Case Study: Facebook

Stage One: Good to Users

Facebook is the perfect place to start. Facebook is a service that Mark Zuckerberg started in his dorm room so that he and his creepy pals could nonconsensually rate the fuckability of their fellow Harvard undergrads.1

It was Zuck’s users who figured out how to make something great out of this inauspicious beginning. They filled up the vessel with themselves and with connections to one another, forging bonds. It was a compelling experience, so much so that many people called it addictive.

But Zuck was determined to bring Facebook back to its origins: a service that treated people as means, not ends—as something for the platform’s managers to toy with and, ultimately, abuse.

When Facebook started, it was available only to American college kids. They loved it, and investors saw potential: with all these users piling it, there would be innumerable opportunities to sell things to them. Hell, you could sell users to one another. Investors poured money into the business. In 2006, Facebook decided to spend some of its investors’ cash to expand to the general population, so it dropped the requirement for users to sign up with the .edu email addresses that only American college students, faculty, and personnel can get.

Back in 2006, Facebook made a simple and compelling pitch to those new users it was hoping to lure onto the platform:

Sure, we understand that most of you already have a social media service that you enjoy using called MySpace. But has it occurred to you that MySpace is owned by an evil, crapulent, senescent Australian billionaire named Rupert Murdoch, and he spies on you with every hour that God sends?

Come to Facebook, where we will never spy on you.2

All we ask of you is that you create a Facebook account and then “articulate your social graph” by telling us which other users you want to connect to you. Then, we will create an automated custom feed, consisting solely of the things that the people you follow have posted for consumption by their followers.3

This is a pretty good deal. You aren’t imagining it: Facebook was fun and useful and valuable, once upon a time.

That was stage one: Facebook had a surplus (its investors’ cash), and it allocated that surplus to its end users by subsidizing a feed of things that users wanted to see, rather than things that businesses would pay to show them.

To sweeten the deal, Facebook gave MySpace refugees a way to eat their cake and have it, too: a bot. Once you joined Facebook, you could give that bot your MySpace login and password, and then, several times a day, the bot would log in to MySpace on your behalf, scraping all the messages your friends there had left for you and pasting them into your Facebook inbox. You could reply to those, and the bot would push them back out to MySpace on your behalf. That way, you didn’t have to choose between Facebook’s superior user experience and the friends you left behind on MySpace.4

So users piled in, and they proceeded to lock themselves into Facebook’s platform. There are lots of ways for digital businesses to lock in their users, but with social media, they don’t even have to try.

Most online businesses enjoy high network effects. This is the economist’s term for a product or service that gets more valuable as it attracts more users. You joined Facebook because the people who were already there made it valuable to you, and once you were there, you made Facebook more valuable to the people who wanted to hang out with you.

But Facebook doesn’t just benefit from large network effects; it also relies on high switching costs.

Switching costs is another useful piece of economics jargon. Switching costs are everything you have to give up when you switch from one product or service to another.

In the case of Facebook, the switching costs of leaving Facebook include the company of everyone you hang out with there, because they’ll still be on Facebook and you won’t be.

Now, hypothetically, you can avoid this switching cost. If you can convince all the people you like on Facebook to quit at the same time as you and transition to a new social network where you can all reestablish your links, you can leave Facebook and keep your friends.

Here’s where the collective action problem comes in. That’s our third and final piece of economics jargon: the collective action problem is the incredibly difficult business of getting other people to do what you want them to do, when you want them to do it.

You experience the collective action problem on a small scale all the time. You know that group chat with half a dozen friends in it? Remember how hard it was to decide what movie or bar to go to, or which board game you should all play? Even when you all agree that you want to do something together, it’s hard to agree what to do and when to do it. That’s the collective action problem.

Adding people to the group makes the collective action problem exponentially more problematic. Once you’re established in a Facebook community with a couple hundred friends, you have to convince them all to leave at the same time as you, and go to the same place as you.

But all those friends have their own groups of people they can’t afford to leave. Maybe they’re in a support group for people with a rare disease. Maybe they use Facebook to organize their kid’s Little League carpool with the other parents. Or maybe that’s how they stay in touch with the people they left behind when they emigrated. Maybe it’s where their customers are.

To get your friends to switch away from Facebook, you not only must convince them that it’s time to go and that you’ve got the right place for them to go next—you also have to convince them to talk all those other people into leaving, too; or you have to convince them to endure the switching costs of leaving their own groups behind.

So Facebook users can’t help but take one another hostage, and Facebook grew progressively more cognizant of this fact. Once it sensed that a critical mass of its users were locked in to its platform, it was time for stage two.

Stage Two: Good to Business Customers

Facebook understood that it could make money from two groups of business customers: advertisers and publishers. There was only one problem: to make the service valuable to them, Facebook would have to reduce the value enjoyed by its users.

So Facebook started to claw back the surplus from those end users and began doling it out to advertisers and publishers. Facebook approached its advertisers and made a pitch: “Hey, do you remember when we told these rubes that we wouldn’t ever spy on them? We were lying. We spy on them from asshole to appetite. If you give us a remarkably reasonable sum of money, we will use that surveillance data to do extraordinarily precisely targeted advertising on your behalf. What’s more, we are such upright, good-natured slobs that we have filled a whole building with engineers who labor day and night to fight ad fraud. If you give us a dollar to show an ad to a specific kind of person, you can be sure that ad is going to be shown to the right person.”5

Then, Facebook approached publishers and made a different pitch: “Hey, do you remember when we told these rubes that we would only show them the things they asked to see? That was a total lie. If you post short excerpts from your own website content to your Facebook account, complete with a link back to that website, we will nonconsensually cram those excerpts into the eyeballs of users who never asked to see them. You will get a free traffic funnel that you can monetize as you see fit.”6

So the publishers and the advertisers piled in, too, and they became dependent on the users—the users who were sticking around because they were dependent on one another, the users who took one another hostage—which meant the publishers and advertisers were now held hostage, too.

Now it was time for stage three of enshittification: putting the screws to the business customers.

Stage Three: A Giant Pile of Shit

For advertisers, this took the form of rising prices and plummeting ad fidelity, along with skyrocketing ad fraud. Gradually, Facebook ramped up the price of targeting an ad to its users, but it also took less care to show ads to the users advertisers had selected.7

Meanwhile, ad fraud was going wild. Advertisers were paying billions for ads that no one ever saw. In 2018, Procter & Gamble zeroed out its $200 million annual “programmatic advertising”8 budget and saw no decline in sales.9 It seems all of those ads were either:

Being shown to random people rather than the people P&G was paying to target; orNot being shown to anyone.It wasn’t any better for publishers. Gradually, Facebook’s content recommendation algorithm started to require longer and longer excerpts in order for posts to qualify for being shoved into strangers’ feeds. The system also started downranking shorter excerpts in subscribers’ feeds, meaning that publishers had to push longer and longer excerpts onto Facebook’s platform in order to be seen at all. The longer a post was, the more substitutive it was for the whole article.

Eventually, publishers were corralled into publishing their whole articles on Facebook, and then, to add insult to injury, Facebook started to suppress posts that linked away from its site, on the grounds that such links might be “malicious.” Worse, even whole articles with no links were often suppressed, even for followers, and publishers had to pay to “boost” their content in order to have it shown to the people who’d explicitly asked to see it.


Copyright © 2025 by Cory Doctorow