Wikipedia Deep Dive

Dunbar's number

10 min read

The 150 People You Actually Know

Here's a test: imagine you're walking into a bar, and you spot someone you recognize. Would you feel comfortable walking up to them and joining their table without an invitation? If yes, that person falls within your Dunbar's number. If the thought makes you cringe with social anxiety, they don't.

That's not a metaphor. That's literally how Robin Dunbar, the British anthropologist who gave this concept its name, explains it.

The number itself is roughly 150. That's supposedly the maximum number of stable social relationships a human being can maintain—relationships where you know who each person is, how they relate to you, and crucially, how they relate to each other. Beyond that threshold, your social brain starts to struggle.

Monkey Brains and the Grooming Problem

Dunbar didn't pull this number from thin air. He started with monkeys.

Primates are intensely social creatures, and they maintain their social bonds through grooming—picking through each other's fur, removing parasites, building trust. It's therapeutic. It's political. It's how you know who has your back when a leopard shows up.

But grooming takes time. A lot of time.

Primatologists noticed something interesting: the size of a primate's neocortex—the wrinkly outer layer of the brain responsible for complex thought—correlates with the size of its social group. Bigger brain, bigger social circle. This makes intuitive sense. Keeping track of who's allied with whom, who slighted whom last Tuesday, and who owes whom a favor requires serious cognitive overhead.

Dunbar took this correlation and ran with it. In 1992, he gathered data on 38 different primate species and plotted neocortex size against group size. Then he asked the obvious question: if we plug in the human neocortex, what number do we get?

The answer was 148. Rounded up for convenience: 150.

A Number That Keeps Showing Up

What makes Dunbar's number compelling isn't just the primate math. It's that 150 seems to pop up everywhere in human society, across cultures and centuries.

Neolithic farming villages? About 150 people.

The Hutterites, a communal religious group that's been carefully studied by anthropologists, have a tradition of splitting their communities when they grow too large. The splitting point? Around 150.

Roman military units? The basic tactical unit was the century, which despite its name contained roughly 80 to 100 soldiers. But when you add in support staff, you're looking at organizational groupings that cluster around that same magic number. Modern military companies follow similar patterns—150 is approximately the size where a commanding officer can still know every soldier by name and reputation.

Academic subfields? About 200 specialists, maximum, before the field starts fragmenting into further subspecialties.

Dunbar went looking for his predicted number in the anthropological literature, examining hunter-gatherer societies—the closest modern analog to how our Pleistocene ancestors lived. He found social structures falling into predictable bands: small groups of 30 to 50 (what anthropologists call bands), medium groups of 100 to 200 (clans or lineage groups), and larger tribes of 500 to 2,500. That middle tier, the natural human community, centers right on 150.

The 42 Percent Problem

There's a catch, though.

Dunbar calculated that for a group of 150 people to remain truly cohesive—for everyone to maintain meaningful relationships with everyone else—members would need to spend about 42 percent of their time on social grooming.

That's nearly half your waking hours spent on nothing but relationship maintenance. No hunting. No gathering. No building. Just endless social upkeep.

Obviously, that's not sustainable. Which is why, Dunbar argues, most human groups don't actually hit 150 unless they're under intense pressure to stick together. Subsistence villages facing harsh environments, nomadic tribes competing for resources, military units whose survival depends on unit cohesion—these are the contexts where 150-person groups actually form and hold together.

For everyone else, the number tends to be lower. Dispersed communities, where people don't see each other daily, maintain smaller networks because relationships decay without regular contact.

Why Language Changed Everything

So how did humans escape the grooming trap? How do we maintain social groups at all if relationship maintenance should consume half our lives?

Dunbar's answer: we invented gossip.

In his book Grooming, Gossip, and the Evolution of Language, Dunbar proposes that human language evolved primarily as a tool for social bonding. Talking is grooming at scale. Instead of picking through someone's hair for an hour, you can bond with them in five minutes of conversation—and you can do it while simultaneously walking, cooking, or working.

Even better, language lets you maintain relationships at a distance. You can talk about people who aren't present, gather intelligence on the social landscape, and keep track of alliances and conflicts throughout your entire network without direct observation.

This theory has mathematical support. Researchers modeling the "social brain hypothesis" have shown that increased brain size alone couldn't produce large groups without some form of complex communication. Language was the key that unlocked human sociality.

Layers Within the Number

The 150 figure is just the outer boundary. Within it, Dunbar has identified concentric circles of intimacy.

Your innermost circle contains about 5 people. These are your closest confidants—the ones you'd call at 3 a.m. in a crisis, the ones whose deaths would devastate you. Dunbar's research suggests we spend about 40 percent of our available social time on just these five people.

The next ring holds about 10 more people, who receive another 20 percent of your social attention. These are close friends, though not quite at the inner-circle level.

Add them together and you get 15 people who absorb roughly two-thirds of your social energy. The remaining third gets distributed among the other 135 people in your network—acquaintances, colleagues, that guy from college you haven't talked to in years but would happily catch up with if you ran into him.

This layered structure explains a common experience: you can have 500 Facebook friends and still feel lonely. The outer layers of your network don't provide the same emotional sustenance as the inner ones. Quantity doesn't substitute for depth.

The Corporate Connection

Business leaders have taken Dunbar's number seriously, sometimes with fascinating results.

Gore-Tex manufacturer W. L. Gore and Associates discovered the 150 limit through trial and error. When more than 150 employees worked in the same building, social problems emerged—cliques formed, communication broke down, the sense of shared purpose eroded. So the company started building facilities with exactly 150 parking spaces. When the lot filled up, they'd build another building, sometimes just a few hundred meters away.

The Swedish Tax Agency reorganized in 2007 with a maximum of 150 employees per office, explicitly citing Dunbar's research.

Flight Centre, an Australian travel company, structured itself into nested Dunbar-sized units: "families" (individual stores), "villages" (clusters of stores), and "tribes" (never exceeding 150 people total).

There's something profound here about the difference between a community and an organization. Below 150, you can run things on trust and personal relationships. Everyone knows everyone; reputation matters; social sanctions work. Above that threshold, you need formal rules, hierarchies, and enforcement mechanisms. The village becomes a city, and cities require police.

When Networks Go Dark

This insight has darker applications too. Security researchers have analyzed terrorist networks and criminal organizations through the lens of Dunbar's number, looking for the structural constraints that limit how such groups can organize and operate.

If even terrorists can't maintain coherent organizations beyond 150 closely connected members, that creates predictable vulnerabilities. Cells must stay small or fragment. Trust becomes harder to establish at scale. The same cognitive limits that make your office need explicit org charts also constrain organizations with less benign purposes.

The Critics Speak Up

Not everyone buys the 150 figure.

Anthropologist H. Russell Bernard and his colleague Peter Killworth conducted field studies across the United States and came up with a substantially higher estimate: 290 social ties on average, with a median of 231. That's roughly double Dunbar's number.

Other researchers have attacked the methodology directly. When scientists attempted to replicate Dunbar's original analysis using updated datasets and different statistical methods, they got wildly inconsistent results. Bayesian analysis suggested average group sizes between 69 and 109. Generalized least-squares methods produced estimates between 16 and 42. More troubling, the confidence intervals—the statistical measure of uncertainty—were enormous, ranging from 4 to 520 in some analyses.

The critique goes deeper than numbers. Primate brains don't all work the same way, and human brains certainly don't process social information identically to chimpanzee or gorilla brains. Moreover, primate group sizes are influenced by factors beyond brain power: diet, predation pressure, habitat. A monkey that eats rare, scattered food sources must range widely and can't afford to maintain a large group regardless of how big its brain is.

Philip Lieberman, a cognitive scientist, raises a different objection. Hunter-gatherer bands top out at 30 to 50 people largely because of nutritional constraints—that's how many mouths you can feed without agriculture. If our Paleolithic ancestors couldn't form 150-person groups anyway, there was no evolutionary pressure to select for brains capable of managing them.

Some species with tiny brains manage impressive social feats. Paper wasps maintain hierarchical societies of about 80 individuals where each wasp knows its place in the pecking order. Computer simulations have produced artificial agents with simple programming that exhibit complex social behaviors resembling primate politics. Maybe you don't need a massive neocortex to run a society.

The Number That Wouldn't Die

Dunbar himself has evolved his position over the years. He's acknowledged that his original analysis, based on primates broadly, might be less reliable than an analysis focused specifically on great apes. The problem is that great apes include only a handful of group-living species—chimpanzees, bonobos, and two species of gorilla—which isn't enough data for rigorous statistics.

Yet despite the methodological critiques, Dunbar's number refuses to disappear. It's been cited in contexts ranging from online gaming communities to social media research. A study of the European professional networking site XING found that users with approximately 157 contacts reported the most success in receiving job offers—close enough to 150 to be intriguing.

Research on mobile phone networks and Facebook friendships has found patterns consistent with Dunbar-sized limitations, even when the technology theoretically allows for unlimited connections.

Perhaps the number's staying power reflects something real, even if the precise figure is fuzzy. We all have intuitive experience of the difference between knowing someone and merely knowing of them. There's a point where your social network exceeds your social cognition, where you can't keep track of everyone anymore, where some people inevitably fade into the periphery.

What This Means for Choosing Your People

If Dunbar is even approximately right, there's a profound implication: your social bandwidth is finite.

You cannot maintain unlimited close relationships. Every person you add to your inner circle displaces someone else, or at least dilutes the time and energy available for existing relationships. Your five closest confidants really are a zero-sum game.

The writer David Wong (of Cracked.com fame) explored this idea in a darkly humorous essay called "What is the Monkeysphere?" His argument: we're all walking around with tribal brains in a global civilization. We can only truly care about our 150 people, our "monkeysphere." Everyone outside that sphere isn't quite real to us—they're abstractions, statistics, characters in news stories.

This doesn't make us evil. It makes us human. But it does explain some of our species' more troubling tendencies. Racism, xenophobia, indifference to distant suffering—these become more comprehensible (if not more excusable) when you realize that our brains simply weren't designed to care about eight billion people.

It also suggests that the question of who you include in your 150 matters enormously. If your social world is inherently limited, then every slot becomes precious. The people you walk beside, as the saying goes, shape your path. They occupy cognitive real estate that could go to someone else.

Choose carefully. You only have 150 spots, and you're probably already over capacity.