customer who says that the capital planning note is one of the three most valuable things the contract provides has said something true about their preferences and nothing at all about whether they will sign a proposal that contains it. Interviews are cheap, they shape the design, and they cannot be the evidence on which a price is set.

The second rung is the one-page description, and it is the least used and most useful test in a small business. The description sets out what the tier contains, what it excludes, what it promises and what it costs, in the customer's language, on one page. It is shown to a handful of customers who match the target building, and the business listens to the questions they ask rather than to the compliments they offer. A page that produces the question "and what happens when the work takes longer than two hours?" has taught the business that the exclusion is not clear. A page that produces the question "is that per building or for the whole portfolio?" has taught it that the portfolio logic is not obvious. This rung costs a morning of writing and produces a list of the misunderstandings the offer will otherwise generate in every conversation for a year.

The third rung measures interest before the service exists, and it takes two forms that behave differently. A letter works in a market of few, large customers, where the offer is a contract with a named building and the recipient can be counted: the case sent a letter describing the Building Check to 40 buildings in the two territories and compared the number of replies with the number of signatures. A landing page works in a market of many small transactions, where the visitor can be counted and a form can be filled in without a conversation. Both measure the same thing, which is whether a description of the offer produces a voluntary step toward it, and neither proves that the buyer would sign. A reply to a letter is a signal; the signature is the test.

The fourth rung is the pilot, described in the next section and told in section 8.8. It is the first rung on which money changes hands at the real price, and it is the first rung that tests the delivery as well as the demand. A pilot can show the business how long the work takes, how often the promise is kept, what the customers ask for in the first weeks and which parts of the visit sheet are never used. What it cannot show, at the scale a small business can run, is whether the offer is better than the one it replaces, because the numbers are too small and the period is too short.

The fifth rung is the controlled experiment, and most small businesses will never need it. It requires enough comparable transactions to compare two groups, and it requires the ability to assign a customer to one group or the other without the assignment itself changing the customer's behavior. A business that sells a few large contracts a year cannot climb this rung, and it should not try to pretend that it has. A business that sells thousands of small transactions through a website can, and Book C2 (Websites, Landing Pages and Conversion) treats the mechanics.

One practical rule orders the ladder. Climb to the rung that can change the decision, and stop. If the decision is whether to include a quarterly review in the Portfolio tier, the one-page description settles it, because the question is whether property managers understand and want the element, and a page answers that in a week. If the decision is whether to build the tier at all, the pilot settles it, because the question is whether buildings sign. If the decision is whether to change the default tier offered at renewal, only an experiment or a full year of renewals can settle it, and a sixty-day pilot cannot. A business that climbs higher than its decision requires spends money it did not need to spend; a business that climbs lower decides on the strength of a compliment.

TOOLS OF THE TRADE What an offer test needs

Five functions carry a test, and none of them is a specialist product. The first is a page of hypotheses and rules, dated, holding the measure, the threshold and the decision for each row, and kept where the people who run the test can see it. The second is a proposal log: one row per proposal and per renewal in the window, with the date, the building, the tier offered, the tier chosen, the reason given for a refusal and the date of the decision. The third is a spreadsheet for the counts, with a column per week and a row per rule, and a formula for the interval around each proportion so that the reading at the end is arithmetic rather than opinion. The fourth is a response form for a letter or a landing page, so that a reply is recorded in one place rather than in an inbox. The fifth, for a business with many small transactions and a website, is an experimentation function that assigns visitors at random and reports the difference between the two groups (Kohavi et al., 2020). No brand name belongs on this list; the requirement is the function.

8.4 Pilots and Controlled Experiments

A pilot and a controlled experiment are both tests on real customers, and they answer different questions. The pilot asks whether the offer works at all in the conditions of the business: whether customers sign it, whether the work can be delivered at the stated cost, whether the promise holds. The controlled experiment asks whether the new offer works better than the offer it replaces, and it answers that question by comparing two groups of customers who differ only in which offer they were shown. The first question is the one a small business usually needs to answer; the second is the one that requires a business to have enough customers to divide them.

The distinction rests on randomization, and randomization is easier to describe than to achieve. In a controlled experiment, each customer who arrives in the window is assigned to the current offer or to the new one by a rule the customer cannot influence and the staff cannot override, and the two groups are then compared on the same measures. Nothing else about the customer's experience changes. Randomization is what makes the comparison attributable to the offer rather than to the period, the region, the weather, the season or the enthusiasm of the person making the sale. A pilot without randomization compares the new offer with the previous period, and that comparison is always contaminated by whatever else changed between the two periods.

The case's pilot is a good illustration of a business that cannot randomize and should not pretend otherwise. Nordec decides 19 proposals in a strategy year. In a sixty-day window, 10 proposals were decided. If the business had split those ten at random into a group shown the tiers and a group shown the old contract, it would have had five proposals in each group, and a close rate computed on five cases tells the reader almost nothing. Worse, the split would have been visible to the customers and to the technicians, because the offer is presented in a conversation with a person who knows both versions. In a market of few, large contracts, the honest test is the pilot against the historic rate, with the limitations of that comparison stated in advance rather than discovered in the discussion.

A business can randomize when it has many small transactions and a channel that assigns them mechanically. A retail counter that serves hundreds of customers a week, an online store, a service booked through a website, a subscription sold at the end of a self-service sign-up: all of these can divide arrivals at random and compare the outcomes, because the assignment is invisible and the volume is sufficient. The threshold the practitioners give is a concrete one: any company with at least a few thousand daily active users can run controlled tests, and a business below that scale should expect the results of an experiment to be too noisy to act on (Kohavi & Thomke, 2017).

For the business that can randomize, the discipline of the experiment is the discipline of the rules in section 8.2, applied to a comparison rather than to a single group. The metrics are defined before the test, and the practice of naming a single overall evaluation criterion — the one number that decides whether the change stays — is what keeps a team from choosing the metric that flatters the result after the fact (Kohavi et al., 2020). The checks that the result can be trusted come before the result is read: whether the two groups were in fact comparable at the start, whether the traffic was divided as intended, whether the period contained an event that affected only one group. And the interpretation comes last, with the change kept, changed or retired according to what the criterion says.

Whether a business randomizes or not, the pilot's design has four features that make the difference between a test and an anecdote. The first is a boundary: the offer changes for new proposals in a stated area and for the renewals falling in a stated window, and for nobody else. The second is a fixed period, decided in advance, of sixty days in the case's pilot, long enough to contain several proposals and several renewals and short enough that a failure does not damage a season. The third is a single source of truth for the counts, kept weekly, so that the reading at the end is a subtraction rather than a recollection. The fourth is that everyone who presents the offer presents the same offer: the same page, the same contents, the same exclusions, the same price, the same three tiers in the same order.

The boundary deserves one more sentence, because it is where pilots go wrong. A pilot that changes the offer for a group of customers and also changes how the business answers the telephone, how quickly it invoices and who visits the building has tested a bundle of changes and can attribute the result to none of them. The case's pilot changed the offer, and it kept everything else — the response commitment, the billing cycle, the dispatch rules, the technician assigned to each territory — exactly as it was. That is what makes the day-60 reading possible to interpret at all.

METHOD A sixty-day offer pilot

First, write the hypotheses, the rules and the decisions they trigger on one dated page, with the measures named as counts the business already produces. Second, choose where the pilot runs and what stays unchanged: name the territory, the window, the customers included and the list of things the pilot will not touch. Third, train the people who present the offer and give them the same page, so that every proposal carries the same contents, the same exclusions and the same price, and rehearse the answers to the three questions customers ask most. Fourth, record every proposal, renewal and delivery measure weekly, in one place, with the date of each decision. Fifth, read the results against the rules at day 60, with the interval around each proportion computed, and record the decision each rule triggers, including the changes the pilot itself suggested.

THE RULE A controlled experiment needs randomization and enough cases

A controlled experiment compares a new offer with the current one by assigning customers at random to two groups and comparing them on the same measures; without the random assignment, the difference between the groups is a difference between two periods or two sets of customers rather than a difference made by the offer. The second requirement is volume: the practice reports that any company with at least a few thousand daily active users can run such tests, and that the large technology firms running more than 10,000 experiments a year are operating at a scale that makes small differences readable (Kohavi & Thomke, 2017). The discipline that goes with the volume is to define the key metrics, and ideally a single overall evaluation criterion, before the test, and to check whether the result can be trusted before reading it (Kohavi et al., 2020). A business that decides ten proposals a year satisfies neither requirement, and its honest instrument is the pilot, read against its own historic rate and against the rules it wrote.

QUÉBEC & CANADA NOTE Testing an offer on real customers

A test that sells something is not a rehearsal; it is a contract. Nordec's contracts are between two businesses in Québec and are governed by the Civil Code of Québec, under which a provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice, and must give the client, before the contract, useful information on the nature of the task and the time it requires (art. 2102), while the client may resiliate a service contract unilaterally on paying for the work performed and the costs incurred (arts. 2125 and 2129). A business that tests an offer on consumers meets a different regime in full: in Québec, the Consumer Protection Act governs every contract the test produces, requiring among other things that goods and services conform to the merchant's statements and advertising, which bind the merchant even when a warranty is not written into the contract (ss. 41 to 43), and that any exclusion from a conventional warranty be set out clearly in separate successive clauses (s. 44). From October 5, 2026, new goods of listed kinds sold to consumers carry a warranty of good working order whose duration is set by regulation (ss. 38.1, 38.2 and 38.8). This book explains the law; it does not replace the advice of a lawyer.

8.5 Reading Small Samples Honestly

The hardest part of testing an offer is not running the test. It is reading the result without claiming more than the test can show, and the difficulty is not a failure of intelligence or of goodwill. It is a feature of small numbers. A business that has just spent sixty days on a pilot has a result it cares about, a team that has worked on it, and a customer base that expects a decision. In that situation, any reading that supports the roll-out feels like realism and any reading that qualifies it feels like hesitation. The arithmetic of the margin of error exists to make the qualification concrete.

A proportion is a share of a set of cases: six signatures out of ten proposals is a proportion of 0.6. The margin of error around it says how far the observed share could sit from the share the business would see if it sent thousands of proposals. It is computed from the observed share, the number of cases, and a multiplier of 1.96, which corresponds to a range that would contain the true value about 95 times out of 100 if the cases were drawn independently. The formula is small enough to do on a page, and its meaning is what matters: the smaller the number of cases, the wider the range.

For the pilot, the arithmetic runs as follows. Six signatures out of ten proposals give an observed share of 0.6; multiplying 0.6 by 0.4 gives 0.24; dividing by ten gives 0.024; the square root of 0.024 is about 0.155; multiplying by 1.96 gives about 0.304, or roughly 30 percentage points. The range around the observed 60% therefore runs from about 30% to about 90%, and the 57.9% of the strategy year sits inside it. The pilot, in other words, cannot show that the three tiers changed the close rate. It can show that the close rate did not collapse, and it can show who said yes and what they chose, and those are different kinds of information from a comparison of rates.

The distinction between the two kinds of information is the practical heart of this section. Rates need numbers; cases need only a story that can be followed. A pilot of ten proposals cannot tell a business that its close rate rose, and it can tell the business a great deal about which customers can say yes to which tier, because each of the ten proposals is a case with a name, a building, a size, a reason and a decision. In the case's pilot, two of the six signatures came from small buildings under 3,500 square metres, of exactly the kind that had been lost in the strategy year to the trades billed by the call. That is a fact about the offer that the close rate cannot express, and it is the fact on which the roll-out decision mainly rested.

The pre-set rules protect this reading. When the team gathered at day 60, the page written before the pilot began was read first, rule by rule, and each result was placed beside its threshold before any opinion was offered. The reservation about the close rate was written into the day-60 note in the same paragraph as the pass, rather than in a footnote added later. That order is the discipline: rules, results, verdicts, and then the discussion of what the numbers cannot say. A team that discusses the two months first and consults the rules afterwards will find reasons for whatever it has already decided.

There is one more honesty problem in a small pilot, and it concerns the period rather than the sample. A sixty-day window contains the season, the weather, the holidays or the vacation of the person who presents the offer, and a business that compares the window with the whole previous year is comparing two things that differ in many ways. The pilot's numbers therefore sit beside the historic rate as a reference point and not as a control. The historic rate is the figure the business actually achieved over 19 decisions in the strategy year, 11 won and 57.9%, and the reason to state it is not that it makes a statistical comparison but that it stops the reader from treating 60% as a discovery.

The following table sets out four questions a business would like its pilot to answer, what the case's sixty-day pilot can and cannot show about each of them, and what would be needed to settle the question properly. Read it as a piece of intellectual housekeeping: it names the questions a small test cannot answer before someone in the business asks them in a meeting and receives an answer built from a favorable reading.

Table 8.2 — What the pilot can and cannot show

WATCH OUT The pilot that proves what everyone hoped

Reading the results without the pre-set rules turns every pilot into a success. The mechanism is always the same and it is not dishonest: the team remembers the two proposals that were won against the odds and forgets the three that were decided without a reply; the rule that a renewal count was supposed to reach eight is quietly replaced by a rule that most renewals stayed where they were; the threshold for credits is treated as an average rather than a maximum. The result is a roll-out decision supported by a story instead of a test, taken with the confidence that comes from having done the work. Two habits close the gap. The first is to read the page of rules aloud, row by row, before anyone describes how the two months felt, so that the thresholds are on the table while the results are still being placed beside them. The second is to write down in the day-60 note what the pilot did not show, in a sentence that will survive into the roll-out file, because the next person to read the file will otherwise inherit the confidence without the reservation.

DO THE MATH How much ten proposals can tell

Illustrative figures; the case is fictional. Step 1 — The pilot closed 6 signatures on 10 proposals, an observed share of 0.6, or 60%. Step 2 — Multiply the share by its complement and divide by the number of cases: 0.6 × 0.4 = 0.24, and 0.24 ÷ 10 = 0.024. Step 3 — Take the square root of 0.024, which is about 0.155, or 15.5 points. Step 4 — Multiply by 1.96 for a range that would contain the true value about 95 times out of 100: 0.155 × 1.96 = 0.304, or about 30 points. Step 5 — The interval around the observed 60% therefore runs from about 30% to 90%, and the 57.9% the business achieved in the strategy year sits inside it. The pilot cannot show that the tiers changed the close rate; it can show which customers said yes, at which tier, and on what grounds.

8.6 Conjoint and Choice Experiments

The tests in the previous sections measure what customers do with an offer that has already been assembled. Conjoint analysis measures something different and earlier: what each part of an offer is worth to a customer, by asking them to choose among combinations that the business has not yet decided to sell. It is the only family of methods in this chapter that can answer a question of trade-off, and it is worth understanding even for a business that will never pay for it, because the questions it asks are the questions a good interview tries to ask and cannot.

The method works by presenting a respondent with a series of alternatives. Each alternative is a combination of the attributes of the offer, and each attribute takes one of several levels. For an offer like Nordec's, the attributes might be the response commitment, the number of preventive visits a year, whether the capital planning note is included, whether a monthly report is included, and the monthly price at four levels. The respondent chooses among the alternatives or rates them, and the analysis estimates the value each level contributes to the choice. The output is a table of relative values: how much the response commitment matters against the monthly report, how much the note matters against the price. The 1990 review of the method describes its developments since the authors' earlier survey, including the approaches available when an offer carries a large number of attributes, which is the situation in which the design of the study becomes a problem in itself (Green & Srinivasan, 1990).

The practical value of conjoint analysis is that it can put a number on a trade-off before a single contract is signed. A business that is deciding whether the Portfolio tier should carry a quarterly review meeting or a tenant request line can ask respondents to choose between offers that contain one or the other and read which combination they prefer. The weakness is that the answers come from hypothetical choices, and a hypothetical choice is a statement of preference rather than a purchase. The method therefore belongs beside the pilot and not instead of it: conjoint tells a business which combination of attributes its market prefers, and the pilot tells it whether the market will buy the combination it chose.

Whether the cost is worth it depends on two conditions, and both have to hold. The first is a market large enough that a study of a few hundred respondents is meaningful, because the method needs a sample and the sample is the expense. The second is an offer with several attributes whose levels can be varied and whose interactions matter, because an offer with one attribute and two levels does not need a study; it needs a letter and a reply. A regional manufacturer choosing among configurations of a machine, a software publisher choosing among feature sets, a retailer choosing among warranty lengths: these are the situations in which the method earns its cost.

Nordec's situation satisfies neither condition, and the reasoning is worth stating because it is the reasoning the reader will apply to their own business. The market is 1,400 buildings within forty kilometres of the depot, of which 46 are under contract. A study large enough to be read would need a substantial share of that market, and the people who own the buildings are the same small set of owners and managers the business already knows, which means the business can do something cheaper: show them the one-page description, ask them to choose between two versions of the Portfolio tier, and count the answers. The design of the tiers in Chapter 4 came from 20 interviews and an analysis of three signals, and the price came from Book A7 (Pricing Strategy and Revenue Models), which sets prices with its own method. Where a market of many buyers and many attributes exists, conjoint is the instrument that answers the question; where the market is a list of names the owner knows, a conversation answers it.

One link deserves a sentence before the next section. The attributes that a conjoint study measures are the same attributes that Chapters 4 and 5 packaged into tiers and guarantees, and the value attached to each level is precisely the information a price should rest on. This book designs the contents of the offer and the promises attached to it; the price of those contents is set by the method of Book A7, and a business that has measured the value of the levels has given its pricing work a better foundation than a business that has not. A study is never a substitute for the knowledge of what a customer will pay at the moment of the signature.

8.7 Iterating and Retiring Offers

A test that produces no change is a test whose result was already known, and the pilot of section 8.8 produced three changes before the roll-out began. The discipline that governs them is the same discipline as the rules: one change at a time, each with a reason attached, each small enough to be described in a sentence. A roll-out that carries five improvements at once has hidden the evidence for all five, and the next test will have to start from a state nobody can explain.

The first change came from a complaint that the pilot itself surfaced. Two Essential customers said the visit sheet arrived too late in the day to be useful, because it was written at the end of the technician's round and sent when the office closed. The response was to shorten the visit sheet to the items that matter to the customer and to have it written before the technician leaves the building, which removes the administrative step from the end of the day and makes the document a record of the visit rather than a report about it. The reason is recorded with the change: two customers raised it unprompted, in the first three weeks, and the change costs a few minutes of technician time and no money.

The second change came from the reply rate to the letter. The Building Check letter produced five buyers out of forty buildings, which met the rule, and the team noticed that the buildings that replied were the ones whose letter mentioned the note they would receive. The letter was rewritten to lead with the capital planning note rather than with the assessment, and the change was recorded with its reason: the note is the element that the interviews of Chapter 2 and the replies of the pilot both pointed to, and the letter should present what the buyer values first rather than what the seller performs first. The third decision was to keep the downgrade threshold exactly as it was, with a written reason of its own: one downgrade in eight renewals is within what the tier structure can absorb, and changing a threshold because a result was comfortably inside it is the habit that section 8.5 warns against.

The pilot also settled a question the design had left open, which was whether the three tiers should be offered in a fixed order. The order was kept as it stands, with Essential first in the conversation only where the building is under 3,500 square metres and Preventive presented as the flagship elsewhere, and the reason was recorded. This is the kind of decision that a test makes easy and a meeting makes endless: the pilot had ten conversations in which the order was used and the results could be read, and a meeting has an argument about customer psychology that nobody can settle.

Retirement is the other half of iteration, and it is harder than addition because a tier or a service that nobody chooses still has an owner. The service menu of Chapter 6 declined three candidates — an energy-efficiency check, non-urgent after-hours dispatch and a filter and consumables subscription — on the grounds that they did not fit the strategy, and those decisions are retirement decisions taken before the offer existed. An offer that is already published needs a different treatment: it needs a measure and a date. The practical rule is to give every published element a review date and a count, and to retire the element whose count stays at zero for two consecutive reviews. A tier with no takers is not neutral; it occupies a line in the offer, it appears in every conversation as an option the customer must consider and reject, and it consumes the time of the person who presents it.

What a business measures after the roll-out is different from what it measured during the pilot, and the difference is volume. A pilot reads counts per proposal; a roll-out reads rates per month. The four measures that matter for an offer of this kind are the take rate of each tier as a share of new signings, the number of downgrades and upgrades at renewal, the attach rate of each add-on against the tiers that can carry it, and the delivery measures that say whether the promises are being kept: first-visit resolution, urgent calls missed, credits issued, and the cost to serve of each tier against the price it carries. The sixth-month reading of Chapter 12 puts those measures side by side for the case, and it is the reading that tells a business whether the offer it tested is the offer it is actually running.

One more practice makes iteration disciplined rather than restless. Every change to an offer should be written on the same page as the hypothesis it is meant to improve, so that the next reading can say whether the change did what it was supposed to do. A change made for a reason and reviewed later is an iteration; a change made because a customer complained on a Thursday is a repair that will be repaired again. The difference is not the size of the change but whether anybody wrote down what it was for.

IN LA MACHINE What a pilot can use in the platform

Three functions of the application copy read for this book carry a pilot's bookkeeping (La Machine, 2026). At Account → Profile → Revenues & Expenses (business/subscriptions) the platform shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet, which is the place to count which package sold and how often; the platform's packages are flat items, so a business running tiers counts its own tiers outside it. A letter or a short survey can be carried by the general form builder at Marketing → Forms (business/marketing/forms), with the answers kept with the form at business/marketing/forms/responses. The pilot's hypotheses, decisions and weekly counts can be carried as Tasks at productivity/tasks, which give each task a verification criterion in the Control center at /taches. What the platform does not have is stated plainly when it matters for this chapter: there is no experiment or A/B testing tool, no tier, bundle or add-on structure for the user's own packages, no guarantee register and no service-level tracking, so the test's design, its randomization and its reading are done outside the platform. Book N2 describes the tasks pages and Book N1 maps the modules.

8.8 The Case's Sixty-Day Pilot

The pilot began with a page written the week before it started. Sophie Nadeau and the administrator sat down with the five rules of Table 8.1, agreed the thresholds, named the measures, and fixed the boundary: the three tiers would be offered on every new proposal in the two territories and on every renewal falling inside the sixty days, and nothing at all would change for any other customer. The existing customers outside the window would continue on the contract they had, at the price they paid, with the same response commitment and the same monthly report. The lead technician would present the offer the same way in both territories, using the same page, and the technician who visited a building would not know whether the customer had been a pilot customer or not.

The page also fixed what the pilot would not touch, and the list is worth stating because it is what makes the reading possible. The dispatch rules stayed as they were, so the 90-minute commitment was answered by the same rotation as before. The billing cycle stayed monthly, on the same date. The visit checklist for the Preventive tier stayed the one the business already used, because the tier is the current contract cleaned up rather than a new service. The only new pieces of work the pilot brought were the entry offer and the two add-ons, and the entry offer was performed by the lead technician, who had the 6 hours on site and the 2 hours of writing that its cost assumes.

The first proposals arrived in the second week. The pilot's log recorded ten proposals decided in the sixty days, and it recorded them with the building, the size, the tier offered, the tier chosen and the reason given when nothing was signed. Three signatures came in the first three weeks, all of them Preventive, and the fourth week produced the two small buildings that the day-60 note would return to: one of about 2,500 square metres and one slightly smaller, both of which had been shown the Essential tier and both of which signed it at $490 a month. They are the customers the strategy year had lost to the trades billed by the call, and they are the reason the tier exists.

The renewal meetings ran in parallel with the proposals, and they were the part of the pilot the team found hardest. Eight renewals fell inside the window, and each one began with an explanation of something the customer had not asked about: three tiers, a new mid-contract note, a premium level for portfolios. One building came from a scattered location and stayed on Preventive at the price it already paid, which is the default the migration rule sets. One owner of a small building asked to move down to Essential, which the rule allowed and the log recorded. Five customers stayed where they were, and two of them — property managers with several buildings each — moved up to the Portfolio level, which is where the $175 increment meets a customer whose job is to report to somebody else.

The Building Check ran through the same window on a different track. The letter describing the $950 assessment went to 40 buildings in the two territories that were not under contract, by letter and by a follow-up call, and five owners bought it. Two of those five signed a contract within the sixty days, and both are counted among the six signatures, which means that the entry offer paid for itself twice over in the sense that matters: the assessment fee was credited against the first three months of each contract, and the business acquired two buildings it had not reached through a proposal.

The delivery numbers were collected weekly and read at the end. Forty-four urgent calls arrived in the sixty days, two were missed, and $300 of credits were issued, which is the rate the guarantee assumed and comfortably inside the two-a-month rule. The two missed calls were the pilot's most honest number, because a guarantee that never pays is a guarantee nobody believes, and a guarantee that pays twice in sixty days is one the customers can see working. The credits were issued automatically, without a claim, as Chapter 5 designed them.

Day 60 fell on a Friday, and the reading was held the following Monday with the page of rules on the table. The rules were read one by one, then the results, then the verdicts. The close rate was 6 of 10, 60%, against the 57.9% of the strategy year, and the interval around 60% runs from about 30% to 90% on ten cases, which means the pilot cannot show that the tiers raised the close rate. The downgrade rule passed at one in eight against a maximum of one in four. Essential took 2 of the 6 signings, 33% against a maximum of 40%. The Building Check converted 2 of 5 buyers against a requirement of one in four. The guarantee held at two credits in sixty days against a maximum of two a month. Every rule passed, and the note recorded in the same paragraph that passing is not proof and that the case the pilot can actually make is about who said yes and what they chose.

The decision was to roll out, and it came with three changes. The Essential visit sheet was shortened and is now written before the technician leaves the building, after two customers said it arrived too late in the day. The Building Check letter was rewritten to lead with the capital planning note, after the replies showed which building owners answered. The downgrade threshold was kept exactly as written, with the reason recorded beside it. The roll-out itself is the subject of Chapter 9, which sets the cadence and the workflows of the offer once it is live, and Chapter 12 reads the six months that followed, when the business held 52 buildings, 8 new signings and 5 upgrades to Portfolio, with contract margin at 52.35% against 51.15% at the start.

CASE IN POINT Nordec Entretien inc.: two small buildings and two upgrades

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The sixty days produced six signatures, and the two that matter most are the smallest. A building of about 2,500 square metres and one slightly smaller had both been lost in the strategy year to the trades billed by the call, and both had told the business, in the words recorded at the time, that the contract was more than the building needed. The Essential tier was designed for exactly that sentence, and in the pilot it produced two signatures at $490 a month from owners who had already refused once. The other half of the story is in the renewal meetings, where two property managers moved to the Portfolio level and one small owner moved down. Two customers chose to pay more per building, one chose to pay less, and five stayed where they were. A close rate on ten proposals cannot be distinguished from the previous year's 57.9%, and the pilot can say with confidence that the offer now contains something a small owner can buy, that a property manager will pay an increment for a view across buildings, and that a customer offered the tiers will sometimes take a smaller one rather than leave.

WHO DOES THIS The administrator keeps the pilot log

In the case followed through this book, the administrator held the pilot log, and the choice of that person is what made the reading possible. The log was a single page of the spreadsheet the business already used, with one row per proposal and per renewal: the date, the building, the size, the tier offered, the tier chosen, the reason given and the date of the decision. It was updated at the end of each week rather than at the end, which is what gave the day-60 reading a set of counts instead of a reconstruction. The administrator also kept the letter log, recording which of the forty buildings was contacted, whether a call followed the letter and what the owner said, and the count of urgent calls and credits came from the same weekly sheet. The rule that made the log work is the rule of the platform's own task pages: every line has a verification criterion attached, so that a row is entered only when the decision it records is final. Sophie read the log weekly, which took ten minutes, and that reading is what allowed the day-60 meeting to be an hour of decisions rather than a morning of archaeology.

Costly Mistakes

• Launching an offer to every customer without testing it on any. A roll-out converts the whole customer base into the sample, and a mistake discovered afterwards has already been discussed at 22 renewal meetings.
• Writing the decision rules after the results arrive. A rule that is chosen once the numbers are known is a description of what happened, and a description of what happened always supports continuing.
• Treating a pilot as a controlled experiment. Ten proposals read against a historic rate cannot attribute a difference to the offer, and a report that says otherwise will be used as proof in the next meeting.
• Reading the close rate and ignoring the cases. The rate on a small sample is a range, while the names, sizes and reasons behind each proposal are facts, and it is the facts that carry the roll-out decision.
• Climbing higher up the ladder of tests than the decision requires. A business that commissions a study for a question a one-page description would answer has spent money on certainty it did not need.
• Changing three things at once after the pilot. A roll-out carrying several simultaneous improvements has hidden the evidence for each of them, and the next reading will compare a state nobody wrote down.
• Leaving a tier or an offer that nobody chooses in the catalog. An element with no takers occupies a line in every conversation and a share of the presenter's attention, and it should be given a count, a review date and, if the count stays at zero, a retirement.

Key Takeaways

• An offer is a set of hypotheses about who buys what, what customers value and what the business can deliver, and the cost of discovering a wrong hypothesis after the roll-out is measured in renewal conversations with the existing customer base.
• Every hypothesis needs a measure, a threshold and the decision the threshold triggers, written and dated before the first result is seen, together with the kill criteria that protect the customers during the test.
• The ladder of tests runs from interviews, through a one-page description and a letter that measures interest, to a pilot with real customers and a controlled experiment, and a business climbs only as high as the decision requires.
• A controlled experiment needs random assignment and enough cases, which means a business with a few thousand daily transactions can run one and a business deciding ten proposals a year cannot; its honest instrument is the pilot read against its own historic rate (Kohavi & Thomke, 2017).
• Ten proposals yielding six signatures give an observed close rate of 60% with an interval of about 30% to 90%, which contains the previous year's 57.9%, so a small pilot can show who says yes and what they choose but cannot show that the offer changed the rate.
• A test that sells something creates real contracts, and the changes that follow a pilot are made one at a time with a reason recorded, while the elements nobody chooses are given a count, a review date and eventually a retirement.

Your Action Plan

1. Write the hypotheses and the rules on one dated page. Take the three beliefs behind the offer — who buys which tier, which element carries the proposal, what the delivery costs — and give each one a measure, a threshold, the decision it triggers and, where the risk is high, a kill criterion. Deliverable: one dated page of rules signed by the people who will read the results. Verification: every row names a count the business already produces, and every row has a decision attached.
2. Choose the test rung that can change the decision and stop there. Walk the ladder of the figure in section 8.3 and select the cheapest rung that can settle the question at hand, from a one-page description for an element to a pilot for a whole tier. Deliverable: a written choice of test with the reason it is sufficient. Verification: the chosen rung is named against the decision it must support, and the rungs above it are named as unnecessary with a reason.
3. Fix the boundary and the period of the pilot. State which territories, proposals, renewals and customers the pilot covers, how long it runs, and what will not change for anybody during it. Deliverable: a boundary of one paragraph and a date for the reading. Verification: every other customer continues to receive exactly what they receive today, and the list of things the pilot will not touch is written down.
4. Keep one weekly log with a verification criterion per line. Record each proposal, renewal, letter and urgent call with its date and its outcome, in one place, updated at the end of each week by one named person. Deliverable: a weekly log covering the whole pilot window. Verification: the person who presents the offer can reconcile the log against their own records, and every line names a final decision.
5. Read the rules first at the end and record what the pilot did not show. Hold the reading with the page of rules on the table, place each result beside its threshold before discussing the period, and note the interval around each proportion. Deliverable: a day-60 note with the verdict per rule and a paragraph naming what the pilot cannot show. Verification: the note lists the changes the pilot produced, each with its reason, and the reservation about the sample size appears in the same paragraph as the pass.

In La Machine

Three routes carry a test's bookkeeping in the application copy read for this book (La Machine, 2026). The page at Account → Profile → Revenues & Expenses (business/subscriptions) shows three tables for an agent or an organization — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet; it is the place to count which package sold and how often, one row per package rather than a tier structure. A letter or a short survey can be carried by the general form builder at Marketing → Forms (business/marketing/forms), with the answers kept with the form at business/marketing/forms/responses, so that the replies to a letter of the kind the pilot sent are recorded in one place. The pilot's hypotheses, its weekly counts and its day-60 decisions can be carried as Tasks at productivity/tasks, where the Control center's plan tasks page at /taches gives each task a verification criterion. Nothing about a tier, a threshold or an interval is specific to the platform; the test is designed and read by the business.

What the platform does not do belongs beside that list. La Machine has no experiment or A/B testing tool, so a business that can randomize does it outside the platform; it has no tiers, bundles or add-ons for the user's own packages, since each package is a single, flat item, so the three-tier pilot of the case is recorded in its own page rather than in the package definitions; it has no guarantee or warranty register and no service-level tracking, so the credits the 90-minute guarantee produces are counted in the weekly sheet; and it has no quote or proposal module, so the proposals the pilot decides live in the business's own document. The page at Business → Motivational Quotes (business/quotes) holds motivational quotations and nothing else, which is worth knowing before looking for a quotation function there. Where a business wants a package structure, the pages at Business → Session Packages (productivity/packages) and Business → Subscription Packages (business/subscription-packages) hold the shape a platform package can take, with a price and a cost per sale shown as the package is edited; Book N1 maps what each module covers.

Self-Check

1. Why is an offer a set of hypotheses, and what are the three parts of the wager a business places when it launches one?
2. What four things does a hypothesis need before a test begins, and why does the threshold have to be chosen before the result is seen?
3. What are the five rungs of the ladder of offer tests, and what can a one-page description prove that an interview cannot?
4. What is the difference between a pilot and a controlled experiment, and when can a small business randomize?
5. What does the case's sixty-day pilot show about its close rate, and what does it show that the close rate cannot express?

Answers. 1. An offer is a set of hypotheses because its contents, its promises and its prices rest on beliefs about what customers will do and what the business can deliver, and none of those beliefs has met a customer; the wager has three parts, which are who buys what, what customers value and what the business can deliver. 2. A hypothesis needs a measure that the business already counts, a threshold that separates a pass from a failure, the decision the threshold triggers and, where the risk is high, a kill criterion that ends the test; the threshold is chosen in advance because a rule written after the result is a description of what happened rather than a rule, and it can no longer fail. 3. The five rungs are interviews with customers, a one-page description of the offer, a letter or a landing page that measures interest before the service exists, a pilot with real customers at a real price, and a controlled experiment that compares randomly assigned groups; a one-page description shows whether customers understand the offer and which exclusions and terms they misunderstand, which an interview about what they value cannot show. 4. A pilot runs the real offer with real customers in a limited place or period and asks whether it works at all, while a controlled experiment assigns customers at random to the current offer and the new one and compares them on the same measures, which is what makes the difference attributable to the offer; a business can randomize when it has many small transactions and a channel that assigns them mechanically, and it cannot when it decides a handful of large contracts a year. 5. On ten proposals, six signatures give an observed close rate of 60% with an interval of about 30% to 90%, which contains the 57.9% of the strategy year, so the pilot cannot show that the tiers changed the close rate; it shows which customers could say yes, including two small buildings of the kind lost to the trades the year before, and what existing customers chose, with two upgrades to the Portfolio level and one downgrade.

Chapter References

Blank, S. (2013, May). Why the lean start-up changes everything. Harvard Business Review. https://hbr.org/2013/05/why-the-lean-start-up-changes-everything

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

Green, P. E., & Srinivasan, V. (1990). Conjoint analysis in marketing: New developments with implications for research and practice. Journal of Marketing, 54(4), 3–19. https://doi.org/10.1177/002224299005400402

Kohavi, R., & Thomke, S. (2017, September–October). The surprising power of online experiments. Harvard Business Review. https://hbr.org/2017/09/the-surprising-power-of-online-experiments

Kohavi, R., Tang, D., & Xu, Y. (2020). Trustworthy online controlled experiments: A practical guide to A/B testing. Cambridge University Press.

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: business/subscriptions, business/marketing/forms and productivity/tasks.

Chapter 9 — Typical Tasks, Cadence and Workflows

An offer is not finished on the day it is designed; it holds its shape only while somebody reads it, which is why this chapter puts the offer's upkeep on a calendar of short, dated reviews rather than in a document that is written once and filed.

What You'll Be Able to Do

• Describe the four rhythms that keep an offer designed — the weekly look at exceptions and credits, the monthly review of the measures, the quarterly offer review and the annual catalog review — and say what each one produces.
• Run the weekly and monthly rhythm in a little over an hour a month, using the exception register, the guarantee credits and a one-page sheet of measures the business already has the data to produce.
• Prepare the quarterly offer review by reading the quarter's measures, the exception register, the win-loss notes and the renewal comments, and by ending with a dated decision for every line.
• Plan the annual catalog review as part of the strategy refresh of Book A3 (Strategy and Competitive Positioning), keeping, changing or retiring each tier, add-on and productized service, and carrying the year's list of changes into the price review of Book A7 (Pricing Strategy and Revenue Models).
• Read complaints, requests, win-loss notes, take rates and guarantee credits as the offer's feedback, and tell the difference between a customer asking for something the offer does not contain and a customer asking for something the business has decided not to sell.
• Recognize the things that break the rhythm — a busy season, a large customer's demand, the owner in the field, a review that ends without a decision — and apply the rule that a skipped review is moved rather than canceled.

Why It Matters

An offer drifts the moment nobody watches it, and it drifts without a single bad decision. A customer asks for a quarterly meeting instead of a monthly report, the meeting is held once, and a promise has been made. A technician replaces a filter on a preventive visit because it is faster than explaining that parts are excluded, and the next customer expects the same. A tier is launched with three elements and ends the year with six, because each of the three additions was small and reasonable on the day it was granted. None of these events looks like a change to the offer. Together they are the whole of it, and by the time anybody notices, the business is selling something it never decided to sell, at a price set for something smaller.

The running case shows what the drift costs. The contract Nordec Entretien inc. sells was built by accretion: two elements in 2022 and four more added over the following years, each given to customers with no price change. The measured cost to serve a building in a territory came to $341.23 a month against $282.73 for the first two elements. The four additions cost $58.50 a month per building, which is $32,292 a year across the 46 buildings under contract, while the price rose from $691 to $712 and brought in $11,592 a year. The difference, $20,700 a year, was never priced, and the contract margin of a territory building fell from 59.08% to 52.08%. No one had decided to reduce the margin. The business had simply been granting small requests for three years without a place to put them.

The remedy presented in this chapter is deliberately small, because the remedy that survives a busy quarter is the only one that counts. Nordec's offer calendar takes the owner 43.5 hours a year, which is 2.1% of a 2,080-hour year and less than an hour a week: fifty weekly looks of fifteen minutes, twelve monthly reviews of an hour, four quarterly reviews of three hours and one annual catalog review of a day. That is the whole apparatus. It is enough to catch the exception in the week it is granted, to see a tier mix move before it has moved for a year, to decide the add-on that nobody is selling, and to retire a service that has stopped earning its place. It is also, and this matters for a business of seven people, small enough to be performed by one person who is already doing something else.

In This Chapter

Section 9.1 explains why an offer drifts and sets out the calendar that keeps it designed, with its cost to the owner and its place beside the strategy calendar of Book A3 (Strategy and Competitive Positioning). Section 9.2 describes the weekly look at exceptions and guarantee credits and the monthly review of the measures, with the table of the whole calendar. Section 9.3 describes the quarterly offer review as a workflow, from its inputs to its dated decisions, and reads the first exception register of the case. Section 9.4 covers the annual catalog review at the strategy refresh. Section 9.5 covers the light instruments by which a business listens to customers continuously. Section 9.6 covers the win-loss and take-rate reading, with the measures the case reads every month and what they showed after six months. Section 9.7 divides the offer tasks among three sizes of business. Section 9.8 names what breaks the cadence and what keeps it running anyway.

9.1 Offer Design as a Recurring Practice

An offer is a set of promises, and promises change quietly. Nobody sits down and decides to alter what the business sells; what happens instead is that a customer asks for something, somebody does it once, and the business has made a new promise without noticing. The promise then repeats, because the next customer hears about it from the first, and the cost of keeping it attaches itself to a contract whose price was set for a smaller offer. This is the drift that this book exists to prevent, and it is not a failure of design. The design was sound on the day it was made, and the person who made it was competent. What was missing was a habit of looking at the design again.

The contract that Nordec Entretien inc. sells grew exactly this way, and the sequence is worth reading closely because almost every small business can recognize itself in it. The contract began in 2022 with two elements: a monthly preventive visit, at a measured cost of $245.60 a month for a building in a territory, and urgent calls answered within ninety minutes with up to two hours of work included, at a measured cost of $37.13 a building a year. In the strategy year, three more elements were added to the contract without a price change: the upkeep of the building health record, at $13.50 a month; the one-page monthly report, at $27.00; and the capital planning note delivered at renewal, at $13.50. At the annual refresh that opens this book, the team decided to add a fourth, the mid-contract note for the buildings whose equipment the first note had flagged, at $4.50 a month for about half the buildings. Adding those four elements to the two original ones gives a cost to serve of $341.23 a month. The first two elements alone cost $282.73. The additions cost $58.50 a month per building, which is $32,292 a year for the 46 buildings under contract, and the price of the contract rose over the same period from $691 to $712, which brought $11,592 a year. The arithmetic of the gap is simple and uncomfortable: $20,700 a year of added cost was never priced, and the contract margin of a territory building fell from 59.08% to 52.08%.

Three doors let the drift in, and a business that watches only one of them will be surprised by the other two. The first door is the request granted. A customer asks for something small, the business says yes because the relationship matters and the cost looks negligible, and the yes becomes a precedent. The second door is the content added to keep a promise competitive: a report improved, a checklist lengthened, a response time shortened, each one a decision made in good faith and none of them brought back to the price. The third door is the price that stays still while the content grows, which is not a decision at all but the absence of one, and it is the door through which most of the $20,700 left the business.

The remedy is a calendar, and the calendar has one property that a document does not: it comes back. A designed offer is re-examined at four speeds, and each speed exists to catch what the others cannot see. The weekly look catches an exception while the person who granted it still remembers why. The monthly review catches a measure that has moved. The quarterly review catches a pattern and turns it into a decision. The annual catalog review asks the question that none of the others is allowed to ask, which is whether a line should exist at all. A business that performs only the annual review will spend the year granting exceptions it then has to unwind in a single day; a business that performs only the weekly look will keep every promise and price none of them.

The annual moment of this calendar already exists at Nordec, because Book A3 (Strategy and Competitive Positioning) gave the business a strategy calendar with an annual refresh. The offer calendar does not compete with it; it rides on it. The annual catalog review happens on the same day as the strategy refresh, for a practical reason: the two questions are the same question asked from two directions. The strategy asks where the business plays and how it wins, and the catalog asks what the customer is actually offered at that position. The three refusals signed in the strategy — no one-off emergency calls for buildings without a contract, no new buildings outside the two territories, and no bundled tenders — are refusals about the offer as much as about the market, and it is at the annual catalog review that they are checked against what the business has actually been selling.

Figure 9.1 draws the year as a listener would hear it described. The figure shows twelve monthly columns, one for each month of the year. Inside each column, a thin mark stands for the weekly fifteen-minute look at the exception register and the guarantee credits, and fifty of those marks run across the year. A larger block in each column stands for the monthly review of the measures. At the end of every third column, a wider block stands for the quarterly offer review, and there are four of them. In the first month, a full-width block stands for the one-day annual catalog review, held on the same day as the strategy refresh. Below the twelve columns, a single line gives the total, 43.5 hours, or 2.1% of a 2,080-hour year.

Figure 9.1 — The offer calendar of a small business. Fifty weekly looks, twelve monthly reviews, four quarterly reviews and one annual catalogue review: about 43.5 hours of the owner's year, or 2.1%.

Two habits make the calendar work, and neither of them is about time. The first is that a review which produces no change is a success rather than a wasted hour. Most monthly reviews confirm that nothing has moved, and a business that treats those months as failures will start looking for something to change in order to justify the meeting, which is how offers acquire features nobody asked for. The second habit is that the calendar is not a plan to redesign the offer four times a year. It is a plan to look at it, and the redesign, when it comes, is the work of one quarter in several. The rest of the year the reviews are short because the answers are short.

KEY CONCEPT An offer stays designed only if someone reviews it

An offer is not a document and it is not a decision; it is a standing arrangement between a business and its customers, and it is altered every time somebody says yes to a small request. That is why an offer cannot be designed once and left. It stays designed only if somebody reads it on a rhythm and writes down what the reading shows. The rhythm does not need to be elaborate, and the case that runs through this book shows how modest it can be: fifty weekly looks of fifteen minutes, twelve monthly reviews of an hour, four quarterly reviews of three hours and one annual catalog review of a day, which together take 43.5 hours of the owner's year, or 2.1%. What the rhythm buys is not a better offer on any given day. It buys the ability to say what the offer is on any given day, which is the precondition for changing it on purpose rather than by accident.

9.2 The Weekly and Monthly Rhythm

The weekly rhythm is a quarter of an hour and it has one job: to write down what happened before it becomes a habit. The owner opens two things, the exception register and the guarantee credit log, and reads them against each other. The exception register holds every request granted outside a tier, written with the customer, the date and the cost. The credit log holds every $150 credit the 90-minute guarantee produced, which is the record of a promise the business made and did not keep. Fifteen minutes, once a week, is enough for both, and the fifty weeks of an ordinary working year make 12.5 hours. The unit of work is deliberately small because the value of the look is entirely in its timing. An exception written down on the Friday of the week it was granted costs three minutes to record. The same exception reconstructed three months later costs an hour of somebody's memory and is usually reconstructed wrongly.

The monthly rhythm is an hour and it has a different job: to read the offer as a set of numbers rather than as a set of events. The owner sits down with one sheet that carries eight measures, described in section 9.6, and fills in the month's figures beside the previous month's. Tier mix, downgrades and upgrades at renewal, add-on attach rate, Building Checks sold and converted, guarantee credits, first-visit resolution, disputed invoices, and the count of requests for something no tier contains. The sheet is short on purpose. A sheet of eight measures that is filled in every month is worth more than a dashboard of forty that is opened twice a year, and the reason is not diligence but memory: the person filling it in is comparing this month against last month, and comparison is what turns a number into a signal.

The sheet is also the place where the exception register stops being a list and starts being a question. At Nordec, the exception register of the first quarter after the roll-out held seven lines, and those seven lines are read at the monthly review as a count before they are read at the quarterly review as decisions. Three of them were requests from Preventive customers for a quarterly review meeting. Three exceptions of the same kind in one quarter is not noise; it is a customer segment telling the business that something the Portfolio tier includes should be purchasable on its own. The monthly review produces the count, and the quarterly review produces the answer.

Two habits keep the monthly hour honest. The first is that the measures are read from records the business already keeps rather than from a special collection effort, because a measure that requires somebody to build a report will be dropped in the second month. The contract list gives the tier mix, the renewals give upgrades and downgrades, the invoices give the add-on sales, the Building Check log gives the entry offer, the credit log gives the guarantee, and the building records give first-visit resolution. None of the eight measures needs a new system. The second habit is that the sheet carries the month's date and is kept, because the value of the exercise after a year is the year's worth of sheets read together, and that is what the annual catalog review reads.

The quarterly rhythm is three hours and it is where the offer changes. It is described in section 9.3 as a workflow, and it is worth saying here what it is not. It is not a longer monthly review. The monthly review reads measures that the business already understands; the quarterly review decides what to do about them, which is a different activity requiring different inputs, including the win-loss notes and the comments gathered at renewals. Nor is it a meeting of everybody who touches the offer. At Nordec the quarterly review is the owner, the administrator and the lead technician, three people, three hours, four times a year.

The following table sets out the four rhythms of the calendar with their duration, their annual frequency, their annual hours and what each one produces. Read it as the offer's maintenance schedule: the four rows are the whole apparatus, the total is under an hour a week, and the column that matters most is the last one, because a rhythm that produces nothing written cannot be checked against the next one.

Table 9.1 — The offer calendar

The rhythm also has a boundary that a business should draw deliberately, because the boundary is what keeps the reviews short. The monthly review does not decide anything about the offer, and the weekly look does not analyze anything. Each rhythm hands its work to the next one and stops there. A monthly review that begins deciding whether to change a tier will take three hours and will therefore happen twice a year, and a business whose monthly review happens twice a year has a quarterly review that happens when somebody remembers. The speed of each rhythm is protected by the modesty of its purpose.

DEFINITION Take rate, attach rate, downgrade and exception register

The take rate of a tier is the share of the customers who were offered it and chose it, and it is read per tier rather than across the catalog, because a tier taken by everybody and a tier taken by nobody are two different problems. The attach rate is the share of the customers eligible for an add-on who buy it, and at Nordec it is read for the capital planning note sold to Essential customers and for the quarterly review sold to Preventive customers. A downgrade is a renewal in which a customer moves to a smaller tier, and an upgrade is a renewal in which a customer moves to a larger one; both are read at the renewal, not at the sale. The exception register is a dated list of every request granted outside the tiers, carrying the customer, the date and the cost of granting it, and it exists so that a pattern of exceptions can be seen before it has become a promise the business is keeping for free.

DO THE MATH What the offer calendar costs the owner

Illustrative figures; the case is fictional. Step 1 — The weekly look takes fifteen minutes, or 0.25 hour, and is performed fifty times a year: 0.25 × 50 = 12.5 hours. Step 2 — The monthly review takes one hour and is performed twelve times: 1 × 12 = 12 hours. Step 3 — The quarterly offer review takes three hours and is performed four times: 3 × 4 = 12 hours. Step 4 — The annual catalog review takes one day, counted as seven hours, and is performed once: 7 hours. Step 5 — The year's total is 12.5 + 12 + 12 + 7 = 43.5 hours, which is 43.5 ÷ 2,080 = 2.09%, or about 2.1%, of a 2,080-hour working year. Read another way, it is 50 minutes a week on average, and the largest single block in it is half a day of preparation and half a day of meeting, once, at the strategy refresh.

TOOLS OF THE TRADE What the offer calendar needs

Five functions carry the calendar, and a business should look for the functions rather than for a product. The first is a shared calendar holding the four rhythms as recurring appointments, visible to the people who attend, with the annual catalog review placed on the same day as the strategy refresh. The second is an exception register: one editable list, with a column for the customer, the date, what was asked, what was granted and what it cost, reachable by whoever grants exceptions. The third is a monthly measures sheet, one page with the eight measures of section 9.6 and a column for each month of the year, fed from records the business already keeps. The fourth is a catalog with the date of every change, so that the version of the offer sold in March can be described in September. The fifth is a customer comment log, where what customers say at renewals and on calls is written down in their words rather than summarized. None of the five needs a brand name attached to it.

9.3 The Quarterly Offer Review

The quarterly review is where the offer changes, and it works as a workflow rather than as a discussion. It has four declared inputs, five steps, and one output: a set of decisions, each with an owner and a date. The inputs are the three monthly measures sheets of the quarter, the exception register, the win-loss notes for the proposals decided in the quarter, and the comments gathered at renewals. Bringing the four inputs to the meeting is the whole of the preparation, and a business that arrives without one of them will spend the first forty minutes reconstructing it.

The reason the inputs are four rather than one is that each of them answers a different question. The measures sheets answer what happened. The exception register answers what customers asked for that the offer does not contain. The win-loss notes answer what the business lost and why, in the customer's words rather than in the salesperson's. The renewal comments answer what the people who stayed think, which is the input most often missing and the one that catches a tier that is quietly being tolerated rather than valued. A quarterly review that reads only the measures will make decisions about the offer the business already sells and will never learn what the business failed to sell.

The first exception register read at Nordec is a good illustration of how much information seven lines can carry. It held three kinds of request. Three Preventive customers asked for a quarterly review meeting, which is a meeting the Portfolio tier includes and the Preventive tier does not. Two customers asked for a technician at night for a non-urgent request, which is the after-hours coverage the business had deliberately refused in Book A3 (Strategy and Competitive Positioning), and which had cost it two proposals in the strategy year. Two customers asked for filter replacement, which the contract excludes because parts are not included, and which the service menu of Chapter 6 had already declined as a service to add.

The three requests are three different answers, and the review reached them in three different ways. The quarterly review meeting was offered to the three Preventive customers as the add-on designed in Chapter 4 rather than as a free exception to the tier, and two of the three bought it at $140 a quarter. That is the outcome the register exists to produce: a request that would have become an unpriced habit became a priced add-on, and the fact that two customers bought it confirms that the request was real rather than casual. The two after-hours requests were refused, and the refusal was recorded with its reason, which is the strategy's reduce, so that the next person asked the same question does not have to reopen the decision. The two filter requests were refused for the same reason the service menu had already given, and the register now carries that reason in writing.

The pattern that the seven lines show is worth naming, because it is the pattern a register is built to reveal. Four of the seven lines asked for something the business already had somewhere in its catalog but was not selling to that segment, and three of them asked for something the business had decided not to sell. The first group is a packaging problem and the second group is a positioning problem, and a business that keeps no register cannot tell them apart. It will treat every request as a packaging problem, which means it will keep adding to the offer, or it will treat every request as a positioning problem, which means it will keep refusing and will eventually lose customers who were asking for something both legitimate and profitable.

The meeting ends with the catalog, which is the document that lists every tier, every add-on and every productized service with the date of its last change. A decision that is taken and not written into the catalog has not been taken; it has been agreed, which is a weaker thing. The catalog is the input to the monthly measures sheet, because the measures are read against what the business believes it sells, and a catalog that is three months stale will make a healthy month look like a drift or a drift look like health.

The quarterly review is also the place where the case's honest reading of small numbers belongs. Chapter 8 showed that a sixty-day pilot with ten proposals and a 60% close rate carries a margin of error of roughly thirty points, which means the pilot cannot prove that the tiers raised the close rate. The same discipline applies to the quarterly review, and it applies to the exception register in particular. Seven lines in one quarter is a direction and not a finding, and a business that reorganizes its tiers around seven lines will do so again next quarter in the other direction. The rule the case adopted is that a pattern must appear in two consecutive quarters before a tier changes, and that a single quarter can produce an add-on, a refusal or a test.

METHOD The quarterly offer review in five steps

First, read the monthly measures of the quarter, three sheets side by side, and mark every measure that moved, without deciding yet what the movement means. Second, go through the exception register line by line, and for each line state what the customer asked for, what was granted and what it cost the business. Third, read the win-loss notes for the quarter's proposals and the comments written at renewals, and note the words customers used rather than the words the business would use. Fourth, decide each line: it becomes part of a tier, it becomes an add-on with a price, or it becomes a stated refusal recorded with its reason, and nothing is left as an exception to be granted again. Fifth, record every decision with an owner and a date, update the catalog the same day, and confirm the next quarterly review in the calendar. Deliverable: a dated decision for every line of the register and every measure that moved. Pitfall: a review that ends with agreement and no owner, which produces the same conversation three months later.

WATCH OUT The exception granted once

An exception granted once becomes a precedent the day the customer tells somebody. The mechanism is not mysterious: a customer says the business did it last spring, and refusing now requires an explanation the person on the call does not have. The exception then gets granted again, more cheaply each time, until it is a feature of the tier and the price of the tier was never set for it. The rule that closes the door is not to refuse exceptions; it is to write them down the same day, with the customer, the date and the cost. Writing them down does three things at once. It makes the cost visible to the person who granted it, which is often enough to end the practice. It gives the quarterly review a line to decide rather than a memory to argue about. And it lets the business spot the exception that is worth turning into an add-on, which is what happened to the quarterly meeting that three Preventive customers asked for in Nordec's first quarter, and which two of them then bought.

CASE IN POINT Nordec Entretien inc.: seven exceptions in a quarter

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The register was opened on the day the three tiers and two add-ons were announced, and it was read for the first time at the quarterly offer review three months later, with seven lines on it. Three Preventive customers had asked for a quarterly review meeting. Two customers had asked for after-hours service on a non-urgent request. Two had asked for filter replacement. The review took forty minutes over the seven lines and produced three decisions. The meeting was offered as the quarterly review add-on at $140 a quarter, and two of the three bought it, which turned a free exception into $280 a quarter of new revenue and a genuine test of whether the add-on would sell. The after-hours requests were refused, with the reason written beside them, because the business had closed that path deliberately a year earlier. The filter requests were refused for the same reason parts are excluded from every tier. Every line ended as a tier, an add-on or a statement, and none of them stayed an unwritten habit.

9.4 The Annual Catalog Review

The annual catalog review is one day, and it is the only occasion in the year when the business is allowed to ask whether a line should exist. Everything about the day is deliberately slow. The weekly look records, the monthly review counts, the quarterly review decides what to do about what was counted, and the annual day decides what the catalog should contain next year. It is the offer's equivalent of the annual inventory, and like an annual inventory it is worth holding on a fixed date rather than when the question becomes urgent.

The day has four tasks, and the first is the reading of the year's changes. The administrator prepares a list of everything added to the offer during the year with the date of each addition and what it cost, and everything removed with the date of its removal. At Nordec, that list for the first year after the roll-out carried the three tiers and the two add-ons, the Building Check with its $950 price and its $507.60 cost, the 90-minute guarantee with its $150 credit, and the first exception register. Read as a list, it is the year's design work in one place, and it is also the document that Book A7 (Pricing Strategy and Revenue Models) needs, because a price review cannot set prices for contents nobody has enumerated.

The second task is the decision on each line, and it has three possible outcomes. A line is kept, which means it stays in the catalog unchanged and its price is carried into the price review. A line is changed, which means the content, the exclusions or the promise around it is altered, and the alteration is written into the catalog with its date. A line is retired, which means it stops being sold and the customers who hold it are given a route to something else. The word retired is doing quiet work in that sentence. Most small businesses do not retire offers; they simply stop mentioning them, and a line that is no longer sold but still held by customers is a cost with no revenue and no owner.

The third task is the light repeat of the value interviews of Chapter 2. Those interviews were the heavy instrument: twenty conversations with owner-operators, property managers and owners who had declined, held once, at length. The annual day does not repeat them at that scale. It repeats the essential question with a handful of customers, and the question is the one Chapter 2 asked in the customer's own words: what job did you hire this for, and what would you do if it did not exist. The answers are compared with the year's answers rather than read on their own, because a job that customers described in the same terms twelve months ago and describe differently now is a job the offer has stopped doing.

The fourth task is the preparation of the price review. The offer calendar carries the year's list of changes, with dates and costs, into the strategy refresh, and from there into Book A7. The division of labor between the two books is exact and worth stating plainly: this book designs the contents, the exclusions and the promises, and Book A7 sets the prices. The annual catalog review is the point where the two meet, and the reason the meeting is placed there is that a price decision taken without the year's list of changes is a price decision taken from memory.

The annual day also carries forward the strategy's refusals, which is a task that belongs at this moment rather than in the quarterly review. Nordec's strategy closed the wider path of energy-efficiency retrofit projects unless at least five owners asked for replacement projects after reading their capital planning note. Four had asked, and the path stayed closed. The annual catalog review is where that count is re-read, because the count is a condition on the catalog rather than a decision about a customer, and because a refusal whose condition is never checked is not a refusal but a delay. The same reading applies to the three refusals signed a year earlier: the business checks that it has not in fact been selling one-off emergency calls to buildings without a contract, or taking on new buildings outside the two territories, or bidding on bundled tenders.

QUÉBEC & CANADA NOTE Keeping the records a review needs

The rhythm described in this chapter is sized for the firm that will actually perform it, and the Canadian numbers make that plain. Small businesses made up 98.2% of employer businesses in Canada in December 2024, and small and medium-sized businesses together produced 47.8% of private-sector GDP on average over 2018 to 2022 (Innovation, Science and Economic Development Canada, 2026a). Almost no business in that population has a product manager, a pricing team or a customer insight function, and the four rhythms of this chapter assume that the owner is the person doing the work. That assumption has a consequence for record keeping. A review is only as good as the record it reads, and the records this calendar needs are unglamorous: the contract list with the tier of each building, the invoices that carry the add-on sales, the call log that carries the response times, the credit log, and the exception register. In a business of seven people those records are kept by the administrator, and the discipline of the calendar is mostly the discipline of keeping them current enough to be read on the first working day of the month.

WHAT CHANGED Nordec's offer since the refresh

The offer of the case changed on four dated occasions in the year this chapter covers. After the annual refresh, the contract that had been a single monthly offer became three tiers — Essential at $490, Preventive from $725 for new customers with current customers keeping $712, and Portfolio at the Preventive price plus $175 per building — with two add-ons, the capital planning note at $420 a year for Essential customers and the quarterly review at $140 a quarter for Preventive customers. The Building Check was introduced as a productized entry offer at a fixed $950, credited in full against the first three months if a contract was signed within sixty days. The 90-minute guarantee was announced to every tier, with a $150 credit applied automatically to the next invoice when a technician was not on site within ninety minutes. And the first exception register was opened on the day of the announcement and read for the first time at the quarterly offer review three months later, where its seven lines became two add-on sales, two written refusals and a stated position on the third request.

9.5 Listening to Customers Continuously

The quarterly review reads what the business wrote down; the listening described in this section is what produces it. Listening continuously does not mean surveying customers constantly, which is the mistake a business makes when it decides to become customer-focused. It means that the conversations the business is already having are treated as data, and that the data is written down in the customer's words rather than in the business's. A service business has more of these conversations than it realizes: at renewal, on every call, at the end of a visit, in the complaint that arrives by email, and in the request that arrives at the dispatcher's desk.

The renewal conversation is the best of the light instruments because it happens anyway. At Nordec, the renewal is the moment a customer is offered the three tiers, and the conversation that follows produces three pieces of information the offer calendar uses. The first is the reason for the choice, which is the sentence a customer says after the answer and which is worth writing down verbatim. The second is the list of what the customer would need in order to move up, which is different from what the customer asks for. The third is the comment about the year just finished, which is where a tier that is being tolerated rather than valued will show itself. None of the three requires a survey; all three require that somebody writes them into the comment log before the next appointment.

The guarantee credits are the second light instrument and the most objective, because they are not opinions. The 90-minute guarantee of Chapter 5 puts a $150 credit on the next invoice, automatically and with no claim to make, whenever a technician is not on site within ninety minutes of an urgent call. Every credit is therefore a measured failure of the delivery promise, recorded without anybody having to admit anything. In the year before the redesign, 96% of urgent contract calls were answered on site within ninety minutes, which is about ten misses a year at 253 calls, and $1,500 of credits. In the six months after the refresh, 138 urgent calls produced five missed calls and $750 of credits, which is 96.4%. Read over time, the credit log is the business's own report on whether the promise it advertises is the promise it keeps, and it is worth reading in the weekly look rather than at the quarterly review, because a credit that is explained on the Friday of the week it happened is a fact, while the same credit in a quarterly pile is a number.

The third instrument is the blueprint of Chapter 3, and it is used in a different way. The blueprint maps an urgent call from the customer's first contact to the closing note, marking the line of interaction, the line of visibility and the line of internal interaction, and it identifies the moments of truth where the customer forms a judgment. Those moments are where listening pays. A business that has drawn its blueprint knows which four or five points on the journey decide whether the customer's memory of the service is good, and it can place its questions there rather than everywhere. At Nordec, the redesign of Chapter 3 added two moments of truth deliberately: the technician calls the owner with an estimate before exceeding the two hours a tier includes, and the closing note is written into the building record on site, in front of the customer. Neither of those is a listening instrument in itself, and both of them generate the comments the comment log needs.

What listening does not do is turn every comment into a change. The distinction that keeps the calendar sane is the one between a comment and a pattern. A comment is a customer saying something once and it belongs in the log, where it is read at the monthly and quarterly reviews as part of a count. A pattern is the same thing said by several customers, or by one customer twice, and it is a candidate for a decision. Nordec's exception register made this distinction visible in its first quarter: three requests for a quarterly review meeting became an offer of an add-on, because three customers asking for the same meeting is a pattern, while the two after-hours requests were refused, because they were a pattern the business had already decided against in its strategy. The two refusals were not a failure to listen. They were listening followed by a decision.

The last thing to say about continuous listening is that it must not become an excuse for holding the reviews. A business that gathers comments constantly and decides nothing has replaced a calendar with a folder. The comments and the credits and the requests all travel to the same two places, the monthly sheets and the exception register, and the quarterly review decides them. That is the whole circulation: the conversations produce the lines, the lines produce the decisions, and the decisions produce the next version of the catalog.

WHO DOES THIS The administrator keeps the register

In Nordec, the exception register is kept by the administrator, and the choice is not arbitrary. The administrator answers the telephone, writes the invoices and sends the monthly reports, which means every request that arrives outside a tier passes across that desk before it reaches anybody who could grant it. The register is opened on the administrator's screen, it is filled in on the day of the request, and it carries the customer, the date, what was asked and what it would cost. The dispatcher contributes the calls that come to the dispatch line, the lead technician contributes what customers said on site, and the owner reads the register every week in the fifteen-minute look. Nobody needs a title for the work; the work attaches itself to whoever already holds the information. In the smallest businesses the owner keeps the register, and in that case the discipline that matters is writing the line down before deciding whether to grant it, because the person who grants an exception and the person who records it being the same person is exactly how exceptions become habits.

9.6 Win-Loss and Take-Rate Reviews

The win-loss reading is where the offer meets the market, and it is short enough to perform in the quarterly review if the notes have been written as they happened. In the strategy year, Nordec decided 19 proposals: 11 were won, which is 57.9%, four were lost on price, two were lost on after-hours coverage, and two ended without a decision. The four lost on price are worth reading carefully, because two of them went to the trades billed by the call and both were small buildings under 3,500 square metres whose owners said the contract was more than the building needed. That single sentence, said by two owners in the same year, is what produced the Essential tier in Chapter 4, and it is worth noticing that it was available in the loss notes all along and was only read as a pattern at the offer review.

The reading by tier is the second half of the exercise, and it cannot be performed in the first months of a tiered offer because there is nothing to read yet. Six months after the refresh, Nordec had 52 buildings under contract: four on Essential, forty-two on Preventive and six on Portfolio. The pilot that preceded the roll-out had produced ten proposals with the three tiers in front of them, of which six signed: two Essential, both small buildings of exactly the kind lost to the trades the year before, three Preventive and one Portfolio. The reading said something the aggregate close rate did not. The 60% close rate against 57.9% for the strategy year was within the margin of error of ten proposals, which Chapter 8 put at roughly thirty points. What the pilot actually showed was who could say yes and what existing customers chose when the tiers were put in front of them.

The renewal reading is where the offer is tested hardest, because a renewal is a customer who already knows what the business does and is deciding whether to keep buying it. In the pilot window, eight renewals produced one downgrade, five customers who stayed on Preventive and two who moved up to Portfolio. Six months after the refresh the picture was one downgrade and five upgrades, and the tier mix was moving in the direction the design had intended. A downgrade is not a failure and should not be read as one. At Nordec the single downgrade was a building whose owner moved to Essential at renewal, and the correct reading is that the business kept a customer it would otherwise have lost to the trades, at a lower price and at a lower cost to serve. The mistake a business makes with downgrades is treating them as a wound; the mistake it makes in the other direction is failing to notice that a downgrade is sometimes the only alternative to a departure.

The third reading is the one that concerns the entry offer. The Building Check was offered by letter and by telephone to forty buildings in the two territories that were not under contract, five bought it, and two signed a contract within the pilot period. Six months after the refresh, eleven Building Checks had been sold for $10,450 and four of them had converted to contracts. The measure to read here is not the revenue the Building Checks bring, which is small against a contract run-rate of $447,012 a year, but the conversion: a fixed-price assessment that credits its fee against the first three months is a way of buying a conversation with a building that would not otherwise take a call, and four conversions out of eleven is the number that justifies the effort or does not.

The following table sets out the eight measures Nordec reads every month, what each one shows, where the figure comes from, and what the measure read six months after the refresh. It is deliberately short, and the last column is deliberately partial: several of the measures were read against the year before the redesign rather than against a six-month figure, and the table says so rather than inventing a comparison.

Table 9.2 — The measures of the offer, read every month

The reading of these eight measures together is a skill that takes a few quarters to acquire, and its first rule is to read the measures in pairs. The tier mix read alone says nothing; read beside upgrades and downgrades it says whether customers are moving toward the richer offer or away from it. The guarantee credits read alone are a quality number; read beside first-visit resolution they say whether the business is losing time on the road or on the job. The add-on attach rate read alone is a sales number; read beside the exception register it says whether the customers were asking for the add-on before it existed. A business that reads eight measures one at a time will draw eight conclusions, and a business that reads them in pairs will draw two or three.

THE RULE Every exception ends in a decision

The rule this chapter sets for the exception register is short: within the quarter, every exception ends as part of a tier, as an add-on with a price, or as a stated refusal recorded with its reason, and none of them stays an unwritten habit. The rule is stated with a deadline of one quarter because a register that is carried from quarter to quarter without decisions is a list of the business's good intentions. It has three outcomes rather than two on purpose. A business that allows only two outcomes will either fold every request into a tier, which is how the case's contract acquired $20,700 a year of unpriced additions, or refuse every request, which is how a business loses customers who were asking for something it would have been happy to sell. The third outcome, a priced add-on, is the one that turns a customer's request into a line of revenue, and the case produced exactly that in its first quarter when two Preventive customers bought the quarterly review add-on they had first asked for as a free meeting.

9.7 Who Does What in Three Sizes of Business

The offer calendar does not change with the size of the business; what changes is how many people touch it and which of them is allowed to decide. In a solo business, one person performs every rhythm and is the single point of failure for all of them. In a small team, like the seven people of Nordec, the rhythms divide between the owner, the administrator, the dispatcher and the lead technician. In an established firm of fifty people or more, the offer has an owner who is not the general manager, and the rhythms are performed by a product or offer manager, with a finance lead reading the numbers and a customer service lead reading the comments.

The solo owner-operator runs the same calendar with one difference: the rhythms are shorter and the register is kept by the person who grants the exceptions. The weekly look is ten minutes rather than fifteen, because a one-person business has fewer lines to read and remembers more of them. The monthly review is an hour and the quarterly review can be two hours rather than three, because there is nobody to consult. What the solo business cannot do is skip the register, because the register is the only thing standing between a helpful owner and an offer that has grown without a price. The solo owner who writes the exception down before deciding whether to grant it has the whole discipline, and a spreadsheet is enough to hold it.

The small team is the size at which the calendar needs named roles, because two people granting exceptions without a shared register is the fastest way to an offer nobody can describe. At Nordec the division is clear. The administrator keeps the exception register and produces the monthly measures sheet from the contract list, the invoices and the call log. The dispatcher writes into the register the requests that arrive on the dispatch line, because the dispatcher hears them first. The lead technician reports what customers said on site and what was found during the preventive visit, which is where the requests for services the business has refused tend to originate. The owner performs the weekly look and the monthly review, chairs the quarterly review and holds the annual catalog review on the day of the strategy refresh.

The established firm of fifty people or more has a division of labor that mirrors the smaller one with more hands. An offer or product manager owns the catalog, keeping the date of every change and preparing the year's list for the price review. A finance lead reads the take rates, the add-on attach rate and the revenue per building, and owns the arithmetic of cost to serve that Chapter 4 uses to check a tier's margin. A customer service lead owns the comment log, the complaint record and the win-loss notes, and brings them to the quarterly review in the customer's words. The general manager or the owner holds the annual catalog review, because the decision to retire a line is a decision about direction rather than about execution, and it belongs with the person who holds the strategy.

The difference between the small team and the established firm is not the volume of work but the distance between the person who hears the customer and the person who decides. In a team of seven that distance is one conversation, and the owner can hear a request the same day it is made. In a firm of fifty the distance is a report, and the risk is that the report carries the business's summary of what the customer said rather than what the customer said. The remedy is the same in both: the customer's own words go into the log, and the summary is written by whoever reads the log rather than by whoever heard the customer.

The following table sets out eight offer tasks and who performs each of them in a solo business, in a small team and in an established firm. Read it as a distribution of responsibility rather than a hierarchy: the tasks are the same in all three columns, and the difference is whether one person carries several of them or several people carry one each.

Table 9.3 — The offer tasks in three sizes of business

Two features of the table are deliberate. The first is that the fourth row is the one that concentrates in the most senior hands as the business grows, because keeping or retiring a line is a decision about where the business is going and it is not delegable to the person who runs the reviews. The second is that the last row is the row on which every other row depends, because a catalog without dates cannot be read against a year's measures, and the business that cannot say what it was selling in March cannot explain what happened in September.

IN LA MACHINE The reviews and the register in the platform

Four pages of the application copy read for this book can carry part of this chapter's calendar (La Machine, 2026). The four rhythm meetings can live in Meetings at productivity/meetings and in the Calendar at productivity/calendar, which is enough to put the weekly look, the twelve monthly reviews, the four quarterly reviews and the annual catalog review in front of the people who attend them. The sales by package can be read at Account → Profile → Revenues & Expenses at business/subscriptions, which shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period and an export to a spreadsheet: the place to count which package sold, and how often, without building a report by hand. The exception register can be carried as a recurring task at productivity/tasks, one line per exception, which gives it a date and a place without turning it into a system. Book N2 (Piloting Tasks, Executions and Projects) describes the tasks pages, and Plan tasks at /taches gives each task a verification criterion, which is where a register line becomes a decision with an owner. What the platform does not have belongs in the same breath: there is no service-level tracking and no service credit register, so the guarantee credits of Chapter 5 are counted outside it, and there are no tiers, bundles or add-ons for the user's own packages, because each package is a single, flat item.

9.8 What Breaks the Cadence

Four things break the offer calendar in a small business, and each of them is ordinary rather than dramatic. The first is the busy season, when the work that pays arrives all at once and the reviews compete with billable hours. The second is a large customer's demand, which is different from a busy season because it is one relationship rather than a season, and because the business is more reluctant to say no to it. The third is the owner spending the month in the field, which removes the person who is supposed to read the register. The fourth is not an interruption at all but a slow failure: a review that happens and decides nothing, or a review that happens without the data it needs. The first three break the schedule and the fourth breaks the purpose, and the fourth is the more dangerous of the two.

The busy season is the break every business expects and few plan for, and it is worth noticing that the season is also when exceptions multiply, because customers ask for more when they are under pressure. The remedy is decided in a quiet month and applied in a busy one. The weekly look is protected and shortened: ten minutes rather than fifteen, reading only the register and the credits. The monthly review moves rather than disappears. The quarterly review keeps its date and drops to two hours, reading the measures and the register and postponing nothing to the following quarter. What the business must not do is reduce everything at once, because a rhythm that has stopped for three months has to be rebuilt rather than resumed, and rebuilding it costs more than the three hours it saved.

A large customer's demand is the break that arrives with a legitimate reason attached. The customer is important, the request is urgent, and the exception is granted in a telephone call by somebody who has the authority to grant it and not the authority to change a tier. The remedy is not to refuse and it is not to escalate every request to the owner. It is the register, written the same day, with the customer, the date and the cost, and read at the next weekly look. At Nordec the mid-contract capital planning note was born exactly this way: the first note flagged equipment approaching the end of its life, customers asked what to do about it, and the answer arrived inside a contract period rather than at renewal. The request was legitimate, it recurred, and the quarterly review turned it into a portfolio element rather than a habit.

The owner in the field is the break that a named second reader prevents, and the second reader has to exist before the absence. The second reader's job is narrow: perform the weekly look and the monthly review, using the register and the sheet, and leave every decision to the person who owns it. The lead technician or the dispatcher can perform the weekly look in ten minutes, because reading a register and a credit log requires no authority, only access. What the second reader must not do is decide whether an exception becomes a tier, which is the work of the quarterly review. A business that installs the second reader changes who reads, not what is read, and the register is a continuous record either way.

The fourth break is the one with no obvious cause and the worst effect. A review that happens without decisions and a review that happens without data look identical from the outside, because both produce a meeting that ended on time. The first remedy is the record of the meeting: a dated decision with an owner, written into the catalog or into the task list, so that the next review can read what the last one decided. The second remedy is the measures sheet, which is the reason the case reads eight figures that the business already produces rather than a dashboard that requires somebody to build one. A business that cannot fill in its own measures sheet has designed a review it cannot perform.

Nordec met the first break once, in the first cold week of the heating season, and the story is worth keeping because it produced the rule the business still uses. The monthly review of that month was skipped outright, and the quarterly review that followed read two months of measures together. Nothing was lost, and the cost was visible anyway: the pair of months had to be compared against a single earlier month, and the tier mix reading of that quarter was weaker than the readings around it. The rule the team adopted is the rule of the strategy calendar of Book A3 (Strategy and Competitive Positioning), and it is one sentence: a skipped review is moved, never canceled. A review moved to the following Tuesday keeps the year's record intact; a review canceled leaves a gap that no later review can fill, because the month it would have read is gone.

The thread that runs through all four breaks is that the calendar must not depend on the memory of the person who keeps it. The dates are on a shared calendar, so a busy season changes what happens at a review rather than whether it happens. The register is written by whoever hears the request, so an absence changes who reads it. The measures come from records the business already keeps, so a review without data is a review that has stopped using its own instrument. And every review ends with a dated decision, so the next review has something to read. A business with those four things has an offer it can describe on any day of the year. A business without them has an offer that is whatever the last exception made it, and a price set for something else.

Costly Mistakes

• Keeping the designed offer in a document instead of on a calendar. An offer is a standing arrangement rather than a finished document, and a design that is written once and filed will be changed by the next exception rather than by the next review.
• Granting an exception without writing it down the same day. An exception recorded on the Friday of the week it was granted costs minutes; the same exception reconstructed a quarter later costs an hour and is usually reconstructed wrongly.
• Reading a long dashboard twice a year instead of eight measures every month. The monthly sheet is short because the person reading it is comparing this month with last month, and comparison is what turns a number into a signal.
• Ending a quarterly review with agreement and no owner. A decision without a name and a date beside it is the same conversation scheduled for three months later, and the catalog will still not carry the change.
• Treating every customer request as a packaging problem. A business that folds every request into a tier is the business that absorbed $20,700 a year of unpriced additions; a business that refuses every request loses customers asking for something profitable.
• Reading a downgrade at renewal as a wound. The single downgrade in the case kept a building that would otherwise have moved to the trades billed by the call, at a lower price and a lower cost to serve.
• Canceling a review instead of moving it. A busy season is a reason to shorten the weekly look and to move the monthly review, never a reason to lose a month that no later review can read.

Key Takeaways

• An offer drifts without a single bad decision, through the request granted, the content added and the price that stays still, and the contract of the case grew to $341.23 a month of cost to serve against a price that had risen from $691 to $712.
• The offer calendar is four rhythms: fifty weekly looks of fifteen minutes, twelve monthly reviews of an hour, four quarterly reviews of three hours and one annual catalog review of a day, which is 43.5 hours and 2.1% of a 2,080-hour year.
• The weekly look reads the exception register and the guarantee credits, and the monthly review fills in eight measures the business already has the data to produce: tier mix, upgrades and downgrades, add-on attach rate, Building Checks sold and converted, guarantee credits, first-visit resolution, disputed invoices and requests for something no tier contains.
• The quarterly review reads the quarter's measures, the exception register, the win-loss notes and the renewal comments, and ends with a dated decision for every line: part of a tier, an add-on with a price, or a stated refusal, as the first register of the case showed when two Preventive customers bought the quarterly review add-on they had first asked for as a free meeting.
• The annual catalog review happens on the day of the strategy refresh and gives every tier, add-on and productized service a kept, changed or retired status, and it prepares the year's list of additions, with dates and costs, for the price review of Book A7 (Pricing Strategy and Revenue Models).
• A skipped review is moved, never canceled, and the record of every review is a dated decision with an owner.

Your Action Plan

1. Open the exception register and write the first line today. Create one editable list with a column for the customer, the date, what was asked, what was granted and what it cost, and enter the next exception before it is answered rather than after. Deliverable: one register with at least one dated line in it. Verification: the person who grants exceptions can reach the register in under a minute and the owner can read it in the weekly look.
2. Put the four rhythms on a shared calendar for the next twelve months. Place the weekly look of fifteen minutes, the twelve monthly reviews of an hour, the four quarterly reviews of three hours and the annual catalog review of a day, with the annual day on the same date as the strategy refresh. Deliverable: dated entries totaling 43.5 hours. Verification: the entries are on the calendar of the person who performs them, and the annual catalog review shares its date with the strategy refresh.
3. Build the one-page measures sheet and fill it in once. Put the eight measures of section 9.6 on a single page with a column for each month, take the figures from the contract list, the invoices, the call log and the register, and fill in the current month beside the previous one. Deliverable: one dated measures sheet. Verification: every figure on the sheet can be traced to a record the business already keeps without any new collection.
4. Hold one quarterly offer review and decide every line of the register. Read the measures, walk the register line by line, read the win-loss notes and the renewal comments, and end every line as part of a tier, an add-on with a price or a stated refusal. Deliverable: a dated decision for every line, with an owner. Verification: the catalog carries each decision with its date, and no line of the register is left open.
5. Prepare the annual catalog review and name what will be kept, changed or retired. Before the strategy refresh, list every tier, add-on and productized service; for each one, state what it costs, when it last changed and what the year's measures showed about it. Deliverable: the year's list of changes, with dates and costs, ready for the price review. Verification: every line of the catalog has a kept, changed or retired status, and the list is handed to whoever holds the price review of Book A7 (Pricing Strategy and Revenue Models).

In La Machine

Four routes in the application copy read for this book can carry part of the offer calendar (La Machine, 2026). The four rhythm meetings fit naturally in Meetings at productivity/meetings and in the Calendar at productivity/calendar, which is enough to place the weekly look, the twelve monthly reviews, the four quarterly reviews and the annual catalog review in front of the people who attend them, with a date and a duration attached. The sales by package can be read at Account → Profile → Revenues & Expenses at business/subscriptions, which presents three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet; that is the place to count which package sold and how often without building a report by hand, and it is the closest the platform comes to the take-rate reading of section 9.6. The exception register can be carried as a recurring task at productivity/tasks, one line per exception, and Plan tasks at /taches gives each task a verification criterion, which is where a register line acquires an owner and a check. Book N2 (Piloting Tasks, Executions and Projects) describes the tasks pages.

What the platform does not do belongs in the same breath, because a business should not design a calendar around functions it does not have. There is no service-level tracking and no service credit register in the application copy read for this book, so the 90-minute guarantee of Chapter 5 is monitored outside the platform, from the call log and the invoices, and the credits are counted by hand. There are no tiers, bundles or add-ons for the user's own packages, since each package is a single, flat item, so the three-tier catalog of Chapter 4 lives in a document rather than in the platform, together with the two add-ons and the Building Check. There is no survey tool beyond a general form builder, so the light listening of section 9.5 is done in conversations whose notes are written into the comment log, and there is no customer journey map and no service blueprint, so the blueprint of Chapter 3 is drawn and kept outside the platform as well.

Self-Check

1. Why does an offer drift even when every decision was reasonable on the day it was made, and what are the three doors through which the drift enters?
2. What are the four rhythms of the offer calendar, how long does each take, and what does the whole calendar cost the owner in a year?
3. What are the inputs to the quarterly offer review, what are its five steps, and what does it produce?
4. What is the exception register for, and what are the three possible outcomes for each line in it?
5. What does the annual catalog review decide, and why is it held on the same day as the strategy refresh?

Answers. 1. An offer drifts because it is a set of promises that changes one small request at a time, without any single bad decision: the request granted becomes a precedent, the content added to keep the offer competitive is not brought back to the price, and the price stays still while the content grows, which is the absence of a decision rather than a decision. 2. The four rhythms are the weekly look at the exception register and the guarantee credits, fifteen minutes, performed fifty times for 12.5 hours; the monthly review of the eight measures, one hour, twelve times for 12 hours; the quarterly offer review, three hours, four times for 12 hours; and the annual catalog review, one day counted as seven hours, once. The total is 43.5 hours, which is 2.1% of a 2,080-hour year, or about fifty minutes a week. 3. The inputs are the three monthly measures sheets of the quarter, the exception register, the win-loss notes for the proposals decided in the quarter and the comments gathered at renewals; the five steps are reading the quarter's measures, walking the register line by line, reading the win-loss notes and the renewal comments, deciding each line as part of a tier, an add-on with a price or a stated refusal, and recording every decision with an owner and a date while updating the catalog; and the output is a set of dated decisions, one for every line and every measure that moved. 4. The exception register records every request granted outside the tiers, with the customer, the date and the cost, so that a pattern can be seen before it becomes a promise the business keeps for free. Within the quarter, each line ends as part of a tier, as an add-on with a price, or as a stated refusal recorded with its reason, and none of them stays an unwritten habit. 5. It gives every tier, add-on and productized service a kept, changed or retired status, repeats the value interviews of Chapter 2 in a light form with a handful of customers, checks that the strategy's refusals are still being honored and that their conditions are still met, and prepares the year's list of additions, with dates and costs, for the price review. It is held on the day of the strategy refresh because the strategy asks where the business plays and how it wins, and the catalog asks what the customer is actually offered at that position, which is the same question asked twice.

Chapter References

Innovation, Science and Economic Development Canada. (2026a). Key small business statistics 2025 (page modified April 23, 2026). Retrieved September 22, 2026, from https://ised-isde.canada.ca/site/sme-research-statistics/en/key-small-business-statistics/key-small-business-statistics-2025

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/meetings, productivity/calendar, business/subscriptions and productivity/tasks.

Chapter 10 — Expertise, Roles and When to Get Help

In a small business the owner designs the offer, but no owner has every skill the work requires — listening to customers without leading them, drawing a service a technician can follow, costing a tier, wording a guarantee that a customer and a court would read the same way — and knowing which help to buy, for which deliverable, is part of the design.

What You'll Be Able to Do

• Name the eight kinds of work behind an offer — deciding, listening, designing the service, costing, wording the promises, selling, delivering and measuring — and say who performs each one in a business of your size.
• Distinguish the owner, who decides what the firm promises and refuses, from the offer owner or offer coordinator, who keeps the catalog, the exceptions and the measures coherent.
• Judge when a service designer, a customer experience specialist or a pricing specialist is worth hiring, and what deliverable to ask for when one is.
• Brief a lawyer, an insurer and an accountant on the parts of an offer that belong to them, and know what each one should hand back.
• Build a small customer panel that reads the offer before the pilot, and state plainly what the panel may decide and what it may not.
• Write a scope for outside help with a named deliverable, a fixed fee or a cap, and a check on the adviser's independence from the products they might recommend.

Why It Matters

An offer is a promise, and a promise is only as good as the weakest piece of work behind it. A firm can listen carefully, design a beautiful service and still lose money on every contract because nobody costed the tier. It can cost the tier precisely and still lose the customer at the first urgent call because the guarantee was worded in a way that promised more than the team could deliver. It can deliver flawlessly and still be unable to sell, because the one-page description of the offer was written by the person who designed it rather than by the person who has to explain it. Each of those failures belongs to a different kind of work, and each of them is invisible from inside the kind of work that is being performed well.

The reason this chapter exists in a book about the content of the offer is that the content is decided by people, and the people are usually too few. A small business does not have a service design department, a pricing department, a legal department and a quality department. It has an owner and a handful of colleagues, and the eight kinds of work are distributed among them by necessity rather than by design. That distribution is itself a design decision, and it is one that most small businesses never make explicitly. The result is a set of roles that exist in practice and nowhere else: the person who always ends up writing the proposal, the person who always ends up calming the customer, the person who is assumed to be keeping the records and is not.

The case followed through this book shows what happens when the distribution is made explicit and when it is not. Before the first annual refresh, Sophie Nadeau did all eight kinds of work herself, and the offer grew by accretion: six elements added over two years, four of them given away free, and $20,700 a year of added cost that was never priced. Nothing was done badly. The listening was done by the person who knew the buildings best, which is the person least able to hear an answer they did not expect. The costing was done from memory. The wording of the guarantee was done in the same afternoon as the decision to offer it. After the refresh, four people held named roles, three pieces of outside help were bought for named deliverables, and a panel of five customers read the offer before it was piloted. The offer did not become more complicated. It became checkable.

Buying help is not a sign that the business cannot do the work, and refusing help is not a sign of discipline. The question is narrower and more practical: for this deliverable, at this moment, is the work better done inside or bought, and what exactly will be handed back? A lawyer asked to review three paragraphs of a guarantee is a small, bounded purchase with a clear output. A consultant asked to "look at the offer" is an open commitment with no output at all, and it is the second kind of engagement that gives outside help a bad name in small businesses. This chapter is about telling the two apart before the money is spent.

In This Chapter

Section 10.1 sets out the eight kinds of work behind an offer and shows how a small business combines them in a few people. Section 10.2 distinguishes the owner, who decides, from the offer owner or offer coordinator, who keeps the catalog coherent. Section 10.3 covers service designers and customer experience specialists, what they produce and when a small business needs one. Section 10.4 covers pricing and packaging specialists, and the line between the packaging this book designs and the pricing Book A7 sets. Section 10.5 covers lawyers, insurers and accountants, with a table of when to call which professional. Section 10.6 covers customers, peers and advisory panels, and what a panel may and may not decide. Section 10.7 covers buying help well: scope, deliverables, fees and independence. Section 10.8 names what a firm should keep in-house whatever else it buys.

10.1 The Roles Behind an Offer

An offer is not designed in a single act, and it is not designed by a single skill. It is the product of eight different kinds of work, each of which can be done well or badly, and each of which has to be done by somebody before a customer can buy anything. The work of deciding settles what the firm will sell and what it will refuse. The work of listening gathers what customers actually say, in their own words, about the job they are trying to get done. The work of designing the service turns that listening into a sequence of steps a technician can follow on a Tuesday morning. The work of costing puts a number on each step. The work of wording the promises writes the guarantee, the service level and the exclusion in language that a customer and a court would read the same way. The work of selling carries the offer to a prospect. The work of delivering performs it. The work of measuring watches whether the promise held.

In a large firm those eight kinds of work are eight departments, and the handoffs between them are the subject of meetings. In a small business they are one person on a Sunday evening, and the handoffs happen inside one head. That is not a defect. It is the reason a small business can change an offer in a week while a large one takes a quarter, and it is also the reason a small business can carry an inconsistency for years without noticing it. The same person who decided that urgent calls would be answered within ninety minutes also wrote the guarantee, also set the price, also trained the technician, and also reads the complaint when the promise fails. Nothing is lost in translation because there is no translation.

What is lost is the check. A role performed by one person is a role that is never challenged, and the challenges are exactly what the eight kinds of work need. The listening is the clearest example. An owner who has spent fourteen years as a technician knows what a building needs, and that knowledge is an asset in every conversation with a customer and a liability in every interview with one. The owner asks a question, hears the beginning of an answer, and finishes the answer internally before the customer has reached the second sentence. The interview produces confirmation rather than information, and the offer that follows is designed for the customer the owner already had in mind.

Naming the roles is therefore not an exercise in organization charts. It is a way of asking, for each of the eight kinds of work, who performs it, what they produce, and who checks it. At Nordec Entretien inc., the fictional case followed through this book, the answer after the first annual refresh was written down. Sophie Nadeau, the owner, decided: she owned the offer and made every decision about what the firm would sell, what it would refuse and what it would promise. The administrator became the offer coordinator, keeping the exception register, the monthly measures, the pilot log and the catalog. The lead technician owned the Building Check and the capital planning notes. The dispatcher owned the urgent-call blueprint and the guarantee credits. Four people, eight kinds of work, and every one of them named.

Figure 10.1 shows the arrangement. The owner sits at the center, because the decisions are the owner's. The offer coordinator sits beside the owner, because the records are what make the decisions repeatable. Around them are the team members who own a service each: the lead technician, who owns the Building Check and the notes, and the dispatcher, who owns the blueprint and the credits. On an outer ring are the four kinds of outside help the firm bought or assembled: a lawyer, linked to the wording of the guarantees; an accountant, linked to the cost of each tier; a facilitator, linked to the working session that produced the tiers; and a customer panel, linked to the reading of the offer before the pilot. Each link is a task, not a relationship.

Figure 10.1 — The roles behind an offer. At Nordec's size, the owner decides, the administrator coordinates, two team members own a service each, and outside help is bought for a named deliverable.

The figure is worth reading in one direction only: from the outside in. The outer ring exists to serve the tasks of the inner ring, and a piece of outside help that cannot name the task it serves is a cost without a deliverable. The lawyer does not design the offer; the lawyer reads the wording of a promise the owner has already decided to make. The accountant does not set the price; the accountant checks the cost of a tier the owner has already designed. The facilitator does not choose the tiers; the facilitator runs a session in which the team chooses them. The customer panel does not approve the offer; the panel reads it and says what it understood.

The listening role deserves a second look, because it is the one a small business most often believes it performs when it does not. Listening, as a role, means producing a record of what customers said, in their words, before the offer was designed. It has a deliverable, which is a set of notes with quotations in them, and it has a check, which is whether the notes contain anything the owner did not already expect. At Nordec, the listening was performed by Sophie and the administrator together, in twenty interviews, and the record is what Chapter 2 of this book describes. The pairing is the check: one of the two interviewers knew the buildings and the other did not, and the questions the second one asked were the questions the first one would never have thought to ask.

The eight kinds of work also explain why the chapters of this book are ordered as they are. Chapter 2 treats listening. Chapter 3 treats the design of the service. Chapter 4 treats packaging, which is the work of deciding what goes in which tier. Chapter 5 treats the wording of promises. Chapter 8 treats the measurement that tells the firm whether the promise held. Chapter 9 treats the cadence that keeps all of it current. This chapter treats the question that runs underneath all of them: who performs each kind of work, and when does it make sense to buy the work from outside rather than perform it inside.

The answer is not a principle about size. A solo owner with a clear method performs all eight kinds of work and produces a coherent offer, and a firm of forty with an unclear method produces eight half-performed roles and an offer that contradicts itself. What matters is that each kind of work has a name, a deliverable and a check, and that the person who performs it knows which of the three they are responsible for. A role with a name and no deliverable is a title. A deliverable with no check is a hope.

DEFINITION Offer owner, offer coordinator and customer panel

The offer owner, called a product owner in some firms, is the person who decides what belongs in the catalog: which tiers exist, what each one contains, which add-ons are sold, which services are refused, and how the offer is described. The offer coordinator is the person who keeps the records that make those decisions repeatable: the catalog itself, the register of exceptions and complaints, the monthly measures, and the log of any pilot. The customer panel is a small group of existing customers who agree to read the offer before it is piloted and to be consulted at a fixed interval afterwards, receiving nothing in exchange but the first look. The three are roles rather than job titles, and in a business of seven people they may sit in two people or even one. What matters is that each role has a named holder, a deliverable and a place in the calendar, because a role that exists only as an understanding is a role that disappears in the first busy month.

10.2 The Owner and the Offer Owner

The owner decides, and this is not a formality. In a business of any size, the question of what the firm promises is the question of what the firm is, and it cannot be delegated to a coordinator, an adviser or a committee. The owner is the person who carries the consequences of a promise that turns out to be too expensive, too slow or impossible to keep, and the owner is therefore the person who must make it. Everything else in this chapter is arranged around that single fact. The offer coordinator keeps the records; the service owner keeps the service; the lawyer reads the wording; the panel reads the description. None of them decides what the firm will promise.

The distinction between the owner and the offer owner is a distinction between deciding and keeping coherent. The offer owner, or product owner, is the role Book A7 names when it discusses who decides what belongs in the catalog. In a firm large enough to have one, that person holds the catalog, knows what each tier contains, knows which add-ons exist and which were refused, and brings the owner a decision when a decision is needed. The role exists because a catalog drifts in the same way a software stack drifts: an exception granted to one customer in March becomes an expectation in June, and by December nobody can say what the standard offer is. The offer owner is the person who notices the drift while it is still one exception.

When is a firm large enough to name an offer owner? The honest answer is that the threshold is not a headcount but a symptom. A firm needs the role when two people can describe the same tier differently, when an exception granted to one customer is repeated to another without a decision, or when the person who sells the offer is not the person who designed it and has begun to improvise. At that point the catalog has stopped being a document and become a set of habits, and a named person with the authority to say "that is not what we sell" is cheaper than the confusion. Below that point, the owner performs the role and the coordinator keeps the records, which is exactly what happens at Nordec.

At Nordec the arrangement is written down in one sentence: Sophie owns the offer and makes every decision, and the administrator is the offer coordinator. The coordinator's work is concrete and it is the work that makes the owner's decisions stick. The exception register records every request that fell outside the standard offer, what was decided, on what date, and by whom, so that the same request arriving six months later is answered in a minute rather than reconsidered from the beginning. The monthly measures record the figures of Chapter 8: the number of proposals, the close rate, the renewals, the downgrades and upgrades, the urgent calls, the misses and the credits. The pilot log records the decision rules of Chapter 8 and the results against them. The catalog is the one-page description of the three tiers that a customer reads and a sales conversation starts from.

None of those four records requires a specialist. All four require a person who has been told that keeping them is their job. That is the whole of the offer coordinator's role at this size, and it is worth stating what it is not: the coordinator does not decide what the firm sells, does not change a tier, does not grant an exception and does not adjust a guarantee. The coordinator makes the owner's decisions visible and durable, which is a different and equally necessary kind of work. A business that has the decisions and not the records will make the same decision three times in a year and believe it has made three decisions.

What never leaves the owner's hands is short enough to be written on a card. The owner decides what the firm promises, which is the guarantee, the service level and the response commitment. The owner decides what the firm refuses, which is the list of work the firm does not take, the customers it does not serve and the tenders it does not bid. The owner decides what each tier contains and what it excludes. The owner decides when a promise changes, and the change is written down on the day it is made rather than remembered. Everything else — the wording, the costing, the description, the measurement, the panel — can be performed by somebody else, and should be, because the owner's attention is the scarcest resource in the business.

The reason the promise cannot be delegated is not hierarchy. It is that a promise is a commitment of the firm's capacity, and only the owner can see the whole of that capacity. A coordinator who grants an exception does not know that the lead technician is already committed to eleven Building Checks in the same month. A salesperson who improvises a response time does not know that the dispatcher has one vehicle in the east territory on the day in question. The owner is the only person in a small business who holds the whole picture, and the promise is the point at which the whole picture has to be consulted. This is also why the offer coordinator's records matter so much: they are what allow the owner to consult the picture in a minute rather than reconstruct it from memory.

CASE IN POINT Nordec Entretien inc.: the administrator becomes the offer coordinator

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The administrator had worked at Nordec since the first year, keeping the schedule, the invoices and the customer files. What changed at the first annual refresh was not the work but its name and its scope. The administrator became the offer coordinator, and four records moved into that role. The exception register was created, because exceptions had been granted for two years and recorded nowhere. The monthly measures were formalized, because the figures existed in three places and were read in none. The pilot log was opened before the sixty-day pilot of Chapter 8 began, with the decision rules written on the first page. The catalog was assembled, which meant turning the three tiers of Chapter 4 into one page a customer could read. None of the four required new software or new hours beyond the cadence of Chapter 9. What they required was that somebody be told, in writing, that keeping them was their job, and that the owner stop being the only person who knew where anything was.

10.3 Service Designers and Customer Experience Specialists

A service designer does for a service what an architect does for a building: turns an intention into a drawing that other people can build from and argue with. The work has a recognizable shape. It begins with research into what customers are trying to get done, which is the listening of Chapter 2 performed with a method rather than a conversation. It continues with a journey map, which lays the customer's experience across time and marks the moments that matter. It produces a blueprint, which is the drawing described in Chapter 3: the customer's actions along the top, the visible employee actions below the line of interaction, the invisible work below the line of visibility, and the support processes beneath that. It ends with a prototype, which is a version of the service tried on a small scale before it is offered to everybody.

A customer experience specialist works on the same material from the customer's side. The distinction is not sharp, and in a small engagement the same person often performs both. What a customer experience specialist brings is attention to the whole of the relationship rather than to a single service: the touchpoints through which a customer meets the firm, the channels they use, and the consistency between them. Customers now meet a firm through many touchpoints, in several channels and media, and their experience of the firm is the sum of those meetings rather than the memory of a single one. Delivering a consistent experience therefore requires several functions of the business to work from the same description of the service, and sometimes requires external partners to be brought into that description. That is a problem a small business has in a mild form and a firm with several locations has in an acute one.

When does a small business need one? Three situations justify the cost. The first is a new service line, where the sequence of steps does not yet exist and somebody has to decide what happens in what order. The second is a service that produces repeated complaints, where the complaints are the symptom and the process is the cause, and where the firm has already tried to fix it twice by exhortation. The third is a redesign that has to hold across several locations or several teams, where the drawing is the only way to make the same service happen in the same way in more than one place. Outside those three situations, a small business usually does better to draw the blueprint itself, badly, and improve it, than to buy a beautiful drawing it cannot maintain.

What to ask for as deliverables is the practical question, and the answer is a short list. Ask for a journey map that names the touchpoints and marks the moments of truth, which are the points at which the customer and the firm meet and the promise is either kept or broken. Ask for a blueprint with the customer's actions drawn first, because the method begins from the customer and works inward, and a blueprint drawn from the firm's process outward will describe the firm rather than the service. Ask for a prototype plan: which customers, how many, for how long, and what will be measured. Ask for the working files, not only the final document, because the firm will have to change the drawing in six months and should not have to buy it again. And ask for a session in which the people who deliver the service redraw part of it themselves, because a blueprint the team has not touched is a blueprint the team will not follow.

The cost of a service design engagement is a project cost, and it should be compared with the cost of the problem rather than with the cost of doing nothing. A firm that loses two proposals a year because its offer is described inconsistently, or that spends three hours a week on complaints that a redesigned sequence would prevent, has a problem with a number attached. The engagement is worth buying when the number is larger than the fee and when the deliverable is something the firm can keep using. It is not worth buying when the deliverable is a presentation, because a presentation is a description of a drawing rather than a drawing.

WATCH OUT The consultant who designs the offer the firm cannot deliver

The most expensive piece of outside help is the one that produces an offer the team cannot perform. It happens in a recognizable way. An adviser is engaged, works with the owner and one or two managers, and produces a design that is elegant on paper: a service level the firm has never met, a tier that includes work nobody has been trained to do, a guarantee whose exceptions are not written down. The team sees the design for the first time at the meeting where it is presented, and the first urgent call after the launch shows what was missed. The remedy is a rule about who drafts and who challenges. The team drafts, because the team knows what the work actually involves, and the adviser challenges, because the adviser knows what the work could be. An adviser who has never watched the service being delivered is a useful critic and a dangerous author, and the difference between the two is decided by who holds the pen. The same rule applies to the wording of a promise: the person who will have to keep it should be in the room when it is written.

10.4 Pricing and Packaging Specialists

The line between packaging and pricing is the line between this book and Book A7, and it is worth drawing precisely because the two are so often bought together. Packaging decides what is in each tier, what is bundled with what, which add-ons exist, what each tier excludes, and how the tiers are described to a customer. Pricing decides what each tier costs, how the price is structured, whether it is a subscription or a one-time amount, and what the firm does when a customer asks for a discount. The two decisions constrain each other, and they are made by different methods. A packaging decision is tested by asking whether a customer can tell the tiers apart and whether the firm can deliver what each one contains. A pricing decision is tested by asking whether the number produces the margin the firm needs at the volume it can serve.

A pricing specialist works on the second question. The work includes research into what customers would be willing to pay, which is usually done by asking them to choose between described offers rather than by asking them for a number, because a customer asked what they would pay answers a different question from the one the firm needs answered. It includes the design of a tier structure, which is the arrangement of the offers relative to one another: how many there are, how far apart they sit, and which one the firm expects most customers to choose. It includes choice research, which tests how real customers behave when the offers are placed side by side. What makes this work more than arithmetic is that a customer's choice depends on the set of alternatives placed in front of them, so the structure itself is part of what is being tested, and an option that sits in the middle of a set is often read differently from one at the extreme.

The same body of practice explains why the number of tiers matters. A firm that offers eleven versions of its service is not being generous; it is making the choice harder and the decision slower. A firm that offers three tiers, with a clear flagship and a clear entry, is doing what the practice of packaging supports, and it is what Nordec did in Chapter 4, with three tiers and two add-ons.

When does the investment in a pricing specialist pay? The honest answer is that it pays when the firm has enough customers for the structure to be tested and enough volume for a small improvement in the mix to matter. A firm with four customers and a bespoke offer for each does not need a tier structure; it needs to know its cost to serve, which is the accountant's work rather than the specialist's. A firm with forty-six buildings under contract, three tiers and a renewal cycle, like Nordec, has exactly the conditions in which the structure can be tested and the mix can be measured. What Nordec did not buy was a pricing consultant, and the reason is stated plainly in the case: the prices were set with the method of Book A7, and the method was enough at that size. What the firm did buy was half a day of the external accountant's time to check the cost-to-serve figures of each tier against the accounting records.

What to ask for when a pricing specialist is engaged is a short list of deliverables. Ask for the structure as a document: how many tiers, what each contains, what the price ratios between them are, and which tier the firm expects to be the flagship. Ask for the evidence behind the structure, which is the record of what customers chose when the offers were put in front of them, not a summary of what they said they preferred. Ask for the assumptions the structure rests on, so that the firm can tell when they stop being true. Ask for the method to be handed over, so that the next revision can be performed in-house. And ask for the boundary: a pricing engagement should not silently redesign the offer, because the contents of a tier are a packaging decision and the firm has to be able to deliver them.

The most common mistake in buying this kind of help is to buy the answer instead of the method. A specialist who hands over a price list and leaves has produced a result that will be wrong within a year, because costs move, the mix moves and the market moves. A specialist who hands over a method, a structure and the assumptions behind it has produced something the firm can maintain. The second engagement is worth more than the first even when it costs the same, and the difference is visible in the deliverable: one is a number, the other is a way of arriving at a number.

KEY CONCEPT Buy expertise for a deliverable, never for a direction

The rule that keeps outside help useful is that it is bought for a deliverable and never for a direction. A deliverable is a thing that can be handed over, read, questioned and used: a reviewed paragraph, a checked cost figure, a drawing, a structure with its assumptions, a session with recorded decisions. A direction is a feeling that the firm is now on the right track, and it cannot be handed over, questioned or used, which is why engagements bought for a direction end with a presentation and a sense that something should have changed. The rule has a second half that matters just as much. The deliverable must be one the firm could not produce itself in the time available, or the purchase is a way of avoiding work rather than a way of buying skill. A lawyer reviewing a guarantee is buying skill the firm does not have. A facilitator running a session the team could run itself is buying time and neutrality, which are legitimate and should be named as such. Naming what is being bought is the whole of the discipline.

10.5 Lawyers, Insurers and Accountants

Three professionals appear in the design of an offer more often than any others, and each of them is called for a different reason. The lawyer is called for the wording of promises: the guarantee, the service level, the exclusions, the cancellation terms and the obligations that apply when the customer is a consumer rather than a business. The insurer is called when a promise creates an exposure the firm could not absorb if it were called often: a service credit, a guarantee that pays money, or a commitment whose failure would cost more than the contract is worth. The accountant is called for the cost of what the firm has designed: the cost to serve each tier, the treatment of credits and refunds, and whether the margin the offer produces is the margin the firm believes it produces.

The lawyer's work on an offer is narrower than small businesses often assume and more valuable than they often expect. It is not the lawyer's role to decide what the firm promises; that decision belongs to the owner. It is the lawyer's role to make sure that the promise, once decided, says what the firm means and means what the firm says. A guarantee that promises a credit "if the firm fails to respond promptly" is a guarantee that will be argued about, because promptly has no number in it. A guarantee that promises a credit of a stated amount if a technician is not on site within ninety minutes of an urgent contract call is a guarantee that can be administered by a dispatcher without a lawyer in the room. The difference between the two is a few hours of drafting, and it is the cheapest risk reduction available to a small business.

The insurer's role is to answer a question the owner cannot answer alone: if this promise is called more often than expected, what does it cost, and can the firm absorb it? A service credit is a small, bounded exposure, and a firm can usually carry it. A guarantee that pays a large amount, or a commitment to a result the firm does not fully control, is a different matter, and the answer may be a policy, a cap on the total credits payable in a period, or a decision not to make the promise at all. The point of the conversation is not to buy insurance; it is to find out what the promise exposes, in a conversation with somebody whose job is to think about the bad year rather than the good one.

The accountant's role is the one most often skipped and the one with the most immediate effect on the offer. A tier is a promise about work, and work has a cost. The cost to serve a tier is the loaded cost of the hours the tier consumes, plus the consumables, plus the share of the travel, plus the cost of the records and reports the tier includes. At Nordec, the cost to serve the Preventive tier is $336.73 a month and the cost to serve the Essential tier is $219.63 a month, and those figures were checked by the external accountant against the accounting records over half a day. The check matters because a cost figure produced from memory is a figure that drifts, and a tier priced against a drifted cost is a tier that quietly loses money.

The following table sets out seven situations that arise while an offer is being designed, the professional each one calls for, the deliverable to ask for, and the scope the engagement usually takes. Read it as a routing table rather than a schedule: most offers call for two or three of the seven, and a firm that finds itself calling for all seven at once is probably designing something larger than it intends to deliver.

Table 10.1 — When to call which professional

QUÉBEC & CANADA NOTE What a lawyer reads in a guarantee

Nordec's contracts are between two businesses, and in Québec they are governed by the Civil Code of Québec. Article 2100 provides that a contractor or provider of services must act in the client's best interests, with prudence and diligence, in accordance with usage and good practice, and, where applicable, ensure that the work or service conforms to the contract; when bound to an obligation of result, they may be relieved of liability only by proving superior force. Article 2102 requires the provider, before the contract, to give the client useful information on the nature of the task and the property and time it requires. Articles 2125 and 2129 allow the client to resiliate a contract for services unilaterally, paying for the work performed and the costs incurred. A reader who sells to consumers faces a different set of rules: under the Consumer Protection Act, statements and advertisements of the merchant or manufacturer bind them (ss. 41 to 43), and an exclusion from a conventional warranty must be set out clearly, in separate successive clauses (s. 44). This book explains these rules and does not replace the advice of a lawyer.

DO THE MATH What Nordec spent on outside help

Illustrative figures; the case is fictional. Step 1 — The lawyer reviewed the wording of the 90-minute guarantee and of the first-90-days exit, which took 3 hours at a fixed fee agreed in advance, and produced a reviewed text with the ambiguities marked. Step 2 — The external accountant checked the cost-to-serve figures of each tier against the accounting records, which took half a day and produced a list of the assumptions behind each figure. Step 3 — The independent facilitator ran the working session on the tiers; the facilitator had already been hired for the strategy refresh described in Book A3, so the session added no new engagement. Step 4 — The customer panel cost nothing: five customers, three owner-operators and two property managers, agreed to read the one-page description of the tiers before the pilot and to be called once a quarter, receiving nothing in exchange but the first look. Step 5 — What was not bought is as instructive as what was: no service designer and no pricing consultant, because the twenty interviews of Chapter 2, the blueprint of Chapter 3 and the method of Book A7 were performed in-house and were enough at Nordec's size.

10.6 Customers, Peers and Advisory Panels

The most useful help a small business can get is usually already paying for the service, and it is usually free. A customer who has agreed to read an offer before it is piloted will say things no adviser can say, because the customer is the person the offer is for. The panel is the structured form of that help: a small group of existing customers who see the offer before the pilot and are consulted at a fixed interval afterwards. At Nordec the panel is five customers, three owner-operators and two property managers, and they receive nothing in exchange but the first look. That last detail is deliberate. A panel paid in discounts is a panel whose answers are shaped by the discount, and a panel paid in nothing but attention gives answers that are worth more than the fee saved.

What a panel may decide is limited, and the limit is what makes it useful. A panel may say whether it understood the offer, which is the single most valuable thing a firm can learn before a launch. A panel may say which tier it would choose and why, which is a test of whether the tiers are distinguishable. A panel may say what it expected to be included and was not, which is the fastest way to find the gap between what the firm means and what the customer hears. A panel may say what it would have to give up to move from one tier to another, which is the test of whether the tiers are actually different. A panel may not decide what the firm sells, may not set a price, and may not approve a promise. Those remain the owner's decisions, informed by what the panel said.

The reason the limit matters is that a panel is not a sample. Five customers chosen because they agreed are five customers with a relationship to the firm, and their answers describe their own situations rather than a market. A firm that treats a panel as evidence about demand will over-read it; a firm that treats it as evidence about comprehension will use it well. The distinction is easy to state and easy to lose. "Four of the five said they would buy the Portfolio tier" is a statement about five people who already buy from the firm. "Four of the five could not explain the difference between Preventive and Portfolio without help" is a statement about the description, and it is actionable on the same afternoon.

Peers and industry associations are the second source of free help, and they are useful in a different way. A peer who runs a similar business in another territory will say what a customer will not: what the work actually costs, what a reasonable response time is, which promises are common in the trade and which are unusual. An association provides the same thing at a distance, through its meetings, its publications and its members. The value of both is that they are outside the firm and inside the industry, which is the combination that makes a comparison possible. The limit is the same as the panel's: a peer describes their own business, and a practice that is common in the trade is not therefore right for this firm.

The panel also has a use that is easy to overlook, which is as a check on the listening. The twenty interviews of Chapter 2 produced a set of elements ranked by how often customers named them among their top three. The panel is where those rankings meet a description: the firm writes down what it concluded and shows it to five people who were not in the interviews, and the places where the panel hesitates are the places where the conclusion was thinner than it looked. That is a cheap check on an expensive piece of work, and it costs a phone call.

The practical arrangements are small. Five to eight customers is enough; more than that turns a conversation into a presentation. One person from the firm runs it, and the same person runs it every time, so that the panel's answers can be compared across quarters. The questions are written before the call and are the same for everybody. The panel is told what happened as a result of what it said, because a panel that never hears back stops answering. And the panel is consulted at a fixed interval, once a quarter at Nordec, because a panel called only when the firm has something to launch is a panel that has been turned into a focus group.

WHAT CHANGED Nordec's roles before and after the refresh

Before the first annual refresh, Sophie did everything and nobody kept the offer's records. She performed the listening, designed the service, costed the tiers from memory, wrote the guarantee in the same afternoon as the decision to offer it, sold the contracts, and read the complaints. The exceptions granted to customers were remembered rather than recorded, the monthly figures existed in three places and were read in none, and the offer had grown by accretion to six elements, four of them given away free. After the refresh, four roles were named and one panel was assembled. The administrator became the offer coordinator, holding the exception register, the monthly measures, the pilot log and the catalog. The lead technician owned the Building Check and the capital planning notes. The dispatcher owned the urgent-call blueprint and the guarantee credits. Five customers agreed to read the offer before the pilot and to be called once a quarter. Sophie still made every decision, and the decisions became visible, dated and repeatable.

10.7 Buying Help: Scope, Deliverables and Independence

Buying help well is a small discipline with five parts, and the first is the one most often skipped. Before an adviser is engaged, the firm names the decision the help must support. Not the topic — the decision. "We need to know whether the 90-minute guarantee can be worded so that a dispatcher can administer it without a lawyer" is a decision. "We need help with our guarantee" is a topic, and a topic produces an engagement that ends when the budget ends rather than when the decision is made. Naming the decision first also makes the second part easy, because a decision has a deliverable attached to it: the reviewed text, the checked figure, the drawing, the structure, the recorded session.

The second part is the written deliverable and the deadline. A deliverable is a thing, and the test of whether it has been named properly is whether the firm can say what it will do with it on the day it arrives. The reviewed guarantee will be published. The checked cost figure will be entered in the tier's record. The blueprint will be used to train the technicians. The structure will be presented to the owner for a decision. A deliverable that has no use on the day it arrives was not a deliverable; it was an activity. The deadline matters for the same reason: an engagement without an end date expands to fill the attention available, and the attention of a small business is the scarcest thing it has.

The third part is the fee. A fixed fee or a cap is the arrangement that suits a small business, and it is usually available for bounded work. A lawyer reviewing three paragraphs can quote a fixed fee. An accountant checking a set of cost figures can quote half a day. A facilitator running a session can quote a day. An engagement that cannot be quoted as a fixed fee is usually an engagement whose scope has not been defined, and the answer is to define the scope more narrowly rather than to accept an open commitment. The fourth part is independence, and it is the part that protects the firm from advice that is really a sale. An adviser who recommends a product they also sell, or who receives a commission on what they recommend, is not giving advice; they are selling, and the firm should know which conversation it is in.

The fifth part is what the firm keeps. The working files, the drafts, the interview notes, the customer's words and the decision record stay with the business, in a place the business controls, with versions that can be told apart. This is not a matter of distrust. It is a matter of continuity: the firm will revise the offer in a year, and the revision starts from the files. A firm that has to buy the same drawing twice has paid for it twice. A firm that keeps the customer's words in its own records has an asset that grows, because the words from three years ago can be compared with the words from this year, and the comparison is where the offer's next change comes from.

The following table sets out eight tasks of offer design and who performs each of them in a solo business, a small team and an established firm. Read it as a distribution of responsibility rather than a hierarchy: the tasks are the same in every column, and what changes is whether one person carries several of them or several people carry one each.

Table 10.2 — Who does what in offer design at three sizes

METHOD Hiring outside help in five steps

First, name the decision the help must support, in one sentence, and write it down before any conversation with an adviser. Second, write the deliverable and the deadline: what will be handed over, in what form, and on what date, and what the firm will do with it on the day it arrives. Third, ask for a fixed fee or a cap, and treat an inability to quote one as a sign that the scope is still undefined. Fourth, check independence from any product the adviser would recommend, and ask directly whether they receive anything from the suppliers they name; if the answer is yes, treat the conversation as a sale. Fifth, keep the working files and the customers' words in-house, in a place the business controls, with versions that can be told apart, so that the next revision starts from what was already paid for. Deliverable: a one-page scope for each piece of help, with the decision, the deliverable, the deadline, the fee and the independence check on it. Pitfall: engaging an adviser for a topic rather than a decision, which produces an engagement that ends when the budget ends.

TOOLS OF THE TRADE What working with outside help needs

Five ordinary things carry the work of buying and using help, and none of them is a specialist product. The first is a written scope template, one page with the decision, the deliverable, the deadline, the fee and the independence check, so that every engagement is described the same way. The second is a shared folder with version control, where drafts, working files and final documents are kept in an order that lets a reader tell which is current. The third is a decision record, a dated list of what was decided, by whom, and on what advice, which is what allows a firm to explain a promise two years later. The fourth is a meeting calendar that holds the panel calls, the review sessions and the deadlines at fixed intervals rather than when somebody remembers. The fifth is a short brief, one page, that the firm sends before any engagement, so that the adviser arrives knowing what the firm sells, what it refuses and what the decision is. No brand name belongs on this list; the requirement is the function.

10.8 What You Keep In-House

Four things should stay inside the business whatever else is bought, and the first is the customer's words. The notes from the interviews, the quotations, the complaints, the requests that came in over the phone — these are the raw material of every future change to the offer, and a firm that lets them live in an adviser's files has rented its own memory. Keeping them is not difficult. It means that the person who performs the listening writes the notes into a place the business controls, in the customer's own words rather than in a summary, and that the notes are dated and kept. Three years of such notes is a record of how the customers' language changed, which is often the earliest signal that the offer needs to change.

The second thing to keep is the decisions. What the firm promises, what it refuses, what each tier contains, what each tier excludes, when a promise was changed and why — these are the owner's decisions, and they are recorded in the catalog and the exception register rather than in a consultant's report. The reason is not ownership but continuity. A decision recorded in the firm's own records can be found, questioned and revised by the people who have to live with it. A decision recorded in a report that sits in a folder is a decision that will be rediscovered, usually by a customer asking why the firm did something different last year.

The third thing to keep is the catalog, which is the one-page description of the offer that a customer reads and a sales conversation starts from. It is the firm's own document, in the firm's own voice, and it is revised by the firm. An adviser can challenge it, a lawyer can read it, a panel can test it, and none of them should own it. The catalog is also the place where the offer's coherence is visible: three tiers, two add-ons, a stated response time, a stated exclusion, and a description short enough that a customer can read it before deciding. A catalog that has grown to nine pages has stopped being a catalog and become a contract, which is a different document with a different purpose.

The fourth thing to keep is the exception register, which is the record of every request that fell outside the standard offer, what was decided, on what date and by whom. It is the most valuable of the four and the least likely to exist. An exception register is what turns a year of improvisation into a list of candidate changes: the requests that recur are the services the firm should consider adding, and the requests that were granted repeatedly are the promises the firm has already made without saying so. At Nordec, the register is what made the service menu of Chapter 6 possible, because the eight candidate services were judged partly on who had asked for them, and the asking had been recorded.

The reason a firm that outsources its listening loses its offer is worth stating plainly. The offer is a set of answers to questions customers asked, and the questions arrive in the customer's language, at the customer's pace, in the customer's building. An adviser who performs the listening returns a summary, and a summary is a set of answers with the questions removed. The firm then designs an offer for the summary rather than for the customers, and the gap between the two is discovered at the first renewal. Listening is the one kind of work that cannot be delegated without losing the thing it produces, which is why the panel of section 10.6 is a check on the listening rather than a substitute for it.

Nordec's choices are a fair illustration of where the line falls at a small size. The firm kept the listening, the decisions, the catalog and the exception register in-house, and it kept the blueprint and the pricing method in-house as well, performing the twenty interviews of Chapter 2 and the blueprint of Chapter 3 with its own people and setting prices with the method of Book A7. It bought three things from outside: three hours of a lawyer's time on the wording of the two guarantees, half a day of the external accountant's time on the cost-to-serve figures, and a working session run by a facilitator who had already been engaged for the strategy refresh of Book A3. It assembled one thing for free: a panel of five customers. The total is small, the deliverables are named, and every one of them is a thing the firm still has.

The general rule that emerges is not about money but about what grows. The listening, the decisions, the catalog and the register all grow in value as they accumulate, because each year's record makes the next year's decision better informed. A drawing, a structure or a reviewed paragraph is valuable when it arrives and does not improve with age. Buy the things that do not improve with age, and keep the things that do. That is the whole of the judgment, and it is the judgment that decides, for a business of any size, which help to buy and which work to perform.

THE RULE The owner keeps the promises

What the firm promises and what the firm refuses are decided by the owner and written down; nobody else changes a guarantee, a tier or an exclusion. The rule has three parts. The first is that the decision is the owner's, whatever help was bought to inform it: a lawyer may redraft a guarantee, an accountant may correct a cost, a facilitator may run the session in which the tiers are chosen, and none of them may change what the firm commits to. The second is that the decision is written down on the day it is made, in the catalog or the exception register, with its date, because a promise that exists only in a conversation is a promise the firm will make differently to the next customer. The third is that a change to a promise is a decision and not an adjustment: when a guarantee is widened, a tier is altered or an exclusion is narrowed, the change is made by the owner, recorded, and reflected in the catalog before the next customer is told about it. A firm that follows the rule can always answer the question of what it sells.

WHO DOES THIS The lead technician and the dispatcher each own a service

Ownership of a service is a small, concrete thing, and Nordec's two examples show what it looks like in practice. The lead technician owns the Building Check: the technician decides what the assessment covers and in what order it is performed, performs the six hours on site and the two hours of writing, signs the capital planning note that goes to the owner of the building, and answers when a note is questioned. The dispatcher owns the urgent-call blueprint: the dispatcher keeps the sequence of the call, decides what the technician records on site, confirms each month how many guarantee credits were due, and checks that each credit was issued automatically without a claim. Neither role required a new title or a new hire. Both required three things that a small business can supply: a written description of what the role produces, a place in the monthly measures that the offer coordinator keeps, and the authority to say no when a request falls outside what the service covers.

IN LA MACHINE Roles and tasks in the platform

The application copy read for this book carries the roles of this chapter in four places (La Machine, 2026). Each role's recurring work can be held as a task at productivity/tasks, and in the Control center the plan tasks page at /taches gives each task a verification criterion, which is the closest the application comes to the "who checks it" column of the role map in section 10.1. The contacts page at business/marketing/contacts holds one record per customer and the people attached to it, with notes, which is where a panel of five customers can be kept together with the record of what each one said. The meetings page at productivity/meetings holds the reviews and the panel calls, so that the quarterly conversation with the panel is an appointment rather than an intention. Read together, those four pages carry the coordination work of an offer: the tasks, their checks, the people and the meetings. What they do not carry is the drawing, the guarantee register or the measures, which stay in the firm's own records.

Costly Mistakes

• Performing all eight kinds of work alone and calling it efficiency. A single person can hold every role in a small business, and the roles still need names, deliverables and checks, or the listening becomes confirmation and the costing becomes memory.
• Letting the offer owner's role go unnamed. When two people describe the same tier differently, or an exception granted in March has become an expectation by June, the catalog has stopped being a document and become a set of habits.
• Buying help for a topic rather than a decision. An engagement described as help with the guarantee or a look at the offer has no deliverable and no end, and it finishes when the budget finishes rather than when the decision is made.
• Letting the adviser write the offer the team has to deliver. A design produced without the people who perform the service is elegant on paper and fails at the first urgent call; the team drafts and the adviser challenges.
• Skipping the accountant on the cost to serve. A tier priced against a cost figure produced from memory is a tier that quietly loses money, and the check costs half a day.
• Publishing a guarantee nobody has read. A promise worded without a lawyer is a promise that will be argued about, and the argument costs more than the hours of drafting it would have taken to avoid.
• Outsourcing the listening. A summary of what customers said is a set of answers with the questions removed, and the firm that designs for the summary discovers the gap at the first renewal.

Key Takeaways

• An offer is the product of eight kinds of work — deciding, listening, designing the service, costing, wording the promises, selling, delivering and measuring — and a small business combines them in a few people, which makes naming each role, its deliverable and its check the first act of offer design.
• The owner decides what the firm promises and refuses, and that decision never leaves the owner's hands; the offer owner or offer coordinator keeps the catalog, the exception register, the measures and the pilot log coherent, and at Nordec's size that is Sophie with the administrator as coordinator.
• A service designer produces a journey map, a blueprint with the customer's actions drawn first, and a prototype plan; the engagement is worth buying for a new service line, for repeated complaints, or for a redesign that has to hold across several locations.
• Packaging decides what is in each tier and pricing decides what each tier costs; a pricing specialist is worth engaging when there are enough customers to test the structure and enough volume for the mix to matter, and the method should be handed over rather than the answer alone.
• Three professionals are called for different reasons: the lawyer for the wording of promises, the insurer for what a promise exposes if it is called often, and the accountant for the cost to serve each tier, which at Nordec was checked over half a day against the accounting records.
• Four things stay in-house whatever else is bought — the customer's words, the decisions, the catalog and the exception register — because they grow in value as they accumulate, while a drawing, a structure or a reviewed paragraph is valuable when it arrives and does not improve with age.

Your Action Plan

1. Name the eight roles and their holders. Write the eight kinds of work behind the offer in a list, and beside each one write the name of the person who performs it, what that person produces, and who checks it. Deliverable: a one-page role map with a named holder for each kind of work. Verification: every role has a deliverable, and no role has two names or none.
2. Separate the owner's decisions from the coordinator's records. Write down the four decisions that stay with the owner — what the firm promises, what it refuses, what each tier contains and what each tier excludes — and name the person who keeps the catalog, the exception register, the monthly measures and the pilot log. Deliverable: a written statement of the owner's decisions and a named offer coordinator. Verification: the coordinator can show the four records, and none of them contains a decision the owner did not make.
3. Write a scope for each piece of outside help before it is engaged. For every adviser the firm intends to use, write one page with the decision the help must support, the deliverable, the deadline, the fee or cap, and the independence check. Deliverable: one scope page per engagement. Verification: each scope names a deliverable the firm can say what it will do with on the day it arrives.
4. Assemble a panel of five to eight customers. Invite existing customers to read the one-page description of the offer before the next pilot and to be called once a quarter, with nothing offered in exchange but the first look, and write the questions before the first call. Deliverable: a named panel and a written question set. Verification: the panel has read the offer and the firm has recorded what the panel could not explain back.
5. Keep the four records in-house and read them. Put the customer's words, the decisions, the catalog and the exception register in a place the business controls, with versions that can be told apart, and read the exception register at the next review. Deliverable: four records in the firm's own keeping. Verification: the exception register shows at least one request that recurred, and the firm can say whether it is a candidate change.

In La Machine

The platform read for this book carries the roles of this chapter in one place and not in another (La Machine, 2026). The tasks page at productivity/tasks holds each role's recurring work as tasks, and in the Control center the plan tasks page at /taches gives each task a verification criterion, which is the closest the application comes to the "who checks it" column of the role map in section 10.1. The contacts page at business/marketing/contacts holds one record per customer and the people attached to it, with notes, which is where a panel of five customers can be kept with the record of what each one said. The meetings page at productivity/meetings holds the reviews and the panel calls, so that the quarterly conversation with the panel is an appointment rather than an intention. Read together, those four pages carry the coordination work of an offer: the tasks, their checks, the people and the meetings.

What the platform does not do belongs in the same breath. La Machine has no service blueprint and no customer journey map, so the drawing of Chapter 3 is made outside it. It has no survey tool and no net promoter tool, only a general form builder at business/marketing/forms, with the answers kept at business/marketing/forms/responses, which can carry a short survey but is not a research instrument. It has no quote or proposal module, so the proposal that carries the offer to a customer is produced elsewhere, and Book E5 (Proposals, Quotes, Business Cases and Contracts) treats that work. It has no guarantee or warranty register and no service-level tracking, so the guarantee credits of Chapter 5 are counted by the dispatcher and recorded in the firm's own measures. The roles of this chapter are therefore named in the platform and performed outside it, which is the ordinary arrangement for a small business and worth stating plainly before a firm assumes that a page exists for every role it has named.

Self-Check

1. What are the eight kinds of work behind an offer, and why does a small business need to name each one even when one person performs them all?
2. What is the difference between the owner and the offer owner, and what four things never leave the owner's hands?
3. When is a service designer worth hiring, and what four deliverables should the firm ask for?
4. What does a lawyer, an insurer and an accountant each contribute to the design of an offer, and what should the firm ask each one to hand back?
5. What may a customer panel decide and what may it not decide, and why does the distinction matter?

Answers. 1. The eight kinds of work are deciding what the firm sells and refuses, listening to customers, designing the service, costing a tier, wording the promises, selling, delivering and measuring; they need names because a role performed by one person is never challenged, and the challenges are what keep the listening from becoming confirmation and the costing from becoming memory. 2. The owner decides what the firm promises and refuses and carries the consequences of the promise; the offer owner, or offer coordinator, keeps the catalog, the exception register, the monthly measures and the pilot log coherent without deciding what the firm sells. The four things that never leave the owner's hands are what the firm promises, what it refuses, what each tier contains and what each tier excludes. 3. A service designer is worth hiring for a new service line, for a service that produces repeated complaints, or for a redesign that has to hold across several locations or teams; the firm should ask for a journey map that names the touchpoints and marks the moments of truth, a blueprint with the customer's actions drawn first, a prototype plan naming which customers and what will be measured, and the working files rather than only the final document. 4. The lawyer makes sure the promise says what the firm means and hands back a reviewed text with the ambiguities marked; the insurer answers what the promise exposes if it is called more often than expected and hands back a view of whether a policy, a cap or a different promise is needed; the accountant checks the cost to serve each tier against the accounting records and hands back the figures with the assumptions behind them listed. 5. A panel may say whether it understood the offer, which tier it would choose and why, what it expected to be included, and what it would have to give up to move between tiers; it may not decide what the firm sells, set a price or approve a promise, because five customers who agreed to help describe their own situations rather than a market, and their answers are evidence about comprehension rather than about demand.

Chapter References

Civil Code of Québec, CQLR c CCQ-1991, arts. 1726, 2100, 2102, 2125 and 2129 (Québec). LégisQuébec, updated to April 7, 2026 — as read for Book H1 (Contracts and Commercial Agreements).

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/tasks, /taches, business/marketing/contacts and productivity/meetings.

Chapter 11 — Tools: Functions and Popular Tools

The tools of offer design are mostly ordinary — a wall, a spreadsheet, a form, a catalog — and a small business chooses them by the function each must perform across the life of an offer, from understanding customers to measuring what they choose, before it looks at a single brand.

What You'll Be Able to Do

• Name the six functions an offer design tool can serve — understand, design, package, sell, deliver and measure — and say which of them the business needs now.
• Choose a surface for the design work, whether a physical wall or a board application, and keep the customer profile, the value map, the journey and the blueprint in a form that survives a staff change.
• Run customer interviews and simple surveys with a guide, a schedule and a place to keep the customers' own words, and recognize the questions a survey cannot answer.
• Build a catalog that holds every tier, add-on and productized service with its contents and its exclusions, and decide when quoting has outgrown a one-page comparison.
• Connect the records a delivery tool produces to the promises the offer makes, so that arrival times feed guarantee credits and service levels are reported rather than believed.
• Judge when a spreadsheet is enough for measuring an offer and when a business has enough transactions to justify an experimentation tool.

Why It Matters

An offer is not designed in a single sitting, and it is not designed in a single place. It is designed across a wall, a document, a spreadsheet and a conversation, over weeks, by people who are also doing the work that pays. That is why the tools matter and why they matter less than they appear to. A business that buys a large platform before it has understood its customers will produce a beautiful map of an offer nobody wants. A business that understands its customers and keeps the understanding in a notebook will produce an offer that works, and will then discover that the notebook cannot be quoted from, billed from or measured from. The tools are the difference between an offer that can be run and an offer that can only be described.

The running case shows what the ordinary version looks like. Nordec Entretien inc. designed three tiers, two add-ons, a productized entry offer and two guarantees, and it did so with a whiteboard on the office wall above the depot, a spreadsheet that already held the budget, a form that already collected requests, and the field service software the firm already used to dispatch technicians. The only new work was a connection: the arrival times the dispatch software recorded had to reach the invoices where the guarantee credits appear. Nothing was bought for the design of the offer, and the design was not less rigorous for it. What the case also shows is where the limit sits: three tiers and two add-ons fit on one page, so a configure-price-quote tool was not needed, and ten proposals in a pilot do not need randomization software, so an experimentation tool was not needed either.

The stakes are practical rather than aesthetic. A tool that cannot hold an add-on will quietly remove the add-on from the offer, because the person writing the quote will choose the path of least resistance. A tool that records arrival times but cannot report them will turn a guarantee into a promise the business cannot prove it kept. A catalog that lives only in the head of the person who built it will produce two different prices for the same building in the same week. None of these failures is dramatic, and all of them are expensive, because each one is discovered by a customer rather than by the business. The function comes first, the data that flows between tools comes second, and the brand comes last, if it comes at all.

In This Chapter

Section 11.1 sets out the six functions an offer design tool can serve and the minimum a small business needs at each one. Section 11.2 covers the surfaces used for the design work — canvases, whiteboards and journey maps — and how to keep them. Section 11.3 covers surveys, interviews and research tools, including the limits of what a survey can tell a business. Section 11.4 covers catalogs, quotes and configure-price-quote tools, and the point at which quoting outgrows a page. Section 11.5 covers field service and service-level tracking, and the records that make a guarantee real. Section 11.6 covers experiments and analytics, from a spreadsheet of take rates to a randomized test. Section 11.7 covers how to choose and combine tools, with the popular families named as examples. Section 11.8 covers data, access and customer information.

11.1 What Offer Design Tools Do

An offer is designed over months, and the tools that carry the work are ordinary. A wall in the office, a spreadsheet, a form, a catalog, a call log, a dashboard: none of them is sold as an offer design tool, and most of them are already in the business for other reasons. That is the first thing to understand about this chapter. The question is not which product a business should buy in order to design an offer. The question is which function the work requires at each moment, and which of the tools already at hand can perform it.

The life of an offer has six stages, and each one asks a different question. To understand is to learn what customers are trying to get done, in their own words, before anything is designed. To design is to turn those words into a customer profile and a value map, and then into a journey and a blueprint of the service. To package is to decide what sits in each tier, what is an add-on, what is excluded and what is promised. To sell is to put the package in front of a customer as a quote, a proposal or a page, with the contents and the exclusions stated plainly. To deliver is to perform what was promised, on the schedule and to the standard the offer set. To measure is to count what customers chose, what it cost to serve, whether the promises were kept and what the offer earned.

Figure 11.1 draws those six stages as a single line, with the family of tools that serves each stage beneath it. The figure runs from left to right. Under understand sit research and interview tools. Under design sit whiteboards and maps. Under package sit spreadsheets and catalogs. Under sell sit quoting and billing tools. Under deliver sit field service software and help desks. Under measure sit analytics and dashboards. A small business needs one modest tool per stage, and the arrows between the stages matter more than the boxes: what a stage receives and what it hands on is the part that decides whether the offer can be run.

Figure 11.1 — Tool families along the life of an offer. A small business needs one modest tool per stage, and the data that flows between them matters more than the brand of any one.

The table below sets out the same six stages as functions, with what a tool must do at each one, the minimum a small business can get away with, and the point at which a business should consider going further. Read it as a shopping list written before any product is considered: the third column is the one most small businesses need, and the fourth column is a threshold rather than an ambition.

Table 11.1 — The functions of offer design tools

The minimum column deserves a second look, because it is where most small businesses should stop. A wall, a spreadsheet and a form will carry the design of an offer for a firm of seven people, and they will carry it well, provided the wall is photographed and the spreadsheet is kept current. What the minimum cannot do is scale past a certain volume. A spreadsheet that holds three tiers and two add-ons is a tool; a spreadsheet that holds forty options across six customer segments is a liability, because the person who maintains it becomes the only person who can quote.

The fourth column is a threshold, not a destination. A business should go further when a specific pressure appears: when several people edit the same map, when options multiply past what one page can hold, when service levels must be reported rather than believed, or when transactions are numerous enough that a difference of two percentage points is invisible without a test. Until one of those pressures is real, a larger tool adds cost and training without adding an answer.

The data that flows between the stages is the part that decides whether the offer can be run at all. Three flows matter more than the rest. The customers' words must reach the design, or the offer is designed from the business's assumptions. The tier a customer bought must reach the invoice, or the customer is billed for something other than what was sold. The arrival times recorded during delivery must reach the credits the guarantee promises, or the guarantee is a sentence rather than a commitment. Each of those flows crosses two tools, and each of them is where a small business most often finds that two good tools do not talk to each other.

The cost of tool proliferation is real, and it is paid in attention rather than in license fees. A business that adds a tool for each stage ends with six places to look, six sets of logins, six exports to test and six renewal dates to remember, and the person who holds them together is the person who should be designing the offer. The discipline this chapter recommends is the opposite of the one the market sells: fewer tools, chosen for the function, connected where the data must flow, and reviewed on the cadence of Chapter 9.

KEY CONCEPT Choose the function before the brand

The central idea of this chapter is that a business chooses an offer tool by naming the function it must perform and the data it must receive and send, and only then by looking at what is for sale. The order matters because the market is organized by product rather than by function, and a buyer who starts from the product will find a plausible answer for every stage and a reason to buy all of them. A buyer who starts from the function will often find that the answer is already in the building: a wall that is already on the office wall, a spreadsheet that already holds the budget, a form that already collects requests, a dispatch tool that already records when a technician arrived. The test is not whether a tool is good. The test is whether the function is needed now, whether the data it must receive exists, and whether the data it produces has somewhere to go. A tool that passes those three questions is worth its cost; a tool that fails them is a subscription with a login.

11.2 Canvases, Whiteboards and Journey Maps

The design stage of an offer is done on a surface, and the surface is the least expensive part of the whole exercise. Chapter 2 sets out what goes on it: a customer profile holding the jobs customers are trying to get done, their pains and their gains, and a value map holding the pain relievers and gain creators the offer provides, with the fit between the two. Chapter 3 sets out the second surface: the customer journey, with its stages and touchpoints, and the service blueprint, with its physical evidence, its customer actions, its onstage and backstage employee actions, its lines of interaction, visibility and internal interaction, and its support processes. Both chapters describe what to draw. This section describes where to draw it and how to keep it.

The choice between a physical wall and a board application is a choice between two kinds of friction, and neither is wrong. A physical wall is visible to everyone who walks past it, it invites a marker in the hand of whoever has an idea, and it costs almost nothing. Its weaknesses are that it cannot be edited by two people at once, that it cannot be searched, and that it disappears when the office is repainted or the lease ends. A board application is editable from anywhere, keeps a version history, and can be shared with a customer or a colleague who is not in the building. Its weaknesses are that it is invisible unless somebody opens it, that it invites tidying rather than thinking, and that it adds a login to a business that may already have too many.

The practice that works for a small business is to use both, in a fixed order. The thinking happens on the wall, because thinking is fast and physical and benefits from being seen. The record happens in the application or in a photograph, because a record must survive. Nordec's whiteboard sits on the office wall above the depot, where the technicians pass it on their way to the vehicles, and it holds the blueprint of an urgent call. It is photographed at the end of each working session and the photograph is filed with a date. The wall is where the argument happens; the photograph is what the business keeps.

Templates are worth using and worth distrusting in equal measure. A template gives a group a shared vocabulary and a place to start, which saves an hour of blank-page negotiation. It also carries assumptions: a template built for a software product will have boxes for features and releases, and a template built for a consumer brand will have boxes for advertising channels. A service business that fills in the wrong template produces a map of somebody else's offer. The remedy is to start from the template, then delete the boxes that do not apply and add the ones that do, and to write the deletions down, because a box that was removed on purpose is a decision the next person needs to know about.

Versions are the part small businesses most often skip, and they are the part that makes the surface useful over time. An offer changes: a tier is added, an exclusion is tightened, a promise is shortened. If the map is redrawn in place, the business loses the reason for the change and the argument that produced it. If each version is dated and kept, the business can answer the question that arrives two years later, which is why the offer is the way it is. The rule is simple and it costs nothing: every version carries a date, and the version it replaced is kept rather than overwritten.

The blueprint deserves one specific note, because it is the map that most directly shapes what a customer experiences. Chapter 3 describes the technique and its origin in the proposal to design services deliberately rather than let them happen. The practical consequence for a small business is that the blueprint is drawn from the customer's point of view first: the customer's actions are placed on the map before the employee's actions, and the line of visibility is drawn afterwards. A blueprint drawn the other way around, from the inside out, produces a description of the business's process rather than a design of the customer's experience, and it will miss the moments that matter most, which are the ones where the customer and the business meet.

TOOLS OF THE TRADE The minimum kit for offer design

Six functions carry the design of an offer, and a small business can perform all six with modest means. The first is a surface for thinking, either a wall with a whiteboard or a board application, used for the customer profile, the value map, the journey and the blueprint, and photographed or exported at the end of each session. The second is a spreadsheet for the catalog of tiers, add-ons and exclusions, with a row per element and columns for its contents, its cost and its margin. The third is a form for collecting replies, requests and short surveys, with the answers kept beside the form. The fourth is a catalog or list that one person maintains and everyone quotes from, so that two people cannot quote two different prices in the same week. The fifth is a call log or field service record that captures what was promised and when it was performed. The sixth is a dashboard or a monthly page that reads the same measures every month. No brand name belongs on this list; the requirement is the function, and the names of well-known examples appear only in Table 11.2.

11.3 Surveys, Interviews and Research Tools

The understand stage is the one where a small business most often substitutes its own opinion for evidence, and the tools of the stage exist to make that harder. Chapter 2 sets out what the business is trying to learn: the jobs customers are trying to get done, the pains they feel and the gains they are after, in their own words. The tools that serve the stage are an interview guide, a schedule, a way to record with consent, a way to keep notes, and a way to group what was said. None of them is complicated, and the discipline is in the use rather than in the purchase.

An interview guide is a page of questions, and its most important property is that it is short. Six to ten open questions will fill an hour, and a guide that runs to thirty questions produces a conversation in which the customer answers the business's assumptions rather than describing their own work. The guide should open with the customer's week and the building or the process they are responsible for, move to the last time something went wrong, and end with what they would change if they could change one thing. The questions that ask a customer to rank the business's own features belong at the end, if they belong at all, because a customer who has been asked to rank has stopped describing and started grading.

Scheduling is the unglamorous part that decides whether twenty interviews happen or six. The tool is a calendar and a form, and the practice is to book the interviews in blocks rather than one at a time, to confirm the day before, and to keep a list of who has been interviewed and who has not. Nordec interviewed twenty people: eight owner-operator customers, six property-manager customers and six owners of small buildings who had declined a proposal in the last two years. The third group is the one most businesses omit, and it is the group that produced the clearest finding in the case, because the owners who had said no were able to say exactly why.

Recording is where the tools touch the law, and the practice is simple enough to state in a sentence: the customer is told that the conversation will be recorded, agrees before the recording starts, and the recording is kept safely and deleted when it is no longer needed. The Québec note below sets out the practice in full. What matters for the tool choice is that a recording is only useful if somebody can find the part that mattered, which means either a transcript or a set of timestamped notes, and either of those is work that has to be planned rather than assumed.

Note-taking is the tool that most small businesses already have and most use badly. The failure mode is a page of summary written in the business's vocabulary, in which the customer's phrase "the tenants call me at seven in the morning" becomes "tenant management issues". The customer's words are the raw material of the offer, and they are the part that cannot be reconstructed later. The practice is to write the customer's phrases verbatim, in quotation marks, beside the business's interpretation, and to keep a running list of the phrases that recur. In the case, the phrases that recurred became the jobs the offer was designed around: the owner-operator who wanted to "keep the building running and the tenants quiet without becoming a maintenance expert", and the property manager who wanted to "show the owners I report to that their buildings are under control".

Grouping is what turns twenty conversations into a design. The tool can be a spreadsheet with one row per person and one column per element, a board with one card per quotation, or a document with headings. What matters is that the grouping is done by the customers' words rather than by the business's categories, and that the count is visible. In the case, the counting was blunt and it changed the offer: urgent response within ninety minutes was named among the top three elements by eighteen of the twenty people interviewed, the monthly preventive visit by fourteen, the capital planning note by twelve, the monthly report by six and the building health record by four. Those five numbers decided what went into the tiers and what did not.

Surveys are the tool of the stage with the clearest limits, and Chapter 2 sets them out. A survey is good at counting and poor at discovering. It can tell a business how many of its customers would consider a quarterly review, provided the question is written in the customer's language and the sample is not the list of people who already like the business. It cannot tell a business what a customer means by "under control", because the answer is a sentence and not a box. The practical rule is to interview first and survey second: the interviews produce the vocabulary, and the survey counts how widely each word is shared.

QUÉBEC & CANADA NOTE Recording an interview

When a customer interview is recorded, the customer is told before the recording starts and agrees to it, and the recording is kept safely and deleted when it is no longer needed. The agreement is obtained for the recording itself, not for the conversation in general, and it covers what will be done with the recording, who will listen to it and how long it will be kept. In Québec, a recording of a person's voice is personal information under the province's private-sector privacy law, and the business that holds it has the obligations that come with holding personal information: a purpose that is stated and limited, access restricted to the people who need it, security proportionate to its sensitivity, and destruction once the purpose is served. The same reasoning applies to a written note that identifies a customer and to a photograph of a building that shows a tenant. Book H2 (Privacy and Personal Information) treats consent, the person in charge of the protection of personal information, and the retention rules in full, and a business that records interviews regularly should read it before it records the next one.

11.4 Catalogs, Quotes and Configure-Price-Quote

The package stage produces a list, and the sell stage turns the list into a document a customer can read and accept. The tool that holds the list is the catalog, and it is the single most important tool in this chapter, because it is the one that decides whether two people can quote the same offer. A catalog is not a marketing document. It is the authoritative list of what the business sells: every tier, every add-on, every productized service, each with its contents, its exclusions, its price and its cost to serve. In the case, the catalog holds three tiers, two add-ons and one productized entry offer, and it fits on a spreadsheet.

The catalog has one rule that makes it work, and it is the rule of the single list. There is one place where the tiers are defined, and every quote, every proposal, every invoice and every page of the website is produced from it. A business with two lists has two offers, and the second one is discovered by a customer. The rule is easy to state and hard to keep, because the pressure to make an exception is constant: a building that is slightly different, an owner who asks for one thing removed, a manager who wants a discount. The catalog can hold those variations, but it must hold them as entries rather than as improvisations, so that the next person who quotes the same building finds the same answer.

Quoting and proposal tools are the next family, and Book E5 (Proposals, Quotes, Business Cases and Contracts) treats the document itself in full. What matters here is the fit between the document and the catalog. A proposal tool that holds its own copy of the tiers will drift from the catalog within a quarter, and the drift will be invisible until a customer is quoted a price that no longer exists. The two ways to prevent it are to produce the proposal from the catalog, or to keep the proposal template free of prices and insert them from the catalog at the moment of writing. Either way, the tier a customer accepts must reach the invoice unchanged, which is the flow named in section 11.1.

A configure-price-quote tool is the family that sits above quoting, and it is the one most small businesses do not need. The tool takes a set of options and their rules and produces a valid configuration and a price for it, which is valuable when the options are numerous and their combinations are constrained: a building system with forty components, a service with a choice of coverage levels, response windows and reporting frequencies. The threshold is not the size of the business but the number of options and the number of people who must quote them correctly. Nordec did not buy one, and the reason is written in the case: three tiers and two add-ons fit on one page, and a one-page tier comparison in every proposal does the work that a configuration engine would do.

Subscription billing is the family that serves an offer sold as a recurring tier, and it is the one that most directly touches the customer's experience of the offer. A recurring offer needs a system that bills the same amount on the same day, that handles a mid-term change of tier, that applies a credit when a promise is missed, and that produces an invoice a customer can read. The last of those is not a detail. An invoice that shows a single line called "service" tells the customer nothing about what they bought, and an invoice that shows the tier, the add-ons and the credit separately is a monthly reminder of the offer's contents. In the case, the guarantee credit of $150 appears on the next invoice automatically, with no claim to make, and the system that produces that invoice is the system that makes the guarantee credible.

DEFINITION Catalogue, configure-price-quote, field service software and service-level report

A catalog is the authoritative list of what a business sells — every tier, add-on and productized service — with each entry's contents, exclusions, price and cost to serve, kept in one place and used as the source for every quote, proposal and invoice. A configure-price-quote tool takes a set of options and the rules that constrain them and produces a valid configuration with its price, which is useful when the options are numerous and their combinations must be checked. Field service software records the work performed away from the office: the call received, the technician assigned, the arrival time, the work done, the parts used and the outcome, and it is the system of record for what was promised and what was delivered. A service-level report is the periodic reading of those records against the promises the offer makes — the share of calls answered within the committed time, the share resolved at the first visit, the credits owed — and it is the evidence a guarantee rests on.

WATCH OUT The tool that decides the offer

A quoting or billing tool that cannot hold an add-on will quietly remove the add-on from the offer. The mechanism is not dramatic and it does not require anyone to decide anything. The person writing the quote reaches the point where the add-on has no field, and the path of least resistance is to leave it out, or to fold it into the tier price, or to write it in the notes where it will not be billed. The same failure appears in a billing system that cannot apply a credit, in which case the guarantee becomes a promise the business makes and does not keep, and in a catalog that cannot hold an exclusion, in which case the exclusion is discovered by a customer who expected the work to be covered. The remedy is the order of work this chapter recommends: design the offer first, write down every element it contains, and only then check whether the tool can hold each one. A tool that cannot hold an element is a reason to change the tool or to change the element deliberately, and it is never a reason to let the tool decide.

11.5 Field Service and Service-Level Tracking

The deliver stage is where the offer stops being a document and becomes something a customer experiences, and it is the stage where the tools of a service business are most likely to be already in place. A firm that dispatches technicians has a scheduling or field service tool, and that tool is the system of record for the offer's most important promise. What this section asks a business to do is not to buy a tool but to look at the one it has and confirm that it records the four things an offer needs: what was promised, when the work was performed, what was found, and what it cost.

The four records are worth naming precisely, because each one feeds a different part of the offer. The call record holds the moment the request arrived, which is the starting point of every response-time promise. The arrival record holds the moment the technician reached the site, which is the measure the guarantee is written against. The resolution record holds whether the problem was solved at the first visit or needed a second, which is the measure of the offer's quality and the source of its cost. The parts and findings record holds what was used and what was seen, which feeds the building record, the capital planning note and the next visit's checklist. A field service tool that holds all four is doing the work of this chapter; a tool that holds only the first two is a dispatch system with a gap.

The gap matters because the offer's promises are written in the language of those records. Chapter 5 sets out the guarantees: a credit when a technician is not on site within ninety minutes of an urgent contract call, and a first-ninety-days exit for a new customer. The first of those is only meaningful if the arrival time is recorded by a system the customer can trust, and the second is only meaningful if the business knows when the ninety days began. Neither promise requires a new tool. Both require that the tool already in use records the time, that the time is not edited after the fact, and that the report can be produced without an afternoon of manual work.

Service-level reporting is the practice that turns those records into evidence. The report is short and it is read on a fixed cadence: the number of urgent calls, the share answered within the committed time, the share resolved at the first visit, the number of second visits and their reasons, the credits owed, and the disputes. In the case, the last twelve months produced 253 urgent contract calls, 243 of them on site within ninety minutes, which is 96%, and 199 resolved at the first visit, which is 78.7%. The fifty-four second visits had two causes: thirty-seven because a part was missing and seventeen because the work exceeded the two hours included and needed the owner's approval. Those two numbers are the reason the redesign in Chapter 3 exists, and neither of them would have been visible without a tool that records outcomes rather than only arrivals.

Help desk software is the family that serves the other half of the deliver stage, and it is the one a business needs when the offer includes a request line. A help desk records a request, assigns it, tracks it to closure and reports how long it took, which is exactly what a promise such as "non-urgent requests logged and scheduled within two working days" requires. A small business that receives a handful of requests a week can perform the same function with a shared inbox and a list, and the threshold for a dedicated tool is the point at which requests arrive through several channels and the business can no longer say how many are open. Book J6 (Customer Service and Support Operations) treats the design of the request process itself.

The connection between the deliver stage and the sell stage is the one that most often fails, and it is worth stating as a single sentence: the times recorded during delivery must reach the credits promised during the sale. In the case, that connection is the only new work the tool set required, and it is the subject of the case note below. A business that records arrival times in one system and applies credits by hand in another has a guarantee that works until the month is busy, and the month is always busy.

THE RULE A guarantee is only as good as the record that proves it

A credit promised on arrival times requires arrival times recorded by a system the customer can trust. Chapter 5 sets out the guarantees a small business can make, and this chapter states the condition that makes them real. The rule has three parts. The first is that the measure must be recorded at the moment it happens, by the person performing the work, in the system that will report it, because a time reconstructed a week later from memory is not evidence. The second is that the record must not be editable after the fact by the person whose performance it measures, or the guarantee becomes a statement about the business's confidence rather than about its performance. The third is that the report must be producible without special effort, because a guarantee whose credits are calculated by hand each month will be calculated correctly in a quiet month and approximately in a busy one. A business that cannot meet those three conditions should not make the promise, and a business that meets them has an offer its competitors cannot copy without building the same records.

CASE IN POINT Nordec Entretien inc.: the credit that needed two tools to talk

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. The 90-minute guarantee promises a credit of $150 on the next invoice when a technician is not on site within ninety minutes of an urgent contract call, automatically, with no claim to make. The promise crosses two tools. The arrival time is recorded in the field service software the firm already used for dispatch, where the technician closes the call from the site. The credit is applied in the accounting system, where the invoice is produced. Neither tool was bought for the guarantee, and neither could make the promise alone. The work the refresh required was a connection between them and a rule about who checks it: the administrator reads the missed calls at the end of each month and applies the credits before the invoices are issued. At 96% of 253 calls, about ten misses a year cost $1,500, which is 0.38% of contract revenue. The number is small, and the reason it is small is that the record exists.

11.6 Experiments and Analytics

The measure stage is where a small business most often relies on a feeling, and the tools of the stage exist to replace the feeling with a count. Chapter 8 sets out what to measure and how to decide whether a result means anything. This section describes the tools that carry the measurement, and it begins with the one that is already on the desk. A spreadsheet holds the take rate of each tier, the cost to serve, the margin, the number of proposals and the number of signings, and it holds them in a form that can be read every month. For a business of seven people making ten proposals in a pilot, that is the whole of the measurement tool set.

The spreadsheet's most valuable column is the one that states the uncertainty. A close rate of 60% on ten proposals is not the same finding as a close rate of 60% on two hundred, and a business that reads the first as though it were the second will make a confident decision on thin evidence. The arithmetic is not difficult and it does not require a tool: with ten proposals, a 60% close rate carries a margin of error of roughly thirty percentage points, which is an interval of about 30% to 90%. Chapter 8 sets out the reasoning in full. The practical consequence for the tool choice is that a spreadsheet with a column for the sample size is more useful than a dashboard that shows a percentage without one.

Web analytics is the family that serves a business whose offer is bought online in many small transactions. The tool records what visitors did, where they stopped and what they bought, and it becomes useful when the volume is high enough that a difference between two versions of a page is visible above the noise. The threshold is not the size of the business but the number of transactions: a business with a few thousand visitors a day can learn from a comparison, and a business with a few dozen cannot. For a service business selling a contract to a building owner, the transaction is a conversation and a proposal, and the analytics tool measures the wrong thing.

Experimentation tools are the family that randomizes a comparison and reports whether the difference is real. They are the right answer for a business with many small transactions and a steady flow of them, and they are the wrong answer for a business making ten proposals in sixty days. Nordec did not buy one, and the reason is the same as the reason it did not buy a configure-price-quote tool: the volume does not support the instrument. What the firm did instead was write its decision rules before the pilot began, so that the result could be read against a standard rather than against a hope. The rules were a close rate of at least 50% on at least ten proposals, downgrades of at most one renewal in four, Essential at most 40% of new signings, the Building Check converting at least one buyer in four within sixty days, and credits of at most two a month. Five rules, written on a page, in a spreadsheet.

Dashboards are the last family and the one most likely to be bought before it is needed. A dashboard reads the measures of Chapter 9 on a fixed cadence and shows them in one place, which is valuable when the measures are read by several people and the reading is a meeting. For a small business, the same function is performed by a single page produced once a month from the spreadsheet, and the page has an advantage the dashboard does not: it can be read in a meeting without a screen, and it can be annotated with the reason a number moved. The threshold for a dashboard is the point at which the measures are read by people who do not have access to the spreadsheet, or the point at which the reading happens often enough that producing the page by hand has become the bottleneck.

DO THE MATH The tools Nordec already had

Illustrative figures; the case is fictional. Step 1 — The whiteboard was already on the office wall above the depot, and it carried the blueprint of an urgent call; the cost of using it was a marker and a photograph at the end of each session. Step 2 — The spreadsheet was already in use for the budget, and it was extended with rows for the tier costs and margins and with the pilot log; the cost of using it was the time to add the columns. Step 3 — The form was already in the platform the firm used, and it was pointed at the replies to the Building Check letter; the cost of using it was writing six fields. Step 4 — The field service software was already recording the arrival times of technicians on urgent calls, and those times are what the 90-minute guarantee is measured against; the cost of using it was nothing, because the record was already being made. Step 5 — The only new work was connecting the call times to the invoice credits, so that a missed call produces a $150 credit on the next invoice with no claim to make, and the work was a monthly reading by the administrator rather than a purchase. The tool set that designed and ran three tiers, two add-ons, a productized entry offer and two guarantees was therefore the tool set the firm already owned, with one connection added.

11.7 Choosing and Combining Tools

Choosing a tool is a decision with five steps, and the order of the steps is what keeps the decision honest. The first step is to write the function and the data, and it is done before any product is looked at. The function is a sentence: "record the arrival time of a technician on an urgent call and report the calls that missed ninety minutes". The data is two lists: what the tool must receive, and what it must send. A tool whose inputs do not exist in the business, or whose outputs have nowhere to go, is not a tool the business can use, however good it is.

The second step is to list the tools already owned that could perform the function, and it is the step that most often ends the search. A business of seven people owns a spreadsheet, a form, a calendar, a document store and a dispatch system, and those five cover a surprising share of what the market sells. The exercise is not a way of avoiding purchases; it is a way of spending the purchase on the function that genuinely is not covered. In the case, the exercise ended with no purchase at all for the design of the offer, and with one connection to build.

The third step is to try the smallest option with a real case, and the word "real" is doing the work in that sentence. A trial performed on invented data tests the interface; a trial performed on one building, one tier and one month of calls tests the tool. The smallest option is usually the one already in the building, used properly for the first time, and the trial should run long enough to include the awkward case: the mid-term change of tier, the credit that must be applied, the exclusion that must appear on the quote. A tool that handles the ordinary case and fails the awkward one has failed, because the awkward case is where the customer is.

The fourth step is to check access, export and where the data lives, and it is the step that a small business most often skips because it is not about features. Access is who can see what, and it should be the smallest set of people who need it. Export is whether the business can get its records out in a form it can use, which is the question that decides whether the business can leave. Where the data lives is the question of which country and which provider, which matters when the records identify customers or employees. Book K1 (The SME Technology Stack) treats the inventory, the renewal dates and the access review that keep these answers current, and Book K2 (CRM and ERP Systems) treats the systems that hold customer records at length.

The fifth step is to decide and to record the decision with its date. The record is a line in the inventory with the tool, its owner, its cost, its renewal date and the people who hold accounts, and it is what makes the decision reviewable rather than remembered. A decision recorded with its date can be revisited on the cadence of Chapter 9; a decision remembered can only be argued about.

Combining tools is the second half of the choice, and it is governed by a single principle: fewer tools, well connected. The connections that matter are the ones the offer depends on, and there are three in a service business. The tier a customer accepts must reach the invoice, so that the customer is billed for what was sold. The arrival times recorded during delivery must reach the credits, so that the guarantee is kept. The findings recorded on site must reach the building record and the next visit's checklist, so that the offer improves with each call. Each connection is a small piece of work, and each of them is worth more than a new tool, because a connection makes two existing tools do the work of three.

The table below lists the families of tools used in offer design, with well-known examples of each and the questions to ask before choosing one. The names are examples as of September 2026; their inclusion is not an endorsement, and the reader must check current features before choosing, because products merge, rename and change what they do.

Table 11.2 — Popular tools by family, as of September 2026

METHOD Choosing an offer tool in five steps

First, write the function in one sentence and the data in two lists: what the tool must receive and what it must send. Second, list the tools the business already owns that could perform the function, and try the smallest of them before looking at anything for sale. Third, try that option with a real case — one building, one tier, one month of calls — including the awkward case of a mid-term change, a credit or an exclusion. Fourth, check access, export and where the data lives: who can see the records, whether they can be taken out in a usable form, and in which country they are held. Fifth, decide, and record the decision with its date as a line in the inventory, with the tool, its owner, its cost, its renewal date and the people who hold accounts. Deliverable: a dated decision and one inventory line. Pitfall: choosing the tool before writing the function, which produces a subscription that performs a function the business did not need.

WHAT CHANGED Tool names as of September 2026

The names in Table 11.2 are examples as of September 2026, and they are the only place in this book where product names appear. They are given so that a reader can recognize the families, not so that a reader can choose a product from a table. The market for these tools changes faster than a book does: products merge, rename, move features between plans and withdraw features that a business depends on. A name that is accurate on the day this chapter was written may describe a different product a year later, and a family that is served by four tools today may be served by three tomorrow. The reader's obligation is therefore twofold. Check the current features of any tool before choosing it, against the function and the data written in the first step of the method above. And keep the business's own records in a form that survives a change of tool, because the catalog, the customers' words and the offer's decisions belong to the business and not to the product that happens to hold them.

11.8 Data, Access and Customer Information

Every tool in this chapter holds something about a customer, and the first task of this section is to name what. The interview notes hold what a customer said about their building, their tenants and their budget. The recordings hold a customer's voice. The catalog holds what each customer bought and what it costs. The field service records hold which buildings had which failures and when. The call logs hold who called, when, and what was wrong. The forms hold the replies of people who may or may not be customers. None of that is unusual, and all of it is personal information when it identifies a person, which means the business that holds it has obligations that do not depend on the size of the business.

Consent is the first of those obligations and the one most often handled by habit rather than by design. A customer who is interviewed knows they are being interviewed, and that is not the same as knowing that the conversation is recorded, who will listen to it, how long it will be kept and what will be done with it. The Québec note in section 11.3 sets out the practice for recordings. The same reasoning applies to a form: a person who fills in a form to request a Building Check has given their contact details for a purpose, and using those details for a different purpose later requires a reason the person would recognize. Book H2 (Privacy and Personal Information) treats consent, the person in charge of the protection of personal information and the retention rules in full.

Access is the second obligation and the one most easily improved. The question is not whether a tool is secure but who inside the business can see the records, and the answer should be the smallest set of people who need them to do their work. A technician needs the building record for the building they are visiting and does not need the interview notes of the owner. The administrator needs the call records and the credits and does not need the recordings. The owner needs the reports and does not need every note. The practical test is to list, for each tool, the people who can see everything in it, and to ask of each name whether the access is needed or inherited. The access review of Chapter 9 is where that list is kept current.

Keeping the records when a tool is replaced is the third obligation and the one that most often fails. A business that changes its field service software must take the call records with it, or it loses the history that makes the guarantee reportable and the building record useful. A business that changes its form tool must take the replies with it, or the customers' words are gone. The test is the export test of Chapter 9: before a tool is chosen, confirm that its records can be taken out in a form the business can use, and confirm it again at the annual review, because a feature that exists at the start of a contract can be changed during it. The records that must survive a change of tool are the catalog, the customers' words, the offer's decisions and the delivery history, and they belong to the business rather than to the product.

The last point is about the customers' words specifically, because they are the records most likely to be lost and the most valuable to keep. An interview conducted two years ago contains the vocabulary that produced the offer, and a business that has lost it will redesign the offer from its own assumptions the next time. The practice that preserves it is modest: a file per customer, a date on every note, the customer's phrases in quotation marks, and a rule that says where the file lives. The rule matters more than the tool, because the tool will change and the rule will not.

WHO DOES THIS The administrator keeps the tools in order

In a small service business, the person who keeps the offer's tools in order is usually the administrator, and the role is worth describing because it is rarely named. The administrator maintains the catalog, so that the tiers, the add-ons and the exclusions are the same in every quote. The administrator reads the missed calls at the end of each month and applies the guarantee credits before the invoices are issued. The administrator keeps the forms and the replies, files the interview notes under the customer's name with a date, and holds the list of who can see what. The administrator produces the monthly page of measures and brings it to the review. None of that work requires a specialist, and all of it requires a person who is named, because a tool set without an owner becomes a set of subscriptions that nobody can explain. In the case, the administrator's loaded cost is $36 an hour, and the work described here is a few hours a month, which is the smallest line in the offer's cost and the one that holds the rest together.

IN LA MACHINE What the platform covers and what it does not

Five pages of the application copy read for this book touch the tool set of this chapter (La Machine, 2026). Business → Session Packages at productivity/packages, also reached at business/packages, defines packages of sessions or minutes with a name, a type, a quantity or an unlimited allowance, a price and an internal cost, and shows the profit and the margin per sale as the package is edited; the same page lists each client's credit balance and its history. Business → Subscription Packages at business/subscription-packages defines packages billed monthly or yearly, each with a name, a billing period, a price and a cost, and shows the profit per period; a disabled package lets existing subscribers keep their access until the end of their current billing period. Account → Profile → Revenues & Expenses at business/subscriptions shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet, which is the place to count which package sold and how often. Forms at business/marketing/forms, with the answers at business/marketing/forms/responses, is a general form builder that can carry an interview sign-up or a short survey, and the answers are kept with the form. What the platform does not have belongs in the same breath: no tiers, bundles or add-ons for the user's own packages, since each package is a single flat item; no guarantee or warranty register; no service-level tracking and no service credits; no refund workflow and no free trial; no service blueprint, no customer journey map, no survey or net promoter tool beyond the general form builder; no experiment or A/B testing tool; and no quote or proposal module, since the page at Business → Motivational Quotes at business/quotes holds motivational quotations.

Costly Mistakes

• Buying the tool before writing the function. A business that starts from what is for sale will find a plausible product for every stage of an offer's life and a reason to buy all of them, and it will spend its attention on six logins instead of on its customers.
• Letting the quoting tool decide the offer. A tool that cannot hold an add-on, apply a credit or carry an exclusion will remove that element from the offer quietly, and the customer will discover the removal rather than the business.
• Recording arrival times that never reach the credits. A guarantee whose credits are calculated by hand each month is kept correctly in a quiet month and approximately in a busy one, and the customer who is missed is the one who tells other owners.
• Keeping the catalog in two places. A business with a list of tiers in a proposal template and another in a spreadsheet has two offers, and the second one is found by a customer who was quoted the wrong price.
• Skipping the export test before choosing a tool. The moment to discover that a system cannot hand the records back in a usable form is before the contract is signed, not in the week the business has decided to change systems.
• Recording an interview without telling the customer. A recording made without the customer's agreement is personal information held without a basis, and the practice of asking first costs one sentence at the start of the conversation.
• Leaving the tools without an owner. A tool set with no named person becomes a set of subscriptions that nobody can explain, and the first question a new owner asks is the one nobody can answer.

Key Takeaways

• An offer's life has six stages — understand, design, package, sell, deliver and measure — and a small business needs one modest tool per stage, chosen for the function and the data rather than for the brand.
• The minimum kit is a surface for thinking, a spreadsheet for the catalog, a form for replies, a call log for what was promised and performed, and a monthly page of measures, and most small businesses already own all five.
• The data that flows between tools decides whether the offer can be run: the tier a customer accepts must reach the invoice, the arrival times recorded during delivery must reach the guarantee credits, and the findings recorded on site must reach the building record.
• A guarantee is only as good as the record that proves it, which means the measure is recorded when it happens, by the person doing the work, in a system that reports it without manual effort.
• A spreadsheet is enough to measure an offer until the transactions are many and small, and a close rate read without its sample size is a number that invites a confident decision on thin evidence.
• The records that must survive a change of tool are the catalog, the customers' words, the offer's decisions and the delivery history, and they belong to the business rather than to the product that holds them.

Your Action Plan

1. Write the function and the data before looking at anything for sale. For the next tool the business thinks it needs, write the function in one sentence and two lists: what the tool must receive and what it must send. Deliverable: a one-page function and data note. Verification: every item on the two lists exists somewhere in the business today, or the note says where it will come from.
2. Inventory the tools already owned against the six functions. Place the tools the business already has under understand, design, package, sell, deliver and measure, and mark the functions with no tool. Deliverable: a six-row map of existing tools. Verification: each function is either covered by a named tool or recorded as a gap with the reason it is not yet covered.
3. Build the catalog in one place and quote from it. Put every tier, add-on and productized service in a single list with its contents, its exclusions, its price and its cost to serve, and produce the next quote from that list. Deliverable: one authoritative catalog. Verification: two people quoting the same building on the same day produce the same price.
4. Connect the delivery record to the promise. Confirm that the tool used to dispatch records the arrival time, the first-visit outcome and the parts used, and that the report needed for the guarantee credits can be produced without manual work. Deliverable: a monthly service-level reading. Verification: the credits owed for the month are calculated from the record and applied before the invoices are issued.
5. Test one export and name one owner. Take the records out of the tool that holds the delivery history in a form the business can use, and name the person who maintains the catalog, the forms and the access list. Deliverable: one tested export and one named owner. Verification: the export opens in a tool other than the one that produced it, and the owner can say who can see what in each tool.

In La Machine

Five routes in the application copy read for this book touch the tool set of this chapter (La Machine, 2026). Business → Session Packages at productivity/packages, also reached at business/packages, defines packages of sessions or minutes with a name, a type, a quantity or an unlimited allowance, a price and an internal cost, and shows the profit and the margin per sale as the package is edited; the same page lists each client's credit balance and its history, and credits added by hand are recorded as free gifts or adjustments, with no payment. Business → Subscription Packages at business/subscription-packages defines packages billed monthly or yearly, each with a name, a billing period, a price and a cost, and shows the profit per period; a package that is disabled lets existing subscribers keep their access until the end of their current billing period. Account → Profile → Revenues & Expenses at business/subscriptions shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet, which is the place to count which package sold and how often. Forms at business/marketing/forms, with the answers at business/marketing/forms/responses, is a general form builder that can carry an interview sign-up or a short survey, and the answers are kept with the form.

What the platform does not do belongs in the same breath, because a reader who counts a family as covered without checking the page will build an offer on a gap. La Machine has no tiers, bundles or add-ons for the user's own packages, since each package is a single flat item; no guarantee or warranty register; no service-level tracking and no service credits; no refund workflow and no free trial; no service blueprint, no customer journey map and no survey or net promoter tool beyond the general form builder; no experiment or A/B testing tool; and no quote or proposal module, since the page at Business → Motivational Quotes at business/quotes holds motivational quotations. The catalog of tiers, the guarantee register and the service-level report this chapter describes therefore live outside the platform, in a spreadsheet and in the field service software the business already uses. Book N1 (Getting Started with La Machine) maps the modules the platform does cover, Book N2 (Piloting Tasks, Executions and Projects) describes the pages that can carry the design work and its reviews, and Book K1 (The SME Technology Stack) treats the inventory that keeps the whole set of tools, inside and outside the platform, under review.

Self-Check

1. What are the six functions an offer design tool can serve, and what does a small business need at each one?
2. Why does the data that flows between tools matter more than the brand of any one of them, and which three flows does a service business depend on?
3. What must a field service tool record for a response-time guarantee to be credible, and why does the record have to be made at the moment it happens?
4. When does a business need a configure-price-quote tool or an experimentation tool, and why did the case need neither?
5. What records must survive a change of tool, and what is the test that confirms they will?

Answers. 1. The six functions are understanding customers, designing the offer, packaging it, selling it, delivering it and measuring it; a small business needs one modest tool per function, and the minimum is a surface for thinking, a spreadsheet for the catalog, a form for replies, a call log for what was promised and performed, and a monthly page of measures. 2. The flows matter because an offer can only be run if what was sold reaches the invoice, what was promised reaches the record and what was learned reaches the next visit; the three flows a service business depends on are the tier into the invoice, the arrival times into the guarantee credits, and the findings on site into the building record. 3. It must record the moment the call arrived, the moment the technician reached the site, whether the problem was resolved at the first visit, and the parts and findings; the record must be made at the moment it happens because a time reconstructed a week later from memory is not evidence, and a guarantee rests on evidence. 4. A configure-price-quote tool is needed when the options are numerous and their combinations must be checked, and an experimentation tool when transactions are many and small enough for a difference to be visible above the noise; the case needed neither because three tiers and two add-ons fit on one page and ten proposals in a pilot do not support randomization. 5. The catalog, the customers' words, the offer's decisions and the delivery history must survive a change of tool, and the test is the export test: the records are taken out in a form the business can use and opened in a tool other than the one that produced them, before the contract is signed and again at the annual review.

Chapter References

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/packages, business/subscription-packages, business/subscriptions, business/marketing/forms and business/quotes.

Chapter 12 — Practical Advice, Pitfalls and Your Action Plan

Redesigning an offer takes sixty days of disciplined work — an inventory, twenty conversations, a blueprint, three tiers, one guarantee, a pilot with its rules written first — and the business that does it ends with an offer its customers can choose, its team can deliver and its owner can price.

What You'll Be Able to Do

• Name the small number of ideas that decide whether an offer redesign changes anything, from the offer as what the customer buys to the guarantee as a cost the business already pays.
• Recognize the ten mistakes that cost the most in offer design, and match each one to the chapter that treats its remedy.
• Run the first fifteen days of a redesign: the offer inventory with dates and costs, the three signals written on one page, and twenty interviews booked.
• Build the middle of the plan, from the interviews written in the customers' own words to the blueprint of the most important moment, the tiers, the add-ons, the guarantee and the pilot's rules.
• Start a pilot in a limited area with its hypotheses and decision rules written before the first proposal, and read it against those rules rather than against the hope that produced it.
• Measure the result at six months on the measures that matter — tier mix, upgrades and downgrades, add-on sales, entry offers converted, guarantee credits and margin by tier — and put the offer on a calendar for its next review.

Why It Matters

Every chapter of this book has ended with work, and work accumulates into the impression that designing an offer is a large project. It is a project, and it is a smaller one than most owners expect. The case followed through this book is a business of seven people with two vehicles and forty-six buildings under contract, and its redesign was carried out by the owner and the administrator in the margins of a normal year: twenty interviews, one blueprint, three tiers, two add-ons, one guarantee, one entry offer and a sixty-day pilot. Nothing in that list required a consultant, a new system or a budget line. What it required was a sequence, and the sequence is the subject of this chapter.

The stakes are easy to state and hard to feel. A business that never redesigns its offer does not stand still; it drifts, because every free addition, every exception granted to a difficult customer and every service performed without a price is a change to the offer that nobody decided. Chapter 1 measured that drift in the case: four elements added to the contract over three years, each given free, costing $58.50 a month per building and $32,292 a year across forty-six buildings, against a price that rose from $691 to $712 and brought $11,592. The margin of a territory building fell from 59.08% to 52.08%. No decision was made to reduce the margin. The margin was reduced by a series of decisions that were each reasonable on the day.

The other side of the same coin is that a redesign pays in ways that are visible within a season. Six months after the refresh, the case held fifty-two buildings instead of forty-six, its revenue had risen 13.7% to $447,012 a year, its contract margin had risen from 51.15% to 52.35%, and its profit on contracts had risen 16.4% to $19,502.30 a month. Those figures come from a pilot of ten proposals and a roll-out at renewals, not from a campaign. What produced them was an offer that a customer could choose, a team that could deliver it and an owner who could price it, which is the whole of what this book has been asking for.

In This Chapter

Section 12.1 gathers the ideas of the book that change results and states them plainly. Section 12.2 lists the ten most costly mistakes in offer design, with what each one costs and the chapter that treats its remedy. Section 12.3 opens the sixty-day plan with days 1 to 15, the offer inventory and the interviews booked. Section 12.4 covers days 16 to 30, the interviews held and counted and the blueprint drawn. Section 12.5 covers days 31 to 45, the tiers, the add-ons, the guarantee and the pilot's rules. Section 12.6 covers days 46 to 60, the pilot started and the weekly log kept. Section 12.7 sets out the measures read at six months. Section 12.8 tells what happened in the case over those six months and closes the book.

12.1 What Matters Most

A book about the content of an offer can be read as a long list of techniques, and the list is not the point. Eleven chapters have described how to inventory what a business already gives away, how to interview customers about what they did rather than what they say they want, how to draw the moment that decides whether a customer comes back, how to build tiers a customer can tell apart, how to write a guarantee the business can afford, how to decide which services to add and which to refuse, how to turn a custom job into something that can be sold twice, and how to test the whole thing before it is rolled out. What matters most is the small number of ideas underneath those techniques, because the techniques are easy to copy and the ideas are what make them work.

The first idea is that the offer is what the customer buys, not what the business does. A maintenance company performs preventive visits, urgent repairs, inspections and reports; what its customer buys is a building that stays open and a year in which nobody has to think about the boiler. The distinction is not philosophical. It decides what belongs in a tier, what belongs in a report, what a guarantee can promise and what a price can be attached to. A business that describes its offer in the language of its own activity sells hours and parts; a business that describes it in the language of the customer's job sells an outcome, and the outcome is what the customer compares with the alternative.

The second idea is that value is judged against the alternative, never in the abstract. A contract at $712 a month is expensive or cheap depending on what the building would otherwise spend, and the case made that comparison explicit: a building of about 4,500 square metres would buy $3,900 a year of preventive work from a patchwork of trades and pay for eleven unplanned interventions at an average of $560, or $10,060 a year, against $8,544 for the contract, a gain of $1,516 before the note, the record and the report. A small building of about 2,500 square metres would spend $6,620 against $7,080 for the same contract, which is $460 more than the alternative, and that is exactly what the two owners who declined on price had said. The same offer was a gain for one building and a loss for another, and only the comparison revealed it.

The third idea is that each tier must be complete for somebody. A tier is not a discount and it is not a stripped version of the flagship; it is a whole offer that a particular customer can buy without feeling that something essential is missing. The Essential tier in the case is complete for a small building: a monthly visit on a core checklist, urgent calls within ninety minutes day and night with up to two hours of work, the building record and a visit sheet after each visit. It omits the monthly report and the capital planning note, and the note can be bought as an add-on, so the customer who wants it is not excluded; the customer who does not want it is not charged. A tier that is complete for nobody is a tier that exists to make another tier look better, and customers notice.

The fourth idea is that a guarantee costs what the business fails to deliver. The ninety-minute guarantee in the case pays $150 automatically when a technician is not on site within ninety minutes of an urgent contract call. At 96% of 253 calls, about ten misses a year cost $1,500, which is 0.38% of contract revenue; at the 88% the plan targeted in its first year, about thirty misses would cost $4,500. The guarantee is not a promise to be perfect. It is a price on the failure the business already has, written down in advance and paid without argument, and the number is small enough to be afforded and large enough to be felt.

The fifth idea is that a free service is a price decision made by default. Every element added to the case's contract without a price was a decision to reduce the margin, taken by somebody who was not thinking about the margin. The four additions cost $58.50 a month per building, and the price rose by $21, which left $20,700 a year of added cost unpriced. The remedy is not to charge for everything; it is to know which elements are given and why, and to write the reason down. A free element with a reason is a marketing decision. A free element without one is a leak.

The sixth idea is that a test is only a test if its rules were written first. The case's pilot had five decision rules agreed before the first proposal: a close rate of at least 50% on at least ten proposals, downgrades of at most one renewal in four, Essential at most 40% of new signings, the Building Check converting at least one buyer in four within sixty days, and credits of at most two a month. All five passed. The honest reading was still that ten proposals cannot show that the tiers raised the close rate, because a 60% result on ten proposals carries a margin of error of about thirty points. A pilot read without its rules is a story; a pilot read against them is evidence, even when the evidence is that the sample was too small.

The seventh idea is that an offer is reviewed on a calendar, not on a feeling. Chapter 9 gave the case its rhythm: a monthly pass, a quarterly review and an annual day, with the offer's own review placed at the annual catalog review. An offer that is reviewed once a year is reviewed often enough to catch the drift that Chapter 1 described, and rarely enough that the review is a decision rather than a habit. The date is what turns the intention into work, and the work is what this chapter now sets out in sixty days.

DEFINITION Offer redesign, roll-out and catalog review

An offer redesign is the deliberate reworking of what a customer buys: the elements included, the tiers that package them, the add-ons sold beside them, the exclusions that bound them, the guarantee that backs them and the entry offer that introduces them. It is not a price change, although it usually precedes one, and it is not a rebrand, although it may change the words. A roll-out is the moment the redesigned offer is presented to customers and prospects: at renewal for existing customers, in proposals for new ones, in a limited area first so that the result can be read. A catalog review is the dated occasion, once a year, at which the business reads the measures of the offer — the tier mix, the upgrades and downgrades, the add-on sales, the entry offers converted, the guarantee credits and the margin by tier — and decides what to change, what to keep and what to retire. The three are a sequence, and the sequence is the subject of this chapter.

KEY CONCEPT Redesign what customers buy, not what the business likes to do

The central idea of this book is that an offer should be redesigned around what the customer buys rather than around what the business enjoys doing or has always done. The distinction sounds soft and behaves hard. A business that designs from its own activity produces tiers that mirror its internal teams, reports that describe its own effort, guarantees about its own process and a price built from its own costs. A business that designs from the customer's side produces tiers that mirror the customer's situation, reports that answer the customer's question, guarantees about the outcome the customer cares about and a price anchored to the customer's alternative. The case's capital planning note is the clearest example: it is not a maintenance activity and it does not reduce the business's cost, but it answers the owner-operator's job of knowing a year ahead what will have to be replaced, and twelve of twenty interviewees named it among their top three elements. Designing from the customer's side is not a matter of taste. It is the difference between an offer that can be chosen and an offer that has to be explained.

The seven ideas are not independent, and the order in which they are applied matters less than the fact that all seven are applied. A business that redesigns around the customer's job but never prices what it gives away will watch its margin fall. A business that prices carefully but builds tiers nobody can tell apart will watch its proposals stall. A business that builds good tiers and tests them without rules will learn nothing it can act on. The sixty-day plan in the sections that follow is simply these seven ideas arranged in a sequence that a small business can perform, and the sequence is deliberately short because a plan that takes a year is a plan that will be interrupted.

12.2 The Ten Most Costly Mistakes

The mistakes that cost the most in offer design are not exotic. They are the ordinary habits of a business that is busy delivering, and each of them has a remedy that another chapter of this book has already described in full. The table below sets out the ten, with what each one costs, the remedy and the chapter that treats it. Read it as a diagnostic rather than a list of accusations: most businesses recognize three or four of the ten in their own history, and the recognition is the useful part, because a mistake that has a name and a remedy is a mistake that can be corrected in the next sixty days.

Table 12.1 — The ten most costly mistakes in offer design

The first group of mistakes is about accumulation. Adding without subtracting is the mistake that Chapter 1 measured, and it is the most common because it is the least visible: each addition is small, each is given for a good reason, and the total appears only when somebody adds it up. Free services that become promises belong to the same family, and they are worse in one respect, because a free element that customers have come to expect cannot be withdrawn without a conversation. The remedy for both is the same discipline: every element of the offer has a date, a cost and a reason, and the reason is reviewed on a calendar.

The second group is about the customer. Asking customers what they want instead of what they did produces a list of wishes that does not survive a purchase decision, and it is the reason the case's twenty interviews were built around what each person had done in the last two years. Designing the service from the inside produces reports and guarantees that describe the business's own effort; the blueprint of Chapter 3 exists to reverse the order, drawing the customer's actions first and the support processes behind them second. One offer for every segment belongs here too, because it is a decision to serve everybody with something that fits nobody, and its cost is visible in the proposals that are lost at both ends.

The third group is about the packaging. Tiers nobody can tell apart turn a sales conversation into a comparison of lists, and the customer who cannot tell the difference defaults to the middle tier or leaves. Too many options is the same problem with more arithmetic: in the studies of choice, people were more likely to buy jams or chocolates, and were more satisfied with their choice, when they were offered six options rather than twenty-four or thirty (Iyengar & Lepper, 2000), and the case's decision to hold three tiers and two add-ons is that finding applied. A guarantee nobody can collect is a packaging failure as well, because a promise that is hard to invoke is not believed, and the strongest guarantees are unconditional, meaningful to the customer, easy to understand, easy to invoke and easy to collect (Hart, 1988).

The fourth group is about the test. Productizing a custom job puts a fixed price on variable work, and servitizing without the numbers is the larger version of the same error: reading data on 10,028 firms, servitized manufacturers were larger in revenue but earned lower profits as a percentage of sales in aggregate, with the effect depending on size (Neely, 2008). Reading a pilot without its rules is the mistake that wastes the whole exercise, because a result that cannot be interpreted cannot be acted on, and the value of a controlled experiment lies in framing the question and the metric before the test rather than after it (Kohavi & Thomke, 2017).

WATCH OUT Rolling out everywhere after a good pilot

The most expensive moment in an offer redesign is the week after a good pilot result, when the business decides to roll the new offer out everywhere. The temptation is strong and the reasoning sounds sound: the pilot worked, so the offer works, so the sooner every customer sees it the better. The trap is that a pilot is a sample, and a sample read without its written criteria is a story. The case's pilot closed six of ten proposals, which is 60%, against 57.9% for the strategy year, and the honest reading was that ten proposals cannot show a difference, because a 60% result on ten proposals carries a margin of error of about thirty points. What the pilot did show was who could say yes and what existing customers chose. A roll-out that begins before the written criteria are read will be justified by the headline number and will not be corrected by the detail underneath it. Read the criteria first, write down what the pilot can and cannot show, and then roll out.

12.3 Redesigning Your Offer in 60 Days: Days 1 to 15

The first fifteen days of the plan produce two things: an inventory of the offer as it stands, and twenty interviews booked. Neither requires a decision about the future, which is deliberate, because a business that begins by deciding what to change tends to decide before it knows. The inventory is the same discipline Chapter 1 applied to the case's contract, and the interviews are the same discipline Chapter 2 applied to its customers. Fifteen days is enough for both if the work is placed on a calendar rather than left to the gaps between other tasks.

The inventory lists every element of the offer as it is delivered today, with three facts attached to each: the date it entered the offer, the cost to serve it per customer per month, and whether it is priced or given. The case's inventory produced six elements: the monthly preventive visit at $245.60 a month per territory building, dating from 2022; urgent calls within ninety minutes with up to two hours of work included, at $37.13, also from 2022; the building health record's upkeep at $13.50, added in the third quarter of the strategy year; the one-page monthly report at $27.00, added the same year; the capital planning note at renewal at $13.50; and the mid-contract note for flagged buildings at $4.50, decided at the refresh and not yet announced. The total was $341.23 a month against $282.73 for the first two elements, and the four additions cost $58.50 a month per building, or $32,292 a year across forty-six buildings.

The second fact, the cost to serve, is what turns the inventory from a description into a decision. A cost per customer per month is not an accounting exercise; it is the number that tells the business what it is giving away. The case's costs were built from a loaded technician rate of $54 an hour and an administrator's loaded cost of $36 an hour, applied to the hours each element consumes. The building health record's upkeep is a quarter-hour a month, which is $13.50; the one-page monthly report is twenty minutes of a technician and fifteen minutes of an administrator, which is $27.00; the capital planning note is three hours a year, which is $13.50 a month. None of those numbers is large on its own, and together they are the difference between a 59.08% margin and a 52.08% margin.

The third fact, priced or given, is the one that produces the three signals the plan asks for on one page. The first signal is the proposals lost for a reason that points at the offer: in the case, two of the four proposals lost on price went to the trades billed by the call, and both were small buildings under 3,500 square metres whose owners said the contract was more than the building needed. The second signal is what customers asked for that the business had nothing to sell them: five asked for a quarterly review meeting, four for help with replacement projects and three property managers for a single report across their buildings, twelve requests in all. The third signal is the unpriced additions, which in the case were $20,700 a year. Written on one page, those three signals are the agenda for the rest of the plan.

The interviews are booked in the same fifteen days, and the booking is a task rather than an intention. Three groups are chosen: current customers of the main segment, customers of a second segment if the business has one, and people who declined a proposal in the last two years. The case chose eight owner-operator customers, six property-manager customers and six owners of small buildings who had declined, twenty in all. The invitation says what the conversation is for and how long it takes, and it is made by the person who will conduct it. A business that books the interviews in the first fifteen days holds them in the second; a business that leaves the booking to the following month holds four and calls the exercise finished.

Figure 12.1 shows the whole sixty days as a single timeline, and it is worth looking at before the plan is described in detail, because the shape of the plan is what makes it performable. The figure is a horizontal timeline divided into four blocks of fifteen days. The first block is labeled with the inventory and the bookings, and its deliverable is an inventory with dates and costs and twenty interviews on the calendar. The second block is labeled with the interviews and the blueprint, and its deliverable is twenty counted interviews and a drawn moment. The third block is labeled with the tiers and the guarantee, and its deliverable is three tiers, two add-ons, one guarantee with its cost and a set of pilot rules. The fourth block is labeled with the pilot started, and its deliverable is a first week of proposals and renewals recorded. Beyond day 60 an arrow runs to the right, labeled with the pilot's reading, because the pilot itself runs sixty days more.

Figure 12.1 — Redesigning an offer in sixty days. Four blocks of fifteen days, each ending with a deliverable someone else can check; the pilot started on day 46 is read against its rules sixty days later.

The table below sets out the sixty-day plan in its four blocks, with the work of each, the deliverable it produces and the verification that tells the business the block is finished. Read it as a contract with the calendar rather than as a description: each block ends with something another person can check, and a block that ends without its deliverable is a block that has not been completed, however much work it contained.

Table 12.2 — The sixty-day plan

METHOD The first fifteen days

First, list every element of the offer as it is delivered today, with the date each one entered the offer and the person who decided it. Second, cost each element per customer per month, using the loaded hourly rates the business already knows and the hours each element consumes, and write the total beside the price the customer pays. Third, mark each element as priced or given, and add the given elements up for the year across the customer base; that total is the size of the decision the business has been making by default. Fourth, choose the three groups to interview — current customers of the main segment, customers of a second segment, and people who declined a proposal in the last two years — and book twenty conversations with a date, a name and a stated purpose. Fifth, write the three signals on one page: the proposals lost for a reason that points at the offer, what customers asked for that the business had nothing to sell them, and the unpriced additions. Deliverable: one page a reader can act on. Pitfall: beginning with the tiers, which decides the answer before the questions are asked.

QUÉBEC & CANADA NOTE Three dates to check before the roll-out

Three dates belong on the wall of any business preparing a roll-out. The first is October 5, 2026, when sections 38.1, 38.2 and 38.8 of the Québec Consumer Protection Act came into force: new goods of listed kinds sold or leased to consumers — among them ranges, refrigerators, washing machines, television sets, computers, cellular telephones, air conditioners and heat pumps, and other goods set by regulation — carry a warranty of good working order whose duration is set by regulation, covering parts and labor, and the merchant must state that duration prominently near the advertised price. A business that sells listed goods to consumers, or that bundles them with a service, must check its offer against that date. The second is the Competition Act consolidation current to July 21, 2026, last amended March 26, 2026, whose sections 52 and 74.01 govern false or misleading representations, performance claims that are not based on an adequate and proper test, and guarantees that are materially misleading or have no reasonable prospect of being carried out. The third is the federal small business statistics page modified April 23, 2026 (Innovation, Science and Economic Development Canada, 2026a), which is where a business checks the size band it belongs to before comparing itself with anyone. This book explains these laws and does not replace the advice of a lawyer.

12.4 Days 16 to 30

The second fifteen days are the ones that change the answer, and they are the ones a busy business is most tempted to shorten. Twenty interviews is not a large number, and it is large enough to separate the elements that customers value from the elements the business assumes they value. The rule that makes the interviews worth holding is that each one is written the same day, in the customer's own words, before the memory of the conversation is replaced by the memory of the business's own opinion.

The interviews are structured around what the customer did rather than what they want. The questions that produce usable answers are concrete: what happened the last time something failed, who was called, how long it took, what it cost, what the customer would have changed, and what they would have done instead if the contract did not exist. The case's twenty interviews produced counts that no internal discussion had produced: urgent response within ninety minutes was named among the top three elements by eighteen of twenty people; the monthly preventive visit by fourteen; the capital planning note by twelve, including seven of the eight owner customers, two of the six managers and three of the six prospects after seeing a sample; the monthly report by six, including one owner and five managers and no prospects; and the building health record by four. The counts are the point. An element named by eighteen of twenty people is a foundation; an element named by four is a feature that may still be worth keeping and is not worth leading with.

The interviews also produce the requests that the business has nothing to sell, and those requests are the raw material of the add-ons. In the case, four of the six managers asked for a consolidated view across their buildings, three asked for a quarterly review, three asked for a line for tenants' non-urgent requests, all six prospects asked for a smaller contract, and four of the eight owners asked for help with replacement projects. The last of those is the most instructive, because it was a request the business decided not to serve: the wider path of energy-efficiency retrofit projects stays closed unless at least five owners ask for replacement projects after reading their note, and four had asked. A request that is counted and refused is a decision; a request that is never counted is a drift.

The interviews are also where the customer's job is written down in the customer's own words, and the two jobs in the case are worth repeating because they are the sentences the whole offer was built around. The owner-operator wants to keep the building running and the tenants quiet without becoming a maintenance expert, and to know a year ahead what will have to be replaced. The property manager wants to show the owners they report to that their buildings are under control, and never to be the last to know about a failure. Those two sentences explain why the capital planning note matters to an owner and why the monthly report and the consolidated view matter to a manager, and no amount of internal discussion would have produced them.

The value equation is built in the same fifteen days, for two typical customers rather than for an average one. The equation compares what the customer pays for the offer with what the customer would spend on the next best alternative, including the failures the alternative does not prevent. In the case, a building of about 4,500 square metres would buy $3,900 a year of preventive work from a patchwork of trades and pay for eleven unplanned interventions at an average of $560, or $10,060 a year, against $8,544 for the contract, a gain of $1,516 before the note, the record and the report. A building of about 2,500 square metres would spend $2,700 plus seven interventions at $560, or $6,620, against $7,080 for the contract at $590 a month, which is $460 more than the alternative; at the Essential price of $490 a month, or $5,880 a year, the same building gains $740. Two equations, two different answers, and the second one is the reason the Essential tier exists.

The blueprint is drawn in the last days of the block, and it is drawn for one moment rather than for the whole service. Chapter 3 set out the method: the customer's actions are drawn first, then the onstage actions separated from the customer by the line of interaction, then the backstage actions below the line of visibility, then the support processes below the line of internal interaction. The case drew the urgent call, because that is the moment the customer judges the whole contract, and the drawing produced the redesign: a van stock of the twenty parts most often missing, a call to the owner with an estimate before the two included hours are exceeded, and a closing note written into the building record on site. A business that draws one moment properly learns more than a business that draws the whole service roughly.

TOOLS OF THE TRADE What sixty days need

The sixty days are carried by five ordinary things, and none of them is a specialist product. The first is a spreadsheet for the offer inventory, with a row per element and columns for its date, its monthly cost per customer, whether it is priced or given, and the person who decided it. The second is a calendar holding the twenty interviews and the four block reviews, with a date and a name against each, because an interview that is not on a calendar is an interview that will not happen. The third is a whiteboard for the blueprint, large enough to hold the customer's actions, the line of interaction, the line of visibility and the line of internal interaction across the width of the drawing. The fourth is a one-page tier comparison, written for the customer rather than for the salesperson, with the inclusions of each tier in the customer's language and the exclusions stated plainly. The fifth is a pilot log, one line per proposal and per renewal, with the date, the outcome and the reason where one is given. No brand name belongs on this list; the requirement is the function.

12.5 Days 31 to 45

The third fifteen days turn the counts into an offer. This is the block where the business decides what it will sell, and the decisions are made in a fixed order because each one constrains the next: the tiers first, then the add-ons, then the exclusions, then the guarantee, then the price, and finally the pilot's rules. A business that sets the price before the tiers has priced a list; a business that writes the pilot's rules after the first proposal has written a description rather than a test.

The tiers are built from the counts of the interviews, and each tier is given a segment, a job and a sentence. The case built three. Essential, at $490 a month and only for buildings under 3,500 square metres, is complete for a small building: a monthly visit on a core checklist of 2.5 hours, urgent calls within ninety minutes day and night with up to two hours of work, the building record and a visit sheet emailed after each visit, with non-urgent requests handled the next working day, and no monthly report and no capital planning note, the note being available as an add-on at $420 a year. Preventive is the flagship and the cleaned-up version of the current contract: the full monthly visit, urgent calls, the record, the one-page monthly report and the capital planning note at renewal, with non-urgent requests handled the next working morning. Portfolio is the Preventive price plus $175 per building per month, for property managers and complex buildings: Preventive plus the mid-contract capital planning note, a quarterly review meeting, a tenant request line and a consolidated monthly report across the manager's buildings.

The add-ons are built from the requests the interviews counted, and they exist so that a customer who wants one element of a higher tier can buy it without buying the tier. The case built two: the capital planning note for Essential customers at $420 a year, and a quarterly review for Preventive customers at $140 a quarter. The add-ons are also what keeps the tiers honest, because a tier that omits an element is only complete for its segment if the element can be bought separately by the customer who wants it. The rule of thumb is that the add-ons carry the variation and the tiers carry the choice, and a business that puts its variation into the tiers ends with five of them and a customer who cannot choose.

The exclusions are written in the same block, and they are written in the customer's language rather than in the business's. The case's contract excludes parts, work beyond two hours per incident, roof and envelope, elevators, and anything requiring a licensed trade the company does not hold. Those exclusions are not fine print; they are the boundary of the promise, and a customer who does not know them will discover them at the worst possible moment. Writing them plainly is also a protection for the business, because a promise that is not bounded is a promise that will be interpreted by the customer, and the interpretation will be generous.

The guarantee is designed next, and it is designed with its cost beside it. The case's ninety-minute guarantee pays a credit of $150 automatically when a technician is not on site within ninety minutes of an urgent contract call, with no claim to make. At 96% of 253 calls, about ten misses a year cost $1,500, which is 0.38% of contract revenue; at the 88% the plan targeted in its first year, about thirty misses would cost $4,500. The second guarantee, for new customers, allows cancellation within the first ninety days with no penalty and keeps the building record; the onboarding at risk is $324, and if one new customer in fifteen used it, it would cost about $21.60 per signing, or about $238 a year at eleven signings. A guarantee designed this way is a number the owner can approve, and a guarantee designed without its cost is a hope.

The price is set in this block, and it is set with the method of Book A7 (Pricing Strategy and Revenue Models), which is where the price of a tier belongs. This book designs the contents, the exclusions and the promises; it does not explain how a price is chosen, and the case's prices were set by Sophie with that method. What this book asks is that the price be set after the contents, so that the price is attached to something a customer can see, and that the margin by tier be calculated before the offer is presented rather than after it has been sold. The case's margins were 55.18% on Essential, 53.56% on Preventive at $725 and 55.09% on Portfolio at $900, with 61.43% on the $175 increment that Portfolio adds.

The pilot's hypotheses and rules are written last, and they are written before the first proposal. The case wrote five rules: a close rate of at least 50% on at least ten proposals; downgrades of at most one renewal in four; Essential at most 40% of new signings; the Building Check converting at least one buyer in four within sixty days; and credits of at most two a month. Each rule names a number, a period and a threshold, so that the reading on day 60 is a comparison rather than a discussion. A rule that cannot fail is not a rule, and a pilot whose rules cannot fail will produce a result that cannot be used.

THE RULE No change to the offer without its reason and its date

The rule that keeps an offer from drifting is short: no change to the offer without its reason and its date. A change is any element added, removed, repriced or reworded, and the reason is written in one sentence beside it, with the date on which it takes effect and the date on which it will be reviewed. The rule exists because of what Chapter 1 measured: four elements added to the case's contract over three years, each given free, each reasonable on the day, costing $58.50 a month per building and $32,292 a year, against a price that rose by $21 and brought $11,592, leaving $20,700 a year of added cost unpriced and a margin that fell from 59.08% to 52.08%. The rule does not forbid giving an element away. It requires the business to know that it is doing so, to say why, and to put a date on the review, so that the decision can be revisited by somebody who was not in the room when it was taken. A change with a reason and a date is a decision; a change without them is the drift this book exists to prevent.

WHO DOES THIS The owner signs the new offer

In a small business, the redesign is performed by several people and signed by one. The interviews are held by whoever can ask a question and stay quiet for the answer, which in the case was Sophie and the administrator. The inventory and the counts are kept by the administrator. The blueprint is drawn by the person who knows the work best, which in the case was the lead technician, because the urgent call is the lead technician's work and the redesign of the van stock came from that bench. The guarantee and the price are decided by the owner, because they are the two decisions that commit the business's money. The pilot's rules are written by the owner with the person who will keep the log, and they are signed before the first proposal. What the owner signs is not the document but the offer: the inclusions, the exclusions, the guarantee and the price, as one decision with one date. A redesign that nobody signs is a redesign that will be reinterpreted by every person who presents it, and the reinterpretation will always be more generous than the original.

12.6 Days 46 to 60

The last fifteen days of the plan start the pilot, and the pilot is deliberately small. It runs in a limited area, on a limited number of proposals and renewals, for a fixed period, against rules written in the previous block. The case ran its pilot for sixty days on new proposals and on the renewals that fell in the window, with the Building Check offered by letter and call to forty buildings. The purpose of the limit is not caution for its own sake; it is that a result read from a small, defined group can be attributed to the change, while a result read from everything at once cannot.

The pilot starts with the people who present the offer, and it starts with one page. The case trained the people who present the offer on a single page holding the three tiers, their inclusions, their exclusions and the two add-ons, written in the customer's language. The page is not a brochure and it is not a script; it is the minimum that a person needs in order to describe the offer without inventing it. A business that trains its team on a folder of documents will find that each person presents a different offer, and the differences will be invisible until the results are compared. One page, one offer, one price.

The weekly log is what makes the pilot readable, and it is kept from the first week. Every proposal is recorded with its date, the tier presented, the outcome and the reason where one is given; every renewal is recorded with its date, the tier chosen and whether it was an upgrade, a downgrade or a hold; every guarantee credit is recorded with its date and its cause; and every add-on sale is recorded separately from the contract. The case's log produced the numbers that the reading would use: ten new proposals with the three tiers and six signed, eight renewals with one downgrade, five holds and two upgrades, forty Building Check letters with five sales and two conversions, forty-four urgent calls with two missed and $300 of credits. A log kept weekly is a record; a log reconstructed at the end of the pilot is a memory.

The pilot also produces the failures that the plan is meant to find, and the case's failures are worth naming because each one was small and each one was instructive. The Essential visit sheet arrived too late in the day to be useful to the customer, and it was shortened. One Portfolio customer found that its tenants did not use the tenant request line and asked to move back to Preventive at the next renewal, which is a downgrade the rules allowed and the business accepted. The Building Check letter drew replies only after it was rewritten to lead with the capital planning note rather than with the inspection. None of those failures required a change to the tiers; all three required a change to a detail, and the details were found because the pilot was small enough to see them.

What to do when a rule fails is decided in advance, and the decision is simpler than it sounds. A rule that fails is read in three steps. The first is to check the arithmetic: a close rate below the threshold on a small number of proposals may be a sample rather than a signal, and the case's own reading of its 60% result on ten proposals was that the margin of error was about thirty points. The second is to check the execution: a rule about downgrades that fails may mean the tiers are wrong or that the people presenting them are describing the flagship as the default, and the two have different remedies. The third is to change one thing and re-run the pilot, rather than changing the tiers, the price and the guarantee at once. A pilot whose rules fail and whose response is a single, dated change has done its job; a pilot whose rules fail and whose response is a new offer has wasted sixty days.

The reading is prepared in the last days of the block, and it is prepared as a written document with the rules on one side and the results on the other. The case's reading listed five rules and five results: close rate at least 50% on at least ten proposals, 60%, pass; downgrades at most one renewal in four, one in eight, pass; Essential at most 40% of new signings, two of six, pass; the Building Check converting at least one buyer in four within sixty days, two of five, pass; credits at most two a month, two in sixty days, pass. The reading then added the sentence that matters more than the five results: with ten proposals, a 60% close rate has a margin of error of about thirty points, so the pilot cannot show that the tiers raised the close rate; what it shows is who could say yes and what existing customers chose. A reading that states what the pilot cannot show is a reading that can be trusted with what it can.

WHAT CHANGED Nordec's offer, before and after

At the refresh, Nordec's offer was one contract with six elements, four of them added over three years and given free: the monthly preventive visit, urgent calls within ninety minutes with up to two hours of work, the building health record's upkeep, the one-page monthly report, the capital planning note at renewal and the mid-contract note decided but not announced. The price was $712 a month, the margin of a territory building was 52.08%, and there was nothing for a small building to buy. After the pilot, the offer was three tiers — Essential at $490 for buildings under 3,500 square metres, Preventive from $725 for new customers with current customers keeping their price, and Portfolio at the Preventive price plus $175 per building per month — with two add-ons, the capital planning note for Essential at $420 a year and a quarterly review for Preventive at $140 a quarter. A productized entry offer, the Building Check, was added at $950 with the fee credited against the first three months of a contract signed within sixty days. A ninety-minute guarantee paying $150 automatically was added to every tier, and a first-90-days exit was added for new customers. Six months later, the business held 52 buildings, 4 Essential, 42 Preventive and 6 Portfolio, with revenue of $37,251 a month and a contract margin of 52.35%.

12.7 Measuring the Result

The measures that tell a business whether its offer redesign worked are the measures of Chapter 9, read on a calendar rather than on a feeling. They are deliberately few, and each one answers a question the redesign asked. The tier mix answers whether customers can tell the tiers apart. The upgrades and downgrades answer whether the tiers fit the customers who already buy. The add-on sales answer whether the variation is being bought. The entry offers converted answer whether the front of the offer works. The guarantee credits answer whether the promise is affordable. The margin by tier answers whether each tier is worth selling. The revenue per customer answers whether the business is growing or only getting busier.

The tier mix is the first measure because it is the one that shows whether the packaging works. A business that ends a period with almost every customer on the middle tier has built one tier and two decorations; a business that ends with a spread across the tiers has built a choice. The case's mix at six months was four Essential, forty-two Preventive and six Portfolio out of fifty-two buildings, which is a spread with a heavy middle, and the middle is heavy because the existing customer base was moved to Preventive by default at renewal. The reading is not that the spread is wrong; it is that the Essential and Portfolio tiers are doing the work they were built for, one at the small end and one at the manager end, and that the middle is where the current base sits.

The upgrades and downgrades are the second measure, and they are the honest one. A redesign that produces new signings and loses existing customers has not succeeded, and the only way to know is to count both. The case counted one downgrade and five upgrades over six months, with the downgrade coming from a Portfolio customer whose tenants did not use the tenant request line. The downgrade is not a failure of the offer; it is the offer working, because a customer who bought a tier they did not need was able to move to one they did, and the alternative would have been a customer who left. A business that treats every downgrade as a loss will build tiers that customers cannot leave, and customers will leave anyway.

The add-on sales are the third measure, and they are counted apart from the contract run-rate because they are a different kind of revenue. The case sold two quarterly reviews to Preventive customers at $140 a quarter each and one capital planning note to an Essential customer at $420 a year in the six months. Those amounts are small against a run-rate of $447,012, and they matter for a different reason: they show that the customers who wanted an element of a higher tier bought it rather than leaving for it. An add-on that sells is a tier that did not have to be built.

The entry offers converted are the fourth measure, and they answer a question that the tiers cannot. The case sold eleven Building Checks for $10,450 in six months, and four of them converted to contracts. The conversion rate is the number that matters, because the Building Check is a productized offer whose purpose is to produce a contract, and a Building Check that does not convert is a $950 assessment that the business performed at a 46.57% margin and no further value. The measure is also the one that tells the business whether the entry offer is priced right: an entry offer that converts at a high rate may be priced too low, and one that converts at a low rate may be priced too high or sold to the wrong buildings.

The guarantee credits are the fifth measure, and they are read against the failure rate rather than against the revenue. The case paid $750 of credits on 138 urgent calls with five missed, which is 96.4% on site within ninety minutes, against a guarantee that costs $150 a miss. The measure to watch is not the total but the trend: a credit count that rises while the call volume stands still is a service problem, and a credit count that rises with the call volume is arithmetic. The margin by tier is the sixth measure, and it is read per tier rather than in total, because a total margin can hide a tier that is sold at a loss. The revenue per customer is the seventh, and it is the simplest: the case's $37,251 a month across fifty-two buildings is $716.37 per building, against an average contract of $712 at the refresh.

The table below sets out the measures of the case at the refresh and six months later, side by side. Read it as a template rather than as a result: the columns are the ones any business can fill, and the value of filling them is that the same table read at the next catalog review will show which of the movements were the offer and which were the season.

Table 12.3 — Nordec six months after the refresh

DO THE MATH Six months, measured

Illustrative figures; the case is fictional. Step 1 — The building count moves from 46 to 46 + 8 − 2 = 52 buildings, which is 4 Essential, 42 Preventive and 6 Portfolio. Step 2 — Revenue is $37,251 a month, which is $37,251 × 12 = $447,012 a year, against $393,024 at the refresh, a rise of 13.7%. Step 3 — Revenue per building is $37,251 ÷ 52 = $716.37 a month, against an average contract of $712.00 at the refresh. Step 4 — Contract margin is 52.35% against 51.15%, and contract profit is $19,502.30 a month against $16,753.25, a rise of 16.4%. Step 5 — The guarantee paid $750 of credits on 138 urgent calls with 5 missed, which is 96.4% on site within ninety minutes; the add-ons sold beside the contracts were two quarterly reviews at $140 a quarter and one capital planning note at $420 a year, counted apart from the run-rate.

12.8 The Case, Six Months Later

The six months that followed the refresh are the last thing this book has to report, and they are worth telling in order, because the order is what a reader can copy. The refresh itself was a decision to add a second capital planning note, mid-contract, for the buildings whose equipment the first note had flagged. That decision was small, and it was the last small decision the business took before it stopped adding and started designing. The twenty interviews came next, then the blueprint of the urgent call, then the three tiers and the two add-ons, then the two guarantees, then the Building Check, and then the pilot with its five rules written before the first proposal.

The pilot ran for sixty days and it was read on day 60 against its rules. Ten new proposals were made with the three tiers and six were signed, which is 60% against 57.9% for the strategy year; the two Essential signings were both small buildings of the kind that had been lost to the trades the year before, which is the most useful single fact the pilot produced. Eight renewals fell in the window: one customer downgraded to Essential, five stayed on Preventive and two moved up to Portfolio. The Building Check was offered by letter and call to forty buildings, five bought it and two signed a contract within the pilot. Forty-four urgent calls produced two misses and $300 of credits. All five rules passed, and the reading said plainly what the pilot could not show.

The roll-out came at renewals, and it was deliberately unspectacular. Every current customer was offered the three tiers at renewal, with Preventive at the customer's current price as the default, which is why the tier mix six months later is heavy in the middle. The eight new signings were three Essential, four Preventive and one Portfolio. Two buildings were lost. One customer downgraded and five upgraded to Portfolio. The Building Checks reached eleven sold for $10,450, with four converted to contracts. The urgent calls reached 138 with five missed and $750 of credits, which is 96.4% on site within ninety minutes. The add-ons sold beside the contracts were two quarterly reviews and one capital planning note, counted apart from the run-rate. No new customer used the first-90-days exit, which is the quietest result in the list and the one that says most about the onboarding.

Three things did not work, and they belong in the report as much as the results do. The Essential visit sheet arrived too late in the day to be useful, and it was shortened. One Portfolio customer found that its tenants did not use the tenant request line and asked to move back to Preventive at the next renewal, and the business accepted. The Building Check letter drew replies only after it was rewritten to lead with the capital planning note rather than with the inspection. None of the three required a change to a tier, a price or a guarantee, and all three were found because the pilot and the log were small enough to see them.

What Sophie wrote on the offer page for the next annual catalog review is the last item, and it is a good model for any business finishing this plan. She wrote the date of the review, the seven measures of section 12.7 with their values at six months, the three things that did not work with the change made against each, and four questions for the year ahead: whether the Essential tier should be offered to buildings above 3,500 square metres now that two small buildings had signed and stayed; whether the tenant request line should be redesigned or retired, given that one Portfolio customer had asked to move back; whether the Building Check should be sold to buildings outside the two territories, which the strategy's refusals currently forbid; and whether the four owners who had asked for help with replacement projects had become five, which is the condition the strategy set for reopening the wider path. Four questions, each with a number attached, and a date on which they will be answered.

The book closes here, and the reader's next steps are outside it. The price of the tiers designed in this book belongs to Book A7 (Pricing Strategy and Revenue Models), which is where the case's prices were set. The words that carry the offer to a customer — the tier page, the Building Check letter, the guarantee sentence — belong to Book B6 (Copywriting and Persuasive Communication). The proposals and the contracts that put the offer in front of a buyer belong to Book E5 (Proposals, Quotes, Business Cases and Contracts), and the sales conversation itself belongs to Book E1 (The Sales Process: From Prospect to Signed Contract). The processes that make the offer repeatable belong to Book J2 (Processes, SOPs and the Definition of Done), and the platform work of carrying the plan belongs to Book N2 (Piloting Tasks, Executions and Projects). What this book has designed is the offer: what the customer receives, in which package, with which promise. Everything else is the work of putting it in front of somebody who can say yes.

CASE IN POINT Nordec Entretien inc.: the offer six months later

Nordec Entretien inc. is a fictional business created for this encyclopedia; its owner, its employees, its customers, its competitors, its figures and the events described are invented, and any resemblance to a real business is coincidental. Six months after the refresh, Nordec held fifty-two buildings: four on Essential, forty-two on Preventive and six on Portfolio. Revenue was $37,251 a month, a run-rate of $447,012 a year, up 13.7% from $393,024, and $716.37 per building. Contract margin was 52.35% against 51.15% at the start, and contract profit was $19,502.30 a month against $16,753.25, up 16.4%. The six months had brought eight new signings, two buildings lost, one downgrade and five upgrades to Portfolio. One hundred and thirty-eight urgent calls produced five misses and $750 of credits, which is 96.4% on site within ninety minutes. Eleven Building Checks were sold for $10,450 and four converted to contracts. Two quarterly reviews and one capital planning note were sold as add-ons, counted apart from the run-rate. No new customer used the first-90-days exit. The Essential visit sheet was shortened, one Portfolio customer asked to move back to Preventive, and the Building Check letter was rewritten to lead with the capital planning note.

IN LA MACHINE Carrying the sixty days in the platform

The application copy read for this book carries the sixty days in four places (La Machine, 2026). The plan itself is a project at productivity/projects, with its tasks at productivity/tasks; in the Control center, Plan tasks at /taches gives each task a verification criterion, which is where the deliverables of the four blocks are written so that a block is finished only when its criterion is met. The twenty interviews and the four block reviews are meetings at productivity/meetings, which holds the date, the people and the notes, and the notes are where the customer's own words are kept. The tiers are packages at productivity/packages, where each package carries a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost — the page shows the profit and the margin per sale as the package is edited — a description and an image. Sales by package are counted at Account → Profile → Revenues & Expenses (business/subscriptions), which shows three tables with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet: the place to count which tier sold, how often. What the platform does not have belongs in the same breath: no tiers, bundles or add-ons for the user's own packages, each package being a single flat item; no guarantee or warranty register; no service-level tracking and no service credits; no refund workflow and no free trial; and no experiment or A/B testing tool. The pilot log and the guarantee credits are kept outside it.

Costly Mistakes

• Beginning with the tiers. A business that decides what to sell before it has inventoried what it gives away and interviewed twenty customers is deciding the answer before it has asked the question, and the tiers it builds will mirror its own assumptions.
• Booking the interviews without a date. Twenty conversations that are left to the gaps between other tasks become four conversations and a conclusion drawn from them, which is how a redesign acquires the shape of what the owner already believed.
• Writing the pilot's rules after the first proposal. A rule written after the result is a description of the result, and a pilot read against a description teaches nothing that can be repeated.
• Rolling out everywhere on a headline number. A close rate of 60% on ten proposals carries a margin of error of about thirty points, and a roll-out justified by the headline will not be corrected by the detail underneath it.
• Training the team on a folder instead of a page. When each person presents a different version of the offer, the differences are invisible until the results are compared, and by then the customers have been given three different promises.
• Treating every downgrade as a loss. A customer who moves from a tier they did not need to one they do is a customer who stayed, and a business that makes downgrades difficult will lose the customer instead.
• Reviewing the offer on a feeling instead of a date. The catalog review is what catches the drift that Chapter 1 measured, and an offer reviewed when somebody notices is an offer reviewed after the margin has already fallen.

Key Takeaways

• The offer is what the customer buys, not what the business does, and value is judged against the customer's next best alternative rather than in the abstract; the case's contract was a gain of $1,516 a year for a building of about 4,500 square metres and a loss of $460 for one of about 2,500 square metres.
• Each tier must be complete for somebody, the add-ons must carry the variation, and the exclusions must be written in the customer's language, because a promise that is not bounded will be interpreted generously by the customer.
• A guarantee costs what the business fails to deliver: the case's ninety-minute guarantee pays $150 a miss, which is about $1,500 a year at 96% and about $4,500 at the 88% the plan targeted, and the strongest guarantees are unconditional, meaningful to the customer, easy to understand, easy to invoke and easy to collect (Hart, 1988).
• Every free element is a price decision made by default; the case's four unpriced additions cost $58.50 a month per building and $32,292 a year, against a price rise that brought $11,592, leaving $20,700 a year of added cost unpriced and a margin that fell from 59.08% to 52.08%.
• A test is only a test if its rules were written first, and a pilot read without them is a story; the case's five rules all passed, and the honest reading was still that ten proposals cannot show that the tiers raised the close rate (Kohavi & Thomke, 2017).
• The sixty days end with a started pilot and a prepared reading, and the offer then goes on a calendar: six months after the refresh the case held 52 buildings, $447,012 a year, a contract margin of 52.35% and $19,502.30 a month of contract profit, up 16.4%.

Your Action Plan

1. Build the offer inventory in the first fifteen days. List every element of the offer as it is delivered today, with the date it entered, its cost per customer per month and whether it is priced or given, and add the given elements up for the year across the customer base. Deliverable: an inventory with dates, costs and a priced-or-given mark for every element. Verification: every element has a date and a monthly cost, and the total of the given elements is written on one page.
2. Book and hold twenty interviews, and write each one the same day. Choose current customers of the main segment, customers of a second segment and people who declined a proposal in the last two years, and write each conversation in the customer's own words before the day ends. Deliverable: twenty written interviews with the elements counted. Verification: the counts are written down, and each interview records what the customer did rather than what they want.
3. Draw one blueprint and build two value equations. Draw the moment the customer judges the whole offer, with the customer's actions first and the support processes behind them, and build the value equation for two typical customers against their next best alternative. Deliverable: one blueprint and two value equations. Verification: the blueprint shows the line of interaction and the line of visibility, and each equation names the alternative and its cost.
4. Build the tiers, the add-ons, the guarantee and the price, then write the pilot's rules. Give each tier a segment, a job and a sentence; write the exclusions plainly; design the guarantee and calculate its cost at the current failure rate; set the price with Book A7 (Pricing Strategy and Revenue Models); and write the pilot's hypotheses and decision rules with their numbers and dates. Deliverable: three tiers, two add-ons, one costed guarantee, a price list and a dated set of pilot rules. Verification: each tier is complete for a named segment, and the pilot's rules are signed before the first proposal.
5. Start the pilot, keep the weekly log and prepare the reading for day 60. Train the people who present the offer on one page, record every proposal, renewal, credit and add-on sale weekly, and write the reading with the rules on one side and the results on the other, including what the pilot cannot show. Deliverable: a started pilot, a weekly log and a written reading. Verification: the log holds a line for every proposal and renewal with its date and outcome, and the reading states the margin of error of its own result.

In La Machine

Four routes in the application copy read for this book carry the sixty days (La Machine, 2026). The plan is a project at productivity/projects, with its tasks at productivity/tasks; in the Control center, Plan tasks at /taches gives each task a verification criterion, which is where the deliverable of each fifteen-day block is written so that a block is finished only when its criterion is met. The twenty interviews and the four block reviews are meetings at productivity/meetings, which holds the date, the people and the notes, and the notes are where the customer's own words are kept rather than a summary of them. The tiers are packages at productivity/packages, where each package carries a name, a type, a quantity or an unlimited allowance, a price in Canadian dollars, an internal cost — the page shows the profit and the margin per sale as the package is edited — a description and an image; the same page lists each client's credit balance and its history. Sales by package are counted at Account → Profile → Revenues & Expenses (business/subscriptions), which shows three tables — subscriptions, session and meeting packages, and product purchases — with filters, a strip of revenues, costs and profit for the period, and an export to a spreadsheet: the place to count which tier sold, how often.

What the platform does not do belongs in the same breath, because a reader who expects the sixty days to be carried entirely by the application will be disappointed. La Machine has no tiers, bundles or add-ons for the user's own packages, each package being a single flat item, so the tier structure designed in this book is described in the package's description and in the one-page comparison rather than enforced by the platform. It has no guarantee or warranty register, no service-level tracking and no service credits, so the ninety-minute guarantee and its $150 credits are recorded outside it. It has no refund workflow and no free trial, so the first-90-days exit is a contractual term rather than a platform feature. It has no experiment or A/B testing tool, so the pilot's rules and its weekly log live in a spreadsheet. The tasks and calendar pages, productivity/tasks and productivity/calendar, can hold the block reviews as tasks, which is a place to put the work rather than a way of scheduling it; Book N2 describes the tasks pages.

Self-Check

1. What are the seven ideas of this book that change results, and why does the order in which they are applied matter less than the fact that all seven are applied?
2. Which of the ten most costly mistakes in offer design does a business most often recognize in its own history, and what does each of the four groups of mistakes have in common?
3. What does each of the four fifteen-day blocks of the sixty-day plan produce, and what verification tells the business a block is finished?
4. Why are the twenty interviews written the same day and in the customer's own words, and what do the counts of the elements named among each person's top three decide?
5. What are the seven measures read at six months, and what did the case's measures show?

Answers. 1. The seven ideas are that the offer is what the customer buys rather than what the business does; that value is judged against the customer's next best alternative; that each tier must be complete for somebody; that a guarantee costs what the business fails to deliver; that a free service is a price decision made by default; that a test is only a test if its rules were written first; and that an offer is reviewed on a calendar rather than on a feeling. The order matters less than the completeness because the ideas fail in combination: a business that designs around the customer's job but never prices what it gives away watches its margin fall, and a business that prices carefully but builds tiers nobody can tell apart watches its proposals stall. 2. Adding without subtracting is the mistake most businesses recognize, because each addition was small and reasonable on the day and the total appears only when somebody adds it up; the four groups are accumulation, the customer, the packaging and the test, and what they have in common is that each one is a decision taken by default rather than by a person who wrote down the reason and the date. 3. Days 1 to 15 produce an offer inventory with dates and costs, twenty booked interviews and a one-page signal sheet; days 16 to 30 produce twenty written interviews with counts, two value equations and one blueprint; days 31 to 45 produce three tiers, two add-ons, one costed guarantee, a price list and a dated set of pilot rules; days 46 to 60 produce a started pilot, a one-page presentation, a weekly log and a prepared reading. Each block is finished when another person can check its deliverable, which is why the verification column names what must be true rather than what must be done. 4. They are written the same day because the memory of the conversation is quickly replaced by the memory of the business's own opinion, and they are written in the customer's own words because the sentence the customer uses is the sentence the offer will be built around; the counts decide which elements are foundations and which are features, as when urgent response was named by eighteen of twenty people and the building health record by four. 5. The seven measures are the tier mix, the upgrades and downgrades, the add-on sales, the entry offers converted, the guarantee credits, the margin by tier and the revenue per customer; six months after the refresh the case held 52 buildings — 4 Essential, 42 Preventive and 6 Portfolio — with revenue of $37,251 a month and $447,012 a year, up 13.7%, $716.37 per building, a contract margin of 52.35% against 51.15%, contract profit of $19,502.30 a month against $16,753.25, up 16.4%, 138 urgent calls with 5 missed and $750 of credits, and 11 Building Checks sold for $10,450 with 4 converted.

Chapter References

Competition Act, R.S.C. 1985, c. C-34, ss. 52 and 74.01 (Canada). Justice Laws Website, current to July 21, 2026 (last amended March 26, 2026) — read September 22, 2026.

Consumer Protection Act, CQLR c P-40.1, ss. 1(e.1), 37, 38, 38.1, 38.2, 38.8, 41 to 44, 53 and 54 (Québec). Sections 1(e.1), 37, 38, 41 to 44, 53 and 54 read September 22, 2026, in the Office de la protection du consommateur's notice on the legal warranty and in a consolidation of October 31, 2020, their wording matching the current text wherever it could be compared; sections 38.1, 38.2 and 38.8, added by S.Q. 2023, c. 21, s. 3, as read for Book H1 (Contracts and Commercial Agreements) in the consolidation of June 12, 2026.

Hart, C. W. L. (1988, July). The power of unconditional service guarantees. Harvard Business Review. https://hbr.org/1988/07/the-power-of-unconditional-service-guarantees

Innovation, Science and Economic Development Canada. (2026a). Key small business statistics 2025 (page modified April 23, 2026). Retrieved September 22, 2026, from https://ised-isde.canada.ca/site/sme-research-statistics/en/key-small-business-statistics/key-small-business-statistics-2025

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too much of a good thing? Journal of Personality and Social Psychology, 79(6), 995–1006. https://doi.org/10.1037/0022-3514.79.6.995

Kohavi, R., & Thomke, S. (2017, September–October). The surprising power of online experiments. Harvard Business Review. https://hbr.org/2017/09/the-surprising-power-of-online-experiments

La Machine. (2026). Documentation — route reference inventory. Internal product documentation, read September 22, 2026, on the application copy of September 7, 2026, together with the code of the pages this book describes: productivity/projects, productivity/tasks, /taches, productivity/meetings, productivity/packages and business/subscriptions.

Neely, A. (2008). Exploring the financial consequences of the servitization of manufacturing. Operations Management Research, 1(2), 103–118. https://doi.org/10.1007/s12063-009-0015-5