The Washington PostDemocracy Dies in Darkness

Opinion Twitter might be replaced, but not by Mastodon or other imitators


By now, every social media network’s primary problem is all the other social media networks. Each platform becomes more valuable with each additional user, because it means more people to connect with. But at this point, fledgling platforms have basically two choices to grow their networks: Persuade users to give up some other app, or persuade us to give up sleep.

This is why Elon Musk’s acquisition of Twitter turned out to be a stroke of luck for Mastodon, a platform that might be described as an open-source, decentralized version of Twitter. In November, people who were disgusted with Musk’s new moderation policies — or just worried that Twitter would stop working after a 75 percent reduction in employee head count — migrated to Mastodon, driving its user total from 300,000 in October to 2.5 million.

This was a spectacular growth rate until it stopped. By early this month, active users had fallen by more than half a million, the Guardian reported. Possibly, this was simply a temporary hiccup, but I doubt it, because the nature of the attention economy makes it hard to create a major social media network that is like an existing network, only a little bit different.

In general, if you are providing a service that is like something else on the market but tweaked in some way, you run the risk that you have changed something important, and your new version will be less popular, or less profitable, than the one you were trying to replace.

Follow Megan McArdle's opinionsFollow

Mastodon is not looking for a big payday; it’s overseen by a nonprofit. But it is (presumably) looking for users. And while the decentralized, open-source model has some real advantages — for instance, it lets communities have more control over their own space, rather than offering a one-size-fits-all corporate solution — it often solves problems that most users don’t care about, while creating ones that bother them quite a bit.

Take bitcoin. It elegantly solves a problem — permissionless payments — that most people in wealthy countries don’t care about. It does so at the cost of speed and some kinds of security; transactions can take more than 10 minutes to clear, and once made, they can’t be reversed, even if they were fraudulent.

Or take Linux, an open-source computer operating system that enthusiasts once hoped would compete with Apple and Microsoft on the desktop, but which has largely remained the province of IT professionals and a few hobbyists. It turns out users didn’t want maximal control over their desktop environment, or even minimal cost; they wanted to pay a corporation to ensure that they never had to think about their operating system.

There are exceptions, including Android, the open-source mobile platform. But this is not a homespun peer-to-peer project; it’s overseen by Google, which has boring capitalist reasons for wanting to deny Apple’s iPhone a monopoly over mobile.

Mastodon is not a corporate project, and it shows, in good ways and bad. Everything has a volunteer, hey-kids-lets-put-on-a-show feel — which appeals to a veteran of the old blogosphere like me. But it also means a clunky interface, and the need to figure out which server to join, and then hunt down the users you want to follow. Also, Mastodon’s lack of a central authority raises questions: If a bunch of server moderators decide to ban you because of a smear campaign against you, or mistaken identity, or just a temporary viral furor that eventually dies down, how do you get unbanned? Do you have to go to hundreds or thousands of moderators and ask them one by one?

Centralization has drawbacks, but also considerable benefits, and it’s uncertain that a decentralized network can provide the features and solve the problems that most users care the most about as well as a centralized service with lots of highly paid engineers. That’s one version of the core problem for any Twitter substitute: Can it solve enough of Twitter’s problems without introducing new ones that most users find worse?

And would it be able to do all this in competition with a much larger network that will be difficult to replicate? It would need to reconstruct not only the lists of people one follows, which might be done manually, but also the large followings that power-users have amassed. Unless everyone jumps at once, these will be lost.

While network effects certainly don’t make a company immortal (just ask Myspace), they do present a hurdle for any new service. And it’s a bigger hurdle than ever because, as I said, our time is already so filled, and it’s hard to add a new service without letting go of another.

So I suspect that when Twitter is eventually displaced, as it undoubtedly will be, its successor will be not Twitter-but-different but rather something not much like the platform at all — a service that can grow its network not by solving Twitter’s problems, but by doing what the platform did originally: fulfilling wants we didn’t realize we had.