Prime's Flywheel Is Real. It's Also Partly Coerced.
Prime members spend roughly 2× non-members and renew at ~93% in year one. The loyalty story is real — but in 2025 the FTC made Amazon pay $2.5B for engineering a slice of that retention with dark patterns and a cancellation maze its own staff called the 'Iliad.'
Comes with a free Switching-Cost Ledger template — plus a worked example for Amazon.
For $79 a year, you used to get one thing: unlimited two-day shipping on eligible items.1 No video, no music, no grocery delivery — those came later. Prime began in 2005 as a bet that if you paid up front to never think about shipping again, you'd stop comparison-shopping and just buy from Amazon. The bet worked almost too well. Today a Prime household spends roughly twice what a non-member spends5, and once you're in, you almost never leave.6 That is the flywheel everyone admires. It is also, in part, a flywheel a federal regulator just fined Amazon $2.5 billion for spinning by force.7
The official story is that Prime is the purest loyalty engine in retail: people love the service, so they stay. The truer story is messier. People do stay — but a documented slice of them stayed because Amazon made joining easy by accident and leaving hard on purpose. The loyalty is real and the coercion is real, and you cannot understand the moat without holding both at once.
The sunk cost that makes you spend more
Here's the mechanism, worked down to the psychology. A membership fee is a sunk cost the moment you pay it, and a paid-up member behaves differently from a free shopper: every purchase elsewhere now feels like leaving money on the table. So you funnel more of your spending to the place you've already paid to belong to. That's why the spend gap is so wide — CIRP's 2024 read puts members at about $1,170 a year against $570 for non-members.5 The fee doesn't just cover shipping; it reframes the customer's entire decision rule. And the more you buy to justify the fee, the more indispensable the fee feels — which is the loop closing on itself.
Amazon then spent fifteen years thickening that loop with reasons to never cancel. Streaming wasn't even a launch feature — the video precursor arrived in 2006 and streaming wasn't folded into Prime until February 2011, six years after the shipping promise.2 Each addition raised the cost of walking away, not by raising the price but by raising the number of things you'd lose. That is what lets Amazon keep nudging the price up — $79 to $99 to $119 to $139 across the years8 — without the wheel slowing down. You're not paying for shipping anymore. You're paying not to unplug your life.
| The loyalty story | The fuller picture | |
|---|---|---|
| Why members stay | They love the service | Love — plus engineered enrollment and exit friction |
| Member spend vs. non-member | ~2× ($1,170 vs $570) | Real, but the headline $1,400 figure is stale |
| Retention ~93% | Voluntary loyalty | Partly manufactured, per FTC documents |
| Tail risk | None worth mentioning | $2.5B settlement; reputational exposure |
The part Amazon called the 'Iliad'
Then the curtain moved. In 2023 the FTC alleged that Amazon used deceptive design — 'dark patterns' — to enroll people in Prime without clear consent, and that it built a cancellation process complicated enough on purpose that staff internally nicknamed it the 'Iliad,' after the long epic.7 In September 2025, Amazon settled for $2.5 billion: a $1 billion civil penalty plus $1.5 billion in refunds to customers, with internal documents showing employees had knowingly discussed the deceptive design.7 The label tells you everything. A cancellation flow is supposed to be a door. Amazon's was named after a war that took ten years to end.
“Amazon used manipulative, coercive, or deceptive user-interface designs known as 'dark patterns' to trick consumers into enrolling in automatically-renewing Prime subscriptions.”7
This is what reframes the famous retention number. A ~93% one-year renewal rate6 reads as adoration only if every member chose to be there and could leave at will. The settlement establishes that some couldn't, easily — and some never meant to join. The wheel was spinning partly because the exit was barricaded. That doesn't erase the genuine loyalty; it just means the headline figure is an alloy of devotion and friction, and Amazon spent fifteen years letting the world assume it was pure metal.
Isn't the flywheel still real — and isn't that all that matters?
The strongest counter is that the lock-in obviously works on the merits, friction or not. The 2× spend gap is too large and too durable to be an artifact of a confusing cancel button5, and subscription-services revenue — the line that includes Prime fees — reached about $44.4 billion in 2024, up from $35.2 billion two years earlier, growing every quarter.4 You don't manufacture a number that size with dark patterns; people genuinely value the shipping, the catalog, the video. All true. But notice what the FTC case actually changed: not whether the flywheel exists, but what kind of asset it is. Engineered retention is borrowed retention. It books like loyalty and pays like loyalty right up until a regulator reprices it — and $2.5 billion is the first installment of that repricing, not necessarily the last. A moat built partly on friction carries a tail Amazon had quietly kept off its own balance sheet.
Every great subscription business runs on a flywheel, and the most dangerous thing about a flywheel is that retention looks identical whether it's earned or engineered. A 93% renewal rate reports the same in a deck whether people stayed because they love you or because leaving was a maze. So pressure-test your own number: if you made cancellation a single, instant click tomorrow, how much of the base would walk? The gap between what you'd keep and what you keep now is your coerced loyalty — and it's the slice a regulator, a journalist, or a competitor with a one-click 'switch to us' button can take from you overnight. Earned stickiness is an asset. Trapped stickiness is a liability wearing the same suit.
Prime is one of the best lock-ins ever built, and it deserves the admiration — for the sunk-cost reframe, the relentless thickening of the bundle, the wheel that turns faster the more you spend. But the cleanest version of the story, the one where 200 million people simply chose Amazon and never looked back3, is the version the FTC just struck through. The real lesson isn't that Prime's loyalty was fake. It's that the most valuable thing a membership owns — the certainty that members are there by choice — is the one thing Amazon couldn't fully claim, and paid $2.5 billion to find out.
Switching-Cost Ledger
A worksheet that prices the exit. It itemizes every cost a customer eats to switch away — the contract penalties, the re-training, the data migration, the muscle memory — so you can see whether lock-in is real or just inertia waiting to break. Blank to audit your own stickiness; filled as the worked example tallying the switching costs the story's customers face.
The worked example unlocks with a subscription. See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1Amazon Prime was announced February 2, 2005, and memberships became available for purchase on April 17, 2005, at $79/year for unlimited 2-day shipping on eligible items.Envision Horizons, When Did Amazon Prime Start? ↗ · 2025-11-12
- 2Prime Video (streaming) was not a launch feature: Amazon Unbox launched September 7, 2006, and streaming access was added to Prime memberships in February 2011.Encyclopedia MDPI, Prime Video ↗ · 2022-11-29
- 3Jeff Bezos disclosed in Amazon's April 2021 shareholder letter (filed as SEC Form 8-K) that Amazon had 'more than 200 million Prime members worldwide'—the last official primary-source member count Amazon has released.
- 4Amazon's subscription services revenue (which includes Prime membership fees, Audible, Kindle Unlimited, and other subscriptions) reached $44.374 billion in 2024, up from $35.22 billion in 2022; the line has grown every quarter since Amazon began reporting it, averaging ~17.6% YoY growth per quarter.
- 5CIRP (Consumer Intelligence Research Partners) estimates that in 2024, U.S. Prime members spent $1,170/year on Amazon vs. $570/year for non-Prime customers—a ~2× ratio that has been consistent for at least five years, though the absolute figures have declined from the widely-cited 2018–2019 peak of $1,400 vs. $600.
- 6Amazon's one-year Prime retention rate is approximately 93%, rising to ~98% after two years, per CIRP surveys—though CIRP also found trial-to-paid conversion declined from over 75% at its peak to under 65% by 2019.
- 7The FTC filed suit in 2023 alleging Amazon used 'dark patterns' to enroll consumers in Prime without consent and deliberately built a complex cancellation process ('Iliad') to prevent cancellations; Amazon settled in September 2025 for $2.5B ($1B civil penalty + $1.5B customer refunds), with internal documents showing employees knowingly discussed deceptive design.
- 8Amazon's US annual Prime price history: $79 (2005 launch) → $99 (March 2014) → $119 (April 2018) → $139 (February 2022); the February 2022 increase was Amazon's first in nearly four years and was attributed to higher labor and shipping costs.