Enshittification Summary and key ideas

Enshittification, by Cory Doctorow.

A platform connects two sides of a market: people using a service and businesses that want to reach or serve them. An intermediary can make those connections easier and create real value. The author says he works with a publisher because it performs useful tasks well and on fair terms, letting him focus on writing. The risk begins when an intermediary gains enough control to act as a gatekeeper, taking over the relationship between the people it connects.

The author calls the resulting decline enshittification. Its central pattern has three stages: a platform first offers value to users, then uses its users to attract business customers, and eventually takes more value from both groups for itself. The same service can therefore begin as something people like and later become something they feel stuck with.

Facebook illustrates how the first stage can build that dependence. When it expanded beyond American college users in 2006, people could import messages from MySpace and reply to them through Facebook. That bridge made trying the new service easier. Facebook also offered a feed of posts from people users chose to follow, rather than a stream shaped around what businesses paid to show. The author says Facebook once was fun, useful, and valuable to its users.

That early usefulness helped draw people in, and each new person made the network more valuable to others who wanted to connect with them. Facebook’s users supplied much of what made it compelling: themselves, their relationships, and the connections between them. In this sense, the platform could spend investor-funded surplus on an experience users valued. The benefit was not just an appealing feed. It was having friends and familiar contacts gathered in one place.

Once Facebook had a critical mass of users, the company could offer advertisers and publishers access to them. It pitched advertisers on targeting based on user data and on fighting ad fraud. Separately, it offered publishers a free traffic funnel by pushing their excerpts into users’ feeds to bring readers to their sites. Facebook began clawing back surplus from users and directing it to these business customers, reducing the value users received.

The businesses, too, became dependent on Facebook. Advertisers wanted its reach, and publishers wanted the traffic it had promised. In the later stage, the author says Facebook raised ad prices while targeting became less accurate and extensive ad fraud continued. He reports that Procter & Gamble eliminated its $200 million annual programmatic-advertising budget in 2018 without a decline in sales. In this discussion of targeting and ad fraud, that example challenges claims about programmatic advertising’s value, though it does not identify the budget as Facebook spending. It does not prove that advertising intermediaries never help businesses; it shows why the value of such spending could be questioned.

Publishers faced a similar shift. Facebook demanded longer excerpts, reduced the reach of shorter posts, and pushed publishers to post full articles on the platform. It then suppressed links sending readers elsewhere and even some full articles. Publishers had to pay to boost their posts in order to have them shown to people who had explicitly asked to see them, and became more like suppliers of content, relying on Facebook’s marketplace to earn from it. At the same time, users saw less from people they had chosen to follow, as advertising and boosted content took up more of the feed. In the author’s account, Facebook kept the value needed to keep users and business customers attached, while more of the remaining surplus went to shareholders and executives.

The network effects that make this arrangement profitable also help explain why users may stay. Leaving can mean losing access to friends who remain. Moving together is difficult because people have different contacts and obligations on the service; some may rely on Facebook groups for family logistics, community support, or customers. A service can therefore become less useful while still being hard to abandon. The cost is not only learning a new tool. It is risking a break in relationships that have not moved with you.

That dependence is not permanent. The author describes Facebook’s position as brittle: if a shock leads enough people to leave, those who stayed for their friends may have less reason to remain once those friends are gone. Departures can then accelerate. The platform’s early usefulness creates a network worth joining; that network can make exit costly, but it can also make a loss of users spread. The value turn rests on dependence, and dependence can weaken when the social ties holding it together begin to move.