Amazon Named Its Cancellation Process “The Iliad”: Subscription Pricing From the Customer’s Side
Note: this segment is from John Oliver’s Last Week Tonight, and it comes with the profanity, crude humor, and political asides the show is known for. Our textbook doesn’t use material like this, and I wouldn’t play it in a classroom without watching it first. That said, the marketing content is unusually good, and several instructors have told me their students will watch a comedy segment they would never watch as a news clip. It works best as an out-of-class assignment or an online discussion board prompt, which is how I’d use it. Your students, your call.

Maggie’s printer kept printing one page. The page said it could not print. There was ink in the cartridge, but her HP Instant Ink subscription had lapsed, and when that happens HP remotely disables the cartridges you thought you had bought. That story is the hinge of a recent Last Week Tonight segment on subscriptions (“Subscriptions,” August 24, 2026; HBO; the full segment is on YouTube). It is twenty-odd minutes on how subscription pricing spread from newspapers and cable to litter boxes, cruise control, and a Taco Bell pass that delivered one taco a day for $5 to $10 a month.
Maggie’s printer kept printing one page. The page said it could not print. There was ink in the cartridge, but her HP Instant Ink subscription had lapsed, and when that happens HP remotely disables the cartridges you thought you had bought. That story is the hinge of a recent Last Week Tonight segment on subscriptions (“Subscriptions,” HBO; the full segment is on YouTube). The business logic is stated plainly by the firms themselves: HP’s chief executive told investors the company makes more money per customer inside the subscription program, and that the money is consistent and recurring. General Motors expects as much as $25 billion a year from subscriptions by 2030 — BMW already charges about $20 a month for enhanced cruise control.
Nearly half of Americans in one survey were still paying for a subscription they had forgotten or did not use. Much of that is design. The segment walks through dark patterns — deceptive design choices that steer people toward decisions they did not intend. The FTC accused Amazon of using them to enroll people in Prime; a former employee told reporters the company had been deliberately confusing, and internal documents show fixes were proposed, tested, found to reduce subscription growth, and shelved. Amazon’s cancellation flow — four pages, six clicks, fifteen options — was known internally as the Iliad. Amazon settled for $2.5 billion and denies intending to deceive. Adobe was cited for burying an early termination fee that one internal email called a bit like heroin for the company. The proposed fix, an FTC rule called click-to-cancel, was vacated by a federal appeals court on procedural grounds days before taking effect. Oliver’s closing argument is the one worth handing to students, because it is a marketing argument rather than a legal one: firms could redirect the effort they spend trapping customers into building products people want to keep paying for.
Claude was used to generate a first draft of this blog post.
Relevant Chapters in Essentials of Marketing
Chapter 18 (price setting in the business world) is the primary anchor, and it defines the term the whole segment is about: with subscription pricing, customers pay on a periodic basis for access to a product. The chapter explains why the model spread beyond newspapers and cable — it gives a seller a predictable revenue stream and enhances customer lifetime value, and breaking the price into smaller pieces changes how customers perceive it. Note how closely the textbook’s reasoning matches what HP’s CEO said out loud to investors. Chapter 17 (pricing objectives and policies) deserves a mention too, and it comes with a nice coincidence: Chapter 17 already uses HelloFresh as an example of introductory price dealing, offering free meals to customers who sign up with an auto-renewing subscription. HelloFresh is also the company whose cancellation maze Oliver walks through step by step. Students can look at the same firm from both ends of the relationship.
Chapter 2 (marketing strategy planning) is the second connection, and it is the one that makes this more than a consumer-complaint story. Customer lifetime value is the total profits a single customer contributes to a firm over the length of the relationship, and Chapter 2 says marketing strategies aim to increase CLV through customer acquisition, retention, and enhanced purchasing. Subscriptions work on all three at once, which is precisely why firms like them. But Chapter 2 grounds CLV in loyal customers who continue to purchase from brands that satisfy them. That gives students a sharp question: retention obtained by making the exit hard is still retention, and it still shows up in this quarter’s subscriber count — but is it the same thing the chapter is describing?
Chapter 14 (personal selling and customer service) is the third connection, and it lands harder than you might expect. The chapter says all types of personal selling help win customers, but that effective customer service is especially critical in keeping them, and it describes service reps as the salespeople who promote the next purchase by making sure the customer is satisfied with the last one. Chapter 14 also covers AI-powered chatbots, and it is specific about their proper role: handling simple, routine requests so human reps are free for complex or sensitive problems. A cancellation is a simple, routine request. The clips in this segment show chatbots deployed to do the opposite. And the chapter’s observation that dissatisfied customers often never reach out — they leave, or they voice their frustrations online — describes the TikTok videos Oliver is playing almost exactly.
Class Discussion Ideas
This is a strong out-of-class assignment and an even better online discussion board prompt, since the segment is funny enough that students will actually watch it and opinionated enough to argue with. The through-line for class is a single uncomfortable question: subscription pricing and customer lifetime value are standard, defensible marketing tools, so where exactly does a legitimate retention strategy turn into something else? The activities below work toward that line rather than assuming students already know where it sits.
In-Class Activities
Audit Your Own Subscriptions. Before class, have each student write down what they think they spend monthly on subscriptions, then check their phone settings and bank statement and write down the real number. Bring both figures. Open class by collecting the gaps on the board and comparing them to the survey finding in the segment. It takes five minutes and it makes every later abstraction concrete. (Chapter 18)
Map the Maze. In small groups, have students diagram the HelloFresh or Amazon cancellation path as a flowchart, labeling each screen with the specific technique used to redirect the customer. Then have each group redesign the flow to satisfy a click-to-cancel standard while still giving the firm one legitimate chance to make a retention offer. Present both versions. (Chapter 14)
Two Ways to Keep a Customer. Teams calculate customer lifetime value for two versions of the same subscriber: one retained because she is satisfied, one retained because canceling is too much trouble. Give them a monthly fee, a margin, and two different assumptions about how long each stays and what each says about the brand afterward. The point is to make students decide what to do with the customer who leaves angry and posts about it. (Chapter 2)
Where Is the Line? Works in class or as an online discussion board. Assign students a position rather than letting them choose: half must defend a specific practice from the segment as legitimate marketing, half must argue it crosses a line. Require both sides to name the principle they are using, not just the outcome they prefer. Free trials that auto-convert, retention offers on the cancellation page, and remotely disabling ink are usefully different cases. (Chapter 18)
Discussion Questions
- What is subscription pricing, and why has it spread to products that never used to be sold that way? (Chapter 18)
- Answer: Chapter 18 defines it as customers paying on a periodic basis for access to a product, long standard for newspapers, magazines, insurance, and cable. The chapter gives two reasons for the spread: it produces a predictable revenue stream, and it enhances customer lifetime value. Breaking the price into smaller pieces also changes how the cost feels to the customer. HP’s CEO stated the same logic to investors almost word for word, which makes this a useful example of a textbook mechanism showing up unedited in a firm’s own language.
- How does a subscription raise customer lifetime value? (Chapter 2)
- Answer: Chapter 2 defines CLV as the total profits a single customer contributes over the length of the relationship, and says strategies increase it through acquisition, retention, and enhanced purchasing. A subscription works on all three: it lowers the barrier to acquisition with a trial, it makes retention the default rather than a decision, and it creates a standing channel for selling more. Students should also notice what the chapter assumes underneath — that loyal customers keep buying because they are satisfied. That assumption is what the rest of this discussion tests.
- The former Salesforce executive describes a connected washing machine that knows how often you wash, when detergent is running low, and that you have had a baby. Make the marketing case for and against. (Chapters 2, 7, and 18)
- Answer: For: this is a real service improvement and a real CLV play — a manufacturer that touches a household once every seven years becomes a weekly presence, and the customer gets detergent before running out. Against: the customer is paying partly in data, and rarely knows the exchange rate. Students should be pushed to say what would make the trade fair rather than just declaring it creepy. Disclosure at purchase? An opt-out that does not disable the appliance? Most will land somewhere between “this is a service” and “this is surveillance,” and that middle is the productive part of the discussion.
- A customer opens a chat window and types “cancel my subscription.” The bot asks for details and offers to help her stay. What does Chapter 14 say the chatbot should have done? (Chapter 14)
- Answer: Chapter 14 is specific that AI chatbots are well suited to simple, routine requests — order status, returns, store policy — so that human reps are freed for complex or sensitive problems. A cancellation is about as routine as a request gets. Using the bot to slow it down inverts the chapter’s logic: the technology is deployed against the customer’s stated intent rather than in service of it. Students might reasonably counter that a single retention offer is normal and expected; the useful follow-up is asking how many offers, and after how many refusals, it stops being service.
- Chapter 14 says effective customer service is especially critical in keeping customers. Is a difficult cancellation process a customer service failure or a retention success? (Chapter 14)
- Answer: This is the question with the least clean answer, which is why it is worth asking. In the short run it is plainly a retention success — the subscriber count is higher this quarter than it would otherwise be, and firms measure that. Chapter 14 frames service reps as people who promote the customer’s next purchase by making sure she was satisfied with the last one, and friction produces no such satisfaction. The chapter also notes that dissatisfied customers frequently do not complain to the company; they leave and voice frustration publicly, which is what the videos in the segment are. Strong answers will point out that both statements are true at once and that the disagreement is really about the time horizon being measured.
- Amazon tested fixes for accidental sign-ups, found they reduced subscription growth, and shelved them. Evaluate that decision on marketing grounds, not legal ones. (Chapters 2 and 18)
- Answer: The test measured one variable — subscription growth — and the decision followed the number, which is defensible process applied to the wrong metric. Chapter 2 would ask about customer lifetime value and customer equity, the earnings stream from all present and future customers, and an accidental subscriber has a poor lifetime value profile: low engagement, high likelihood of cancellation, and a meaningful chance of becoming a public detractor. Students should try to name the metric Amazon should have run the test against. That is a harder and more useful exercise than declaring the decision unethical.
- Should a rule requiring symmetric sign-up and cancellation be necessary at all? (Chapters 17 and 18)
- Answer: Open-ended, and the right note to end on. Chapter 17’s coverage of the legal environment for pricing establishes that a good deal of pricing law exists precisely because some practices are profitable and hard for individual customers to counter. Students who oppose the rule should be pressed on what market mechanism they expect to correct the behavior, given that switching costs are the point. Students who support it should be pressed on what it costs firms with legitimate retention programs. One observation worth surfacing: a rule of this kind binds only firms whose retention depends on friction. For everyone else it changes nothing, which is itself an argument.
