The Friendship Apps Are Solving the Wrong Problem

Here is the assumption baked into almost every app that promises to fix loneliness: the problem is that you don’t know enough people. Widen the net, meet someone new, repeat. Swipe, match, dinner, done.

It’s a reasonable-sounding theory. It is also mostly wrong.

The evidence on how friendships actually form points somewhere else entirely. What predicts a close friendship isn’t the number of new people you meet. It’s whether you see the same person, in the same low-stakes context, over and over, without having to plan it each time. Robert Putnam documented decades ago that the informal, repeated contact — the bowling league, the church group, the regular’s seat at the bar — was the actual machinery of social life, not the occasional big event. Take that machinery away, as third places have largely been taken away, and you don’t get fewer friendships because people meet fewer strangers. You get fewer friendships because nobody sees anybody twice.

This matters for how you should actually evaluate the apps in this space, because most of them are optimised for the wrong metric. They measure themselves on how good the first meeting is. Almost none of them measure, or even try to build for, the fifth one.

What “no close friends” actually looks like

A 2021 survey from the Survey Center on American Life found that 12% of Americans report having no close friends at all, up from just 3% in 1990. Among men, it’s 15% — a fivefold increase in three decades. These aren’t people who never meet anyone. Most of them go to work, see family, are on apps. What collapsed wasn’t the opportunity to meet people. It was the structure that used to turn meetings into relationships automatically, without anyone having to decide to make it happen.

Harvard’s Making Caring Common project found something consistent with this in 2021: about half of lonely young adults said no one had taken more than a few minutes recently to ask how they were doing in a way that felt genuine. That’s not a meeting-people problem. That’s a nobody-followed-up problem.

Judged against that, here’s how the tools actually stack up

Meetup is the oldest tool in this category and, oddly, the one best aligned with what the research says works — if you use it right. Its actual design is recurring groups: a running club, a board game night, a book club that meets every third Thursday. Reported at around 60 million members, it’s free at the basic tier. The catch is that it depends entirely on which group you pick. Join a one-off event and you’ve just had a pleasant evening with strangers you’ll likely never see again. Join a group that meets weekly and show up four times running, and you’ve built the actual condition friendship needs. The app doesn’t force the second choice on you. Most people default to the first.

Timeleft organises weekly dinners that seat small groups of strangers matched by a personality questionnaire, with no messaging beforehand — you just turn up. It’s had real reach, seating more than 3 million guests globally. But by design, it’s built for novelty: a new table, a new set of strangers, most weeks. That’s a genuinely good night out, and TechCrunch’s reporting and Timeleft’s own model both frame it that way rather than as a route to a standing friendship. If you want it to function as one, you have to do the unglamorous thing the app doesn’t ask of you: notice who you got on with, and ask them to get dinner again — outside the app, on a normal Tuesday, with no matching algorithm involved.

Buffet, which launched in Los Angeles earlier this year, works differently again. TechCrunch’s coverage in April described it as pairing two people using five stated interests and a short personality questionnaire, then suggesting a venue for them to meet. It’s a one-to-one introduction tool, closer in spirit to a blind date than a group hangout. That makes it good at solving the actual first bottleneck — not knowing who to ask — but it has nothing built in for what happens after the first coffee. Whether that meeting becomes a friendship is entirely up to what the two of you do next, with no structure from the app to make the second meeting easier than the first.

None of this makes any of these apps bad. It means they’re solving introduction, which is a real and legitimate problem, but a different one from repetition, which is the harder and more important one.

The actual fix is duller than an app

If repetition is what does the work, the most useful thing any of these tools can do is get you into something that happens on a fixed schedule whether you feel like going or not — a run club, a library programme, a Meetup group with a standing weekly slot. The app’s job is to get you through the door once. Your job, which no app can outsource, is going back the fourth time you don’t feel like it.

That’s the least exciting advice in this whole category, and it’s also the only part with real evidence behind it. The people who report having close friends aren’t the ones who met the most new people. They’re the ones who kept showing up to see the same ones.