The Enshittification Cycle
Diagnose platform decay as a three-phase transfer of value from users to shareholders
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 93%
Enshittification is both an observation and a causal theory. Observationally, a platform runs three phases: first it is good to end users while quietly building lock-in; second it degrades the user experience, relying on lock-in to prevent exit, and redirects the surplus to business customers; third it locks in those business customers too and claws back their surplus, leaving only the minimum residue needed to keep both sides attached. The causal half asks the inverse question: not why firms are greedy now, but why they were ever restrained. The answer is four disciplines — competitive pressure, regulators, interoperability (rivals building plugins and bridges that let users escape a rent-extraction scheme), and skilled workers who fight for users. Dismantle those constraints and you create a policy environment that rewards the worst actors, the worst factions inside firms, and the worst instincts of CEOs.
Origin
Developed by Cory Doctorow across a quarter-century at the Electronic Frontier Foundation and formalised in his book on the subject, as a framing device to make abstract digital-rights harms feel urgent before people's houses are already on fire.
Core principles
- 01Lock-in is the precondition for abuse, not the reward for quality.
- 02Firms always wanted to charge more and deliver less; what changed is what stopped them.
- 03Four disciplines constrain a firm: competition, regulation, interoperability, and labour.
- 04Decay is a policy outcome, not a personality defect in founders.
- 05The end state is a platform that keeps only the minimum residue needed to hold everyone hostage.
How to run it
- 1
Map phase one: good to users, building the trap
Identify the period where the platform subsidised or delighted end users, and find what it was accumulating alongside that goodwill — social graph, purchase history, files, reputation, or format-locked media.
Pro tip The asset that creates lock-in is rarely the product; it is the thing you cannot take with you when you leave.
- 2
Locate the switching cost
Name precisely what a user forfeits by leaving: their audience, their library, their history, or their livelihood. If you cannot name it, the platform is not yet enshittifying, it is just competing.
Watch out Switching costs are often legal rather than technical — a format is readable, but reading it is unlawful.
- 3
Track the surplus moving to business customers
Watch for the moment user experience visibly degrades while sellers, advertisers, or creators get better reach and terms. The degradation is funded by the lock-in established in phase one.
Pro tip Recommendation feeds that stop showing what you subscribed to are the clearest tell.
- 4
Track the clawback from business customers
Once sellers depend on the platform for their livelihood, fees, mandatory advertising, and pay-for-placement schemes rise. Measure fees as a share of gross revenue, not as a headline rate.
Pro tip Compare 'optional' fees against what actually happens to visibility if you decline them.
Watch out Firms present extraction as optional services; treat any fee you cannot realistically refuse as compulsory.
- 5
Audit the four disciplines
Score competition, regulation, interoperability, and labour power as intact, weakened, or gone. The count tells you how much further the platform can decay before something pushes back.
Pro tip Interoperability is the fastest-acting discipline because it needs no election and no lawsuit.
- 6
Forecast the next move from the weakest discipline
Extraction always advances along the path of least resistance. Where no discipline binds, expect the next fee, the next ad load, or the next data grab.
Watch out Do not attribute the next move to a named executive; the same people ran these firms during the good years.
In the wild
Doctorow traces Amazon's growth to rent rather than products. Its advertising business — payment for placement in a search result — grew from roughly thirty-two billion dollars when he wrote the book to over fifty billion. The consequence is that the top search result is on average twenty-nine percent more expensive than the actual best match, with genuinely good results pushed to around position seventeen. Sellers who spend more on placement have less surplus left for product quality. On the merchant side, junk fees consumed forty-five to fifty-one percent of merchant revenue at the time of writing, and fifty to sixty percent by the time of the interview.
→ Both sides of the marketplace are worse off while platform revenue from rent keeps compounding.
After thirteen years of subsidised rides funded by a thirty-one billion dollar SoftBank round, Uber had frozen out competitors and reclassified drivers. It then used private data and algorithmic pricing to offer different wages to different drivers based on estimated desperation. A driver who accepts a below-historic fare sets a new personal ceiling and gets nudged lower still. Drivers self-classify as 'pickers' and 'ants'; the system's goal is to convert every picker into an ant, because an ant stays on the road regardless of fare quality.
→ Prices rose for riders and wages fell for drivers simultaneously, with the spread captured by the platform.
Common mistakes
Blaming founder greed rather than policy
The same people ran these companies during their good years, and greed was not invented mid-decade. Attributing decay to individual villainy hides the removal of the constraints that previously made restraint profitable.
Confusing losing a market with enshittification
MySpace decayed and died; the platforms Doctorow interrogates decay and stay wildly profitable. If the firm is punished by exit, the disciplines still work and the model does not apply.
Treating quality decline as a UX problem
Degradation is a deliberate reallocation of surplus, not a design failure. Fixing the interface does nothing while the lock-in that permits the reallocation stays intact.
Is it for you?
Best for
Analysts, founders, and policymakers trying to predict where a dominant platform's next value grab will land.
Not ideal for
Explaining ordinary product failure where a company simply loses to a better competitor and dies.
From the transcript
“in the first phase you have firms that are good to their end users, but find a way to lock those users in. In the…”
“You have to labor under the discipline of competition of regulators, and if you're in tech, you had other sources of discipline, like interoperability, the…”
“my thesis is that when we dismantled these sources of discipline, we created what you could call the eshridogenic policy environment”
From the episode
BONUS: How the Internet Got Worse with Cory Doctorow
Cory Doctorow