The Recurring Lunch: What Actually Works When You Work From Home

Most remote workers eat lunch at the desk they were already sitting at, off a plate balanced on the keyboard tray, while reading something that has nothing to do with lunch. This is not a failure of willpower. It is what happens when the thing that used to force you out of the building — a colleague poking their head round the door, a shared kitchen, a walk to the sandwich place with someone else who also needed a break — simply isn’t there anymore.

Gallup’s 2024 workplace data put a number on the gap: fully remote employees report loneliness at 25%, hybrid workers at 21%, and people who work on-site at 16%. That’s not a small difference. A study in Science this year went further, estimating that the shift to remote work explains roughly a third of the rise in isolation and mental distress recorded between 2011–2019 and 2022–2024. The same study found something specific worth sitting with: remote workers who live alone saw their number of solitary days rise roughly ten times faster than remote workers who live with someone else. Working from home doesn’t create loneliness evenly. It concentrates it on people who were already at risk of an empty day.

So the habit worth examining properly isn’t “socialise more.” It’s smaller and more specific than that: getting out of the house, at a fixed time, with the same person, on a regular basis, for something as unglamorous as lunch.

Why the habit is harder than it sounds

The obstacle isn’t finding someone to eat with once. Most people can manage that. The obstacle is that “let’s get lunch sometime” almost never survives contact with two full calendars. Someone has to remember, someone has to propose a date, someone has to reply, and if either person is tired that week the whole thing quietly dies. Arranging things is real effort, and it falls apart exactly when you’re too depleted to notice it’s falling apart.

This is why a fixed slot beats a floating intention. A recurring lunch at 12:30 on Thursdays removes the negotiation. You’re not deciding whether to see someone this week — you already decided that, weeks ago, when you set it up. The decision has already been made by a version of you with more energy than the one showing up on any given Thursday. That’s the whole trick, and it’s also why the tools built to solve “meeting people” often don’t solve this particular problem, because they’re optimised for the first meeting, not the fortieth.

What the apps are actually good at

Meetup is the oldest and most literal fit for the habit of repetition. It’s built around recurring groups rather than one-off events, and a group that meets weekly to run, knit, or play chess is, structurally, exactly the kind of thing that produces the same faces on the same day. Its weakness is that attendance at any given session is voluntary and variable — the group persists, but you can’t guarantee the same three people you liked last week will be there this week. You get a recurring occasion, not necessarily a recurring companion.

Timeleft runs weekly dinners that match strangers by personality, with no messaging beforehand. It’s a genuinely good route into a room of people you wouldn’t otherwise meet, and it removes almost all of the arranging effort — you show up, the table exists. But it is explicitly designed against repetition: you’re seated with different strangers most weeks, and there’s no mechanism for saying “actually, put me back with that group.” It solves for novelty, deliberately, which is a different problem from the one this article is about.

Buffet now works as a set of curated experiences posted weekly, where you claim a seat, a table locks in once three people have joined, and groups cap at six; there’s also a one-to-one mode. Like Timeleft, it’s good at lowering the cost of the first lunch or dinner with a stranger — the venue and headcount are handled for you. But unless you and someone from that table actively decide to swap details and set up something recurring yourselves, the app’s job ends when the meal does. It hands you a seed. It doesn’t water it.

None of this is a criticism of these products doing something wrong. They’re built to solve the harder-seeming problem — getting a stranger into a room with you at all — and they do that reasonably well. The quieter problem, the one Oldenburg wrote about back in 1989 when he described the “third place” as the ordinary, repeated backdrop of informal social life, is that connection used to happen by default, through the same coffee shop, the same regulars, the same shift. Nobody had to plan the second lunch. It just recurred, because the structure recurred. Apps that match you with someone new each week are, in a sense, manufacturing first lunches indefinitely. That’s useful. It is not the same as building a third place of your own.

The version that actually holds up

The habit that survives is the boring one: you meet someone — through an app, through work, through a class — and then you personally move the relationship onto a fixed slot that doesn’t depend on an app matching algorithm ever again. Second Thursday of the month becomes every Thursday. The venue gets simplified, usually to whatever’s closest, because novelty was never the point.

This matters more for remote workers specifically because the research on remote isolation keeps turning up the same detail: it’s not that remote workers dislike people, it’s that the ordinary mechanisms that used to put people in the same room at the same time have gone quiet. A 2025 study in the Journal of Affective Disorders found a measurable association between remote work arrangements and loneliness in a nationally representative U.S. sample, and separate research on healthcare workers made a useful distinction — workplace isolation and loneliness are not the same thing, and perceived social support changes how much one turns into the other. A recurring lunch is, in effect, a homemade dose of that support. It doesn’t need to be interesting. It needs to happen on schedule.

Cigna’s 2025 survey found 52% of workers report feeling lonely, a number too large to be explained by anything unusual about the people asked. This is not a niche problem being solved by a niche fix. It’s ordinary, and the fix is ordinary too.

What to actually do this week

Pick one person — a former colleague, a neighbour, someone from a Meetup group or a Buffet table you liked once — and propose a specific day, a specific time, and a specific place, monthly at minimum, weekly if you can manage it. Put a name in the calendar entry, not “lunch tbd.” Say out loud, or in the message, that you’d like it to be a standing thing rather than a one-off, because the ask itself is what turns a nice afternoon into a habit.

Expect the first few to feel unnecessary — you’ll be busy, they’ll be busy, it will seem easier to cancel. That week is the test. The lunches that matter are the ones that happen anyway, because the slot already existed and neither of you had to decide, again, whether the other person was worth the effort of asking.