The One-Off Dinner Apps: What They're Actually Good For

Timeleft has seated more than 3 million people at dinners with strangers since 2020. Buffet, which launched in Los Angeles in 2024, matches one person with another person and a place to meet, based on five interests and a short personality questionnaire. Meetup, the oldest of the three, has been doing some version of “organise strangers around a shared interest” since 2002 and still counts roughly 60 million members.

All three exist because something real broke. The places where people used to bump into each other without trying — churches, union halls, diners, senior centres — have been closing for years, and the closures have hit poorer and more rural areas hardest. Research on third-place availability across the United States found losses in every category tracked between 2019 and 2021, concentrated in places that already had the least. If you don’t have a corner bar or a standing card game, an app that hands you a stranger and a table is doing a job that used to happen by accident.

The question worth asking isn’t whether these apps work. It’s what they’re actually for, and what they aren’t.

What one dinner can and can’t do

Here’s the honest limitation: a single dinner with a stranger, however pleasant, is not friendship. It’s the raw material for friendship, if you’re willing to do something with it.

Research on college students returning to social life after the pandemic found that what predicted wellbeing wasn’t the number of people someone met — it was the quality and regularity of the connections they kept. That tracks with the broader pattern in this field: repetition beats novelty, almost every time. Seeing the same handful of people on a schedule does more for a person than meeting twenty new people once each.

None of the one-off matching apps are built around repetition by default. Timeleft seats you with a new group of strangers, matched by personality, once a week if you keep booking — but it’s a new group each time, with no messaging beforehand and no built-in mechanism for turning Tuesday’s dinner into a standing Tuesday dinner. Buffet’s original model is even more singular: one match, one person, one venue, and then it’s on you to make it happen again. Meetup is the outlier here, because its actual structure is the recurring group — a running club that meets every Saturday, a board game night that happens monthly — rather than a one-time introduction. If the goal is the kind of contact that actually protects against loneliness, Meetup’s format is closer to what the research says matters. The trade-off is that Meetup groups vary wildly in quality, some are inactive, and walking into an established group as a stranger can feel more exposing than a blind date designed for exactly that.

The real friction isn’t finding people

The quiet failure mode of all three apps is the same, and it isn’t about matching algorithms. It’s what happens after the good dinner.

You have a genuinely nice evening. You exchange numbers, or you don’t, because the app doesn’t especially encourage it. Nothing happens for three weeks, because everyone is busy and nobody wants to be the person who suggests something and gets a lukewarm answer. This is the exact mechanism the friendship literature keeps flagging: it isn’t that people don’t want connection, it’s that initiating feels like imposing, and that ambiguity kills more potential friendships than any lack of chemistry does. Robert Putnam’s account of declining civic participation in Bowling Alone wasn’t really about people disliking each other more than they used to. It was about the erosion of the structures — regular meetings, standing commitments, recurring venues — that used to make follow-through automatic instead of a decision each person had to make while tired.

A survey from the American Enterprise Institute’s Survey Center on American Life found that 12% of Americans report having no close friends at all, up from 3% in 1990, and the collapse has been sharpest among men: 15% report zero close friends, a fivefold increase since 1990. Those are not people who lack the opportunity to meet someone new. Apps solve the opportunity problem well. They do almost nothing about the follow-through problem, which is the harder one.

What actually makes a dinner app worth the download

Used well, a matching app is a low-stakes way to practise showing up, and a way to meet someone you’d genuinely want to see again — which is a real and useful thing, just not the whole thing.

A few practical notes, from what these tools are actually built to do:

Book the same slot again on purpose. If Timeleft or a Meetup group works, treat the good dinner as an audition for a standing arrangement, not an event in itself. Ask, before the night ends, whether anyone wants to do a monthly version of this. The app got you in the room. It won’t turn that into a rhythm for you.

Treat a one-to-one match differently from a group. Buffet’s original format pairs exactly two people. That’s higher variance — one bad match and the evening is just awkward — but it’s also the format most likely to produce something you’d call a friendship, because you actually talked to one person for two hours instead of splitting attention across five. If the personality-questionnaire matching feels gimmicky, that’s a fair complaint; it hasn’t been independently tested for whether it predicts good matches any better than random pairing.

Expect Meetup to require more of you upfront. There’s no algorithm doing the social work. You have to pick a group, show up as the unknown person in an established room, and go back enough times that you’re not the unknown person anymore. That’s a bigger ask on night one and a better payoff by month three, which is the trade-off recurring in-person groups have always made.

None of this replaces what third places did. A library programme, a running club with a fixed weekly time, a regular seat at the same coffee shop — these cost nothing, don’t require a profile, and produce the same repetition effect these apps are chasing. The apps are worth trying because the free version of a third place is often gone from where you live. They’re not worth trying instead of the free version, if the free version still exists near you.

The dinner is not the point. What happens in the three weeks after it is.