How to make friends as an adult (why it got so hard)

Try to name the last friend you made. Not a colleague you like. Not someone's partner you see at dinners. A friend you chose, who chose you back.

If the answer is "someone from school, a decade ago," you are in the majority, and almost nobody says so out loud.

Here is the part worth sitting with: you already know how to make friends. You did it effortlessly at seven, at twelve, at nineteen. No books, no strategies, no apps. Then somewhere in your mid-twenties the supply stopped — for nearly everyone, at nearly the same age.

When millions of people develop the same problem on the same schedule, that isn't a personality flaw. That's a system failing.

The size of the hole

Surveys suggest a large decline. In 1990 Gallup found about 3% of Americans reported no close friends at all; by 2021 the American Perspectives Survey put it near 12%, and around 15% among men. The share reporting ten or more close friends fell from roughly a third to about one in eight.

Treat the exact gap loosely — the 1990 survey was conducted by phone and the 2021 one online, so some of the difference is survey method rather than real change. But every way researchers slice it points the same direction.

The time data is firmer, and it's the number that matters. The American Time Use Survey tracks how people actually spend their hours. In 2003 the average American spent about 45 minutes a day socialising. Twenty years later it was closer to 35 — a drop of more than a fifth. For 15-to-24-year-olds it was steeper: from about an hour down to roughly 35 minutes.

And this decline was well underway by the mid-2010s. Lockdowns accelerated a curve already pointed at the floor.

Why it matters more than it sounds

In 2010 Julianne Holt-Lunstad pooled 148 studies covering more than 300,000 people, and found that people with stronger social relationships had a 50% higher likelihood of survival across the study periods — a risk magnitude comparable to quitting smoking. Her later work put social isolation's effect on early death odds at roughly 26–32%.

The caveats belong here rather than in the footnotes: these are correlations gathered across many studies, and the comparison is about the size of the risk, not the biology. Nobody is claiming loneliness gives you the diseases smoking does. They knock roughly the same amount off your odds.

The machine that used to do this for you

Since the 1950s, sociologists have pointed to three conditions wherever friendships form — a triad the sociologist Rebecca Adams is most often associated with:

Proximity. You have to physically be near the same people, repeatedly.
Repeated, unplanned interaction. You have to bump into them without arranging it.
A setting that lets your guard down. Somewhere you can be a little unimpressive.

Read that list again and notice what it describes. It's school.

Every school day put you within arm's reach of hundreds of peers, collided you with them at lockers and lunch tables with no planning required, and threw you into shared struggle that cracked everyone's guard open. School wasn't just where your friends were. School was a friendship factory — and nobody mentioned that when you graduated, you were walking off the factory floor.

It's more mechanical than you'd like

In 1950 Leon Festinger and colleagues mapped friendships in an MIT student housing complex called Westgate. Among residents assigned to their apartments, the single best predictor of who became friends wasn't personality, values or shared interests. It was the distance between front doors. People living near stairwells and mailboxes — the collision points — made the most friends in the building.

Four decades later Moreland and Beach made it stranger. They had women attend a university class a different number of times — some five sessions, some fifteen — never speaking to anyone. At the end, students rated the women who had simply shown up more often as more likeable. Without a single conversation. It's a shift, not a transformation, but it's real: your brain reads I keep seeing this person as this person is safe.

Friendship is less a meeting of minds than an accumulation of exposures.

So how many exposures?

Someone counted. Jeffrey Hall surveyed people who had recently moved to a new city and estimated the shared hours needed to climb the friendship ladder: roughly 50 hours to go from acquaintance to casual friend, around 90 to become a genuine friend, and past 200 for a close one.

Hold those numbers loosely. It's one study built on people's own retrospective reports, from modest samples, and it hasn't been independently replicated. They're the best estimate anyone has, not a law.

But now put that beside the other number. Half an hour a day of total social time, spread across everyone you know.

At that pace, 200 hours with one new person isn't a project for the autumn. It's a project for years.

At school pace you sat beside the same people thirty-plus hours a week. Two hundred hours took a term. That is the whole mystery, solved with arithmetic. Nothing about your personality changed at twenty-five. Your hours changed.

What modern life removed, one condition at a time

Proximity went first. Careers that expect relocation, suburbs that put a lawn between neighbours, hours sealed inside a car. Americans actually move less than ever — around 11% a year, a record low — but the moves that do happen cluster in your twenties, exactly when adult networks should be forming, and every move resets the exposure clock to zero.

Repetition went next. The commute replaced the walk, the drive-through replaced the counter, delivery replaced the errand. The collision points of daily life were optimised away as inefficiencies. A study of 61,000 Microsoft employees during the 2020 shift to remote work found cross-group collaboration fell by about 25% and networks became more siloed — the weak ties thinning first.

Unguarded settings went last. Participation in organised community life — leagues, service clubs, congregations, the things Robert Putnam counted — has been sliding for two generations. What replaced all of it is scheduled, high-intention socialising: the calendar invite three weeks out, dinner at seven, everyone performing their best self.

Efficient. Pleasant. And structurally incapable of manufacturing a friend, because it strips out the repetition and the unguardedness that do the actual work.

You are not bad at this

You are running social hardware built for a fire circle, inside an environment that removed the fire, fenced the circle, and sold the time back to you in half-hour portions.

For most of human history nobody "made friends" — the category barely makes sense. People lived in small, flexible camps of a few dozen, nested inside much larger regional networks. Food was gathered together, meals were shared, evenings pooled around a fire. Proximity, repetition and unguardedness weren't things you sought out. They were the default condition of being alive.

Friendship wasn't an achievement. It was weather. You were never designed to construct a social life — you were designed to inherit one.

Which means the fix isn't confidence, and it definitely isn't an app. It's rebuilding a small, deliberate version of the old environment inside this one.

The protocol

Install an anchor, not a plan. Stop scheduling friendship one meeting at a time. An anchor is anything that repeats on its own: same people, same place, same time, no weekly decision required. A run club. A weekly pickup game. A standing Sunday dinner you host badly. A class every Tuesday. A fixed volunteering shift.

The content barely matters. The structure is the entire point, because an anchor generates proximity and repetition automatically — exactly what school did. One hour a week, held for a year, is fifty hours with the same faces. You'll have crossed the casual-friend threshold without once "putting yourself out there."

Choose it with the arithmetic in mind: weekly beats monthly by a factor of four, and showing up every time beats showing up enthusiastically sometimes.

Then correct two bugs in your own perception. After a conversation with someone new, people systematically underestimate how much the other person liked them. Boothby, Cooney, Sandstrom and Clark named it the liking gap in 2018, and it has replicated repeatedly. So when you replay an interaction and cringe, treat the cringe as faulty instrumentation, not information.

Second, people predict that talking to near-strangers will be awkward and draining, then consistently report the opposite afterwards — better mood, more belonging, for both parties. Epley and Schroeder showed this in 2014, and Sandstrom's work finds even minimal contact with the loose outer ring of your life measurably lifts belonging. Those weak ties aren't a consolation prize; they're the on-ramp every close friendship once used.

Which leads to the last move: be the one who escalates. Someone has to send the "want to grab food after?" message, and the liking gap means you're both waiting, each assuming the other isn't interested. Assume interest. You'll be right more often than your instincts predict.

Expect the quiet stretch

So you don't quit at the worst moment: expect the first month of any anchor to feel like polite small talk with strangers, because that's what hour eight feels like. Expect a name-knowing, inside-joke stage somewhere in the second or third month. Expect the first hangout outside the anchor past that, usually because one of you finally escalated.

The hours research doesn't promise a schedule — people and groups vary. What it promises is a direction. Keep showing up and the line only moves one way.

None of this requires becoming a different person. It requires noticing that every friend you have ever had was manufactured by an environment — a hallway, a dorm, a shift, a fire — and that environments can be built on purpose.

Your great-grandparents never needed instructions for this, because their world was the instructions. Yours isn't. So now you have them.

The full argument, with every source on screen, is in the film — watch it on YouTube.