Enshittification: The Three-Stage Way Every Platform Rots

You already know the feeling, even if you have never heard the word. The app that used to be brilliant is now infuriating. The search box that once found what you wanted now buries it under sponsored junk. The feed that showed you your friends now shows you strangers you did not ask for. The word for this slow rot is enshittification, coined by the writer Cory Doctorow, and it names something most of us have lived through without a vocabulary for it: the specific, patterned way that almost every digital platform gets worse over time. Not by accident, and not because the engineers got lazy. By design, in three predictable stages, for reasons that are structural rather than personal.

I find the word useful precisely because it is crude. It refuses to let anyone dress up the process in the polite language of “monetisation” or “evolving the product.” It says plainly what is happening: the thing is being made worse on purpose, and someone is profiting from the worsening. Once you can see the pattern, you cannot unsee it. It is everywhere.

The three stages of decay

Doctorow’s description is elegant because it is so simple. A platform sits between two groups — the people who use it and the businesses that want to reach those people — and it rots in three moves.

First, it is good to its users. This is the seduction phase, and it is often genuinely wonderful. A new search engine returns clean, relevant results with barely any advertising. A new social network shows you a chronological feed of exactly the people you chose to follow. A new marketplace undercuts everyone on price and ships in a day. A new ride-hailing app is cheap for you and pays the driver well. The platform runs at a loss, subsidised by investors, because its only goal right now is to get you in and keep you. And it works. You come, you stay, you tell your friends.

Second, it abuses its users to please its business customers. Once you are locked in — and I will come to why you are locked in — the platform turns the dials. The search results now lead with ads dressed up to look like results. The chronological feed becomes an algorithmic one, throttling the reach of the accounts you followed so that businesses must pay to reach the audience they already built. The marketplace starts charging sellers for placement and stuffs its own house brands to the top. Your experience degrades, but only just enough that you grumble and stay. The value that was flowing to you now flows to the advertisers, sellers, and merchants who pay the platform.

Third, it abuses those business customers too, and claws all the value back for itself. The sellers who came for cheap reach now pay ever-rising fees for ever-worse placement. The advertisers pay more for less. The creators who built audiences find their reach strangled unless they pay, and paying no longer guarantees much. The drivers who were paid well are now squeezed on every ride. By this stage the platform has extracted the value from both sides — the users and the businesses — and funnelled it upward, to its owners and shareholders. What is left is a hollow thing that everybody complains about and nobody can leave.

The word refuses to let anyone dress the process up in the polite language of “monetisation.” It says plainly what is happening: the thing is being made worse on purpose, and someone is profiting from the worsening.

Why doesn’t competition stop it?

Here is the question that matters. In a healthy market, a business that degrades its product loses customers to a rival that does not. The threat of losing you is what keeps a company honest. So why can a platform get worse and worse and still keep you?

The answer is that platforms are engineered to remove the exit. Two forces do the work: network effects and switching costs.

Network effects mean the platform is valuable to you mainly because everyone else is on it. You do not stay on a social network because its software is good; you stay because your friends, your family, your customers, and your professional contacts are there. A better-designed rival with nobody on it is useless to you, however good the design. The value is not in the product — it is in the people, and the people are hostage to each other. Everyone would leave if everyone left, but nobody can be the first to go.

Switching costs are everything you would lose or have to redo if you left. The years of photos. The followers you spent a decade building. The reviews and the seller rating you cannot take with you. The playlists, the purchase history, the saved addresses, the muscle memory. Platforms deliberately make it hard to export your data or take your audience elsewhere, because every gram of friction on the way out is a gram of freedom to mistreat you once you are in.

Together these produce lock-in: the state of being unable to leave even when you want to. And lock-in is the whole game. It removes the competitive pressure that kept the platform honest, and once that pressure is gone, degrading the product is no longer a risk — it is a strategy. This is the same logic I traced in how technology gets captured: a tool that begins by serving us is steadily turned to serve its owners, and the turning is possible precisely because we can no longer walk away.

The pattern in the wild

You do not need me to name companies for the examples to land. Think of the search engine you used ten years ago and the one you use now — the same box, but the answer you want is further down the page, behind a wall of ads and auto-generated filler. Think of the social feed that once showed your friends in order and now shows you an algorithmic slurry optimised to keep you scrolling and to sell that scroll to advertisers. Think of the online marketplace where the top results are increasingly whatever the seller paid to put there, not what is best or cheapest. Think of the ride-hailing app that was cheap for you and generous to drivers, and is now expensive for you and mean to drivers, with the difference kept by neither of you.

In every case the arc is the same: good, then mediocre-but-sticky, then extractive. And in every case the mechanism is the same: get everyone in, raise the walls, then turn the dials. The details differ — ads here, fees there, throttled reach somewhere else — but the shape does not.

Enshittification is the attention economy’s endgame

It helps to see this word alongside a few others, because enshittification is not a standalone insult — it is one name for a larger machine. When a platform is free to you, you are not the customer; your attention and your data are the product being sold. That arrangement is the engine of the attention economy, and enshittification is simply what happens to it over time: the moment your attention is the thing being sold, the platform’s incentive is to harvest more of it and return less to you. The famous line about free products — that if you are not paying, you are the product — is really a warning about the second and third stages of this decay, and I have written more plainly about that bargain in the real cost of free apps.

Zoom out further and enshittification is the customer-facing symptom of a deeper shift some thinkers call technofeudalism — a world where a handful of platform owners control the digital land the rest of us must live and trade on, and collect rent for the privilege. Enshittification is what the rent-collection feels like from inside. The platform is the fief, the users are the peasants, the businesses are the tenant farmers, and the owner takes an ever-larger cut of everything grown on the land, because where else is anyone going to farm?

Get everyone in, raise the walls, then turn the dials. This is the capture pattern rendered in software: not a betrayal by bad people, but the predictable output of a system built to reward it.

Why calling it a “pattern” matters

The most important thing about enshittification is that it is a pattern, not a personality flaw. It is tempting to explain the rot by pointing at a villain — a greedy founder, a cynical executive, a board that lost its way. Sometimes those people exist. But the pattern would happen without them, because the pattern is what the incentives produce. Give any company a locked-in audience, remove the threat of them leaving, and add shareholders who expect the numbers to go up every quarter forever, and enshittification is not a temptation the company might resist. It is the path of least resistance. It is what the machine does when you leave it running.

That is why I think of it as the capture pattern rendered in software. The same dynamic that let landlords enclose common fields, or let a company town pay its workers in scrip they could only spend at the company store, is at work when a platform locks you in and then makes you worse off. The technology is new; the move is ancient. Concentrate control over something people need, eliminate their alternatives, and then extract. Software just lets it happen faster, at planetary scale, and with a friendlier interface.

Can a platform escape it?

Not on its own, and this is the uncomfortable part. As long as the three conditions hold — lock-in, no meaningful competition, and relentless pressure to extract more — the pull toward enshittification is constant. A well-meaning team can slow it, but they are swimming against the current of their own business model, and they usually lose. The people who founded the good version rarely run the extractive one; by then it belongs to whoever bought it, and to the shareholders whose only relationship to you is the value they can pull out.

What actually breaks the pattern is anything that restores the exit. The right to take your data and your audience elsewhere, so switching costs fall. Interoperability, so a rival does not have to recreate the whole network to compete. Real competition, so degrading the product once again means losing you. Ownership structures — cooperatives, public utilities, non-profits — that are not chained to the quarterly demand for more extraction. None of these is a fantasy; each already exists somewhere. What they have in common is that they attack the lock-in, because the lock-in is the disease and the enshittification is only the fever.

So the next time an app you love starts to turn — the ads creeping in, the good features moved behind a paywall, the reach you built quietly throttled — you will know you are not imagining it, and you are not being ungrateful. You are watching a three-stage process run its course. The value that used to flow to you is being redirected, on purpose, because the door behind you has been quietly locked. Naming it is the first small act of refusing it.