Enshittification is the process by which a two-sided platform declines in quality in a predictable order: good to its users first, then degraded in favour of its business customers, then degraded for those business customers too, until the value has been moved almost entirely to the platform itself.
The word is crude on purpose, which has caused people to underrate it. What makes it useful is not that it insults platforms — plenty of words do that — but that it names a sequence. It tells you what comes next, which a complaint cannot.
Where the word comes from
Cory Doctorow popularised the term in 2022. It travelled fast enough that the American Dialect Society named it the 2023 Word of the Year, and Australia’s Macquarie Dictionary did the same for 2024.
He did not coin it from nothing. The Oxford English Dictionary attests uses going back to 2013. What Doctorow added was the three-stage structure — turning an insult into a description of a mechanism. That is the whole reason the term outlived the news cycle that carried it.
The three stages
| Stage | Who the platform serves | What holds people in place |
|---|---|---|
| 1. Acquisition | Users. The experience is genuinely good, often subsidised by investor money. | Nothing yet — that is the point. It has to be good. |
| 2. Extraction from users | Business customers: advertisers, sellers, publishers. | Your data, your social graph, your livelihood. Leaving now costs something. |
| 3. Extraction from everyone | The platform itself, and its shareholders. | Switching costs on both sides. Everyone is stuck at once. |
The third stage is unstable, and that is the interesting part. A platform that is bad for everybody retains people only through the cost of leaving. Nothing else is holding the arrangement together.
Why it happens in that order
Because the thing that made the platform good was usually a subsidy, not a business model. Cheap capital paid for the early experience. When the subsidy ends, the value has to come from somewhere, and lock-in decides who absorbs it: the harder it is to leave, the more can be taken before anyone does.
That is why this reads as a sequence rather than a run of unrelated bad decisions. The incentive arrives at a predictable moment — when growth slows and switching costs are high — and the same incentive produces the same behaviour across companies that have nothing else in common.
What it is not
This is where the term gets misused, and the misuse is worth resisting if you want it to keep meaning anything.
A product can get worse for entirely ordinary reasons: a redesign people dislike, accumulated technical debt, a change in taste, a bad hire. None of that is enshittification. The term describes value being moved between parties in a particular order. If something is worse because someone made a mistake, that is a mistake. If it is worse because the deterioration is how the money is now made, that is the thing the word is for.
The same distinction separates it from dopamine culture, which describes what media is optimised for. Enshittification describes how a platform’s economics change over time. They frequently appear together, and they are not the same claim.
Is it reversible?
Doctorow’s own position is that the decay is not a law of nature: the constraints that used to prevent it — competition, regulation, interoperability, and workers willing to refuse — were dismantled rather than outgrown.
The individual version of that is smaller and less satisfying, but it is the part you can act on. The leverage is entirely in switching costs. Anything that lowers yours — data you can export, an audience you can contact directly, work that does not evaporate when you leave — reduces how much can be extracted from you before you go. That is not a solution to the phenomenon. It is the only part of it under your control.
Sources
- Enshittification — Wikipedia — for the definition, Doctorow’s 2022 popularisation, the American Dialect Society’s 2023 Word of the Year and Macquarie Dictionary’s 2024 selection, and the OED’s attestation of earlier uses from 2013.
Checked against the source above on 5 August 2026.
Frequently Asked Questions
What is enshittification?
Enshittification, also called platform decay, is the process by which two-sided online platforms decline in quality over time in a predictable order. A platform first offers users a genuinely good experience to attract them, then degrades that experience to favour its business customers — advertisers, sellers, publishers — and finally degrades things for those business customers too, extracting as much value as possible for shareholders. The term was popularised by the Canadian writer Cory Doctorow in 2022. Its usefulness is that it names a sequence rather than a mood: it predicts what happens next, not merely that something feels worse.
Who coined the word enshittification?
Cory Doctorow popularised it in 2022, and it spread quickly enough that the American Dialect Society selected it as its 2023 Word of the Year, with Australia's Macquarie Dictionary doing the same for 2024. Doctorow did not invent the word from nothing — the Oxford English Dictionary attests uses dating back to 2013 — but he gave it the specific three-stage meaning that made it analytically useful rather than merely rude. That distinction matters: the word existed as an insult before it existed as a description of a mechanism.
What are the three stages of enshittification?
First, the platform is good to its users, often subsidised by investor money, to build an audience that would be costly to leave. Second, once users are locked in by their data, their social graph or their livelihood, the platform degrades their experience to favour business customers who pay it — more ads, worse ranking, pay-to-reach. Third, once those business customers are also locked in, the platform degrades their experience too, taking a larger share of the value for itself. The final stage is unstable, because a platform that is bad for everyone has nothing holding anyone in place except switching costs.
Why do platforms enshittify instead of staying good?
Because the thing that made them good in the first place was usually a subsidy rather than a business model. Cheap capital paid for the early experience, and when the subsidy ends the platform must extract value from somewhere. Lock-in decides who absorbs that extraction: the harder it is to leave, the more can be taken before anyone does. That is why the process reads as a sequence rather than a series of unrelated bad decisions — the incentive appears at a predictable moment, when growth slows and switching costs are high.
Is enshittification the same as a platform just getting worse?
No, and collapsing the two is the most common misuse. A product can get worse for ordinary reasons: bad design decisions, technical debt, a redesign people dislike, or simply a shift in taste. Enshittification is narrower — it describes value being deliberately moved between parties in a specific order, from users to business customers to the platform itself. If a service is worse because someone made a mistake, that is not enshittification. If it is worse because the deterioration is how the money is now made, it is.
Can enshittification be reversed?
Doctorow's own argument is that it is not inevitable and that the constraints which used to prevent it — competition, regulation, interoperability, and the willingness of workers to refuse — were removed rather than lost to nature. The practical implication for an individual is unglamorous: the leverage sits in switching costs, so anything that lowers yours (exportable data, an audience you can contact directly, a service you can leave without losing your work) reduces how much can be extracted from you. That is a smaller claim than the theory makes, and it is the one you can act on.