How Many Friends Do You Really Need?

Image credit: Desktop2Wallpaper.

Chimpanzees are our closest ancestors in the evolutionary chain – they share more than 98% of our DNA. This means that their behaviour in certain situations might well be a good indicator of our own.

Professor of Evolutionary Anthropology at the University of Oxford, Robin Dunbar, has spent much of his life studying chimpanzees to help us better understand questions about ourselves and our world.

One of the theories that has come out of his research is the concept of Dunbar’s number. He claims that this is the maximum number of friends any one person can have. [1]

In this blog post I aim to explain why this theory is important in the increasingly social world of digital.

The chimpanzee and another ape, the bonobo, are humans’ closest living relatives.
Humans and chimps share a surprising 98.8 percent of their DNA.
By studying how chimpanzees behave we can potentially understand more about human behaviour. [2]
(image: (c) AMNH Exhibitions. Banding Patterns)

To arrive at Dunbar’s number, first we have to look at something called social intelligence theory.

Social intelligence theory

Various scientists over the centuries – biologists, zoologists and anthropolgogists – have studied the brain size and relative intelligence of different animals. What is now widely accepted is that the animals which are deemed the most intelligent (and have the largest brain size relative to their body) – corvids (birds of the crow family), dolphins, primates – apes, chimpanzees and humans – all have more complex social lives than less intelligent animals. [4]

There is a strong correlation between how social a species is and how big the species’ brain is.

This is known as social intelligence theory or the social brain hypothesis. [5]

Average social group size plotted against relative brain size (neocortex size) – The Social Brain Hypothesis, Dunbar 1998. [6]

From social intelligence theory to Dunbar’s number  

Professor Dunbar crunched the numbers on limits of group size (i.e. number of friends) for various species of ape – and found that as brain size increases so does the maximum group size.

Given that we know the maximum size of the human brain, it is straightforward to work backwards and conclude the upper limit on human group size. This is how Professor Dunbar reached the number 150 – the maximum number of friends/relationships one person can have. 

Primates are above all social animals: that is their big evolutionary breakthrough. It’s what has made them as successful as they have been and, by extension of course, it is what makes humans so successful – we have inherited the same social expertise. What marks primates (or at least monkeys and apes) out as different from all other species of amimals is the sheer intensity of their social interactions. The difference between the rest of our primate cousins and us is that we have simply taken this trend to a whole new level.[1] 

Examples of Dunbar’s number throughout history

He then looked back in history for typical social group sizes to see if his number cropped up anywhere. This is where it started to get interesting.

Tribal groups

Of twenty tribal societies that had data held about them, the average clan group size for the tribe was 153.

Village groups

Villages from the Middle East in 6000 B.C. to England in 1086 were all about 150 people in size.

English villages in the 11th Century were on average 150 people in size – matching Dunbar’s number [6]

Military Groups

In military units a company – the smallest fighting unit – contains 130-150 men including commanding officers and support. This is similar to what it was in Roman times.

Business Groups

An example from the world of business: the owner of GoreTex always capped his factory sizes at 150 staff. Above that limit he said that people became unfamiliar with each other and were unable to work as productively.

Why Dunbar’s number?

Why would there be an upper limit on the number of relationships/friends one person can have? One reason is that the quality of relationship is even more important than the quantity. In a complex social group – knowing who did what to whom, what the hierarchies are, where the allegiances are today – and tomorrow – is important to an individual’s survival and success.

From the “sympathy group” to the ideal size for a democracy

Further research concluded that there are other patterns to group sizes for the average human. Most people have a small group – of three to five very close friends. The various layers of friendship – which increase in number but decrease in intimacy and frequency of contact are on average:

Layer 0, Nucleus / Very close friends – those who you could turn to in a crisis, ask for money, lean on for support: on average 3 to 5 people. It is most likely you keep in touch with them once a week.

Layer 1, Close friends: 12-15 people (social psychologists name this the “sympathy group” [3] – which incidentally is also the number of Apostles, members on a jury etc.). You are likely to contact them once a month.

Layer 2, Distant friends: 45 to 50 people

Layer 3, Maximum number of friends/acquaintances: 150 people (Dunbar’s number)

Layer 4, 500 people

Layer 5, 1500 people

Layer 6, Plato’s ideal size for a democracy – 5300 people

Does this research have any implication for social software design?


There is very likely a measurable limit of the maximum number of meaningful relationships in humans. As with anything else relating to brain function there will be a normal statistical distribution of those in the population with fewer friends, those with more, but the majority will have a similar average number. One of the key determinants of how many friends we have is our fixed mental capacity – our fixed brain size.

Perhaps this number is as Professor Dunbar claims: one hundred and fifty?

Implications of Dunbar’s number for digital design

What does this mean for digital design, user experience, and online community?

Firstly, if Dunbar’s number and the layers of friendship are true, it seems to point to an emphasis in humans on quality of relationships rather than number of relationships.

Accumulating thousands of digital friends is surely futile if the maximum number of meaningful relationships we can maintain in our lives is limited.

Can our brains – which have not changed in thousands of years – manage more than a limited number of relationships? [9]

Therefore, designing applications which enable more than a certain number of connections per user could rapidly diminish the potential quality of any of those relationships. This does not mean that social software must be limited to 150 friends or connections. However, bearing in mind the cognitive limit on how useful those friendships / relationships can be, it is worth taking into consideration. Especially if meaningful engagement and retention is an important metric of success for you and your organisation.

Secondly, the circles of close relationships are even smaller: 5, 15, 45, and 150 people – are the upper boundaries of numbers of closer relationships for the human brain.

Given that, as far as we know, our brain has not evolved or increased in size or complexity in the last few thousand years, it makes little sense to design software that does not take these natural constraints into account.

Perhaps our cousins the chimpanzees have an important contribution to make to the design of social technology?


  1. “How many friends does one person need?” Dunbar, R. 2010, Faber & Faber, London.
  2. DNA: comparing humans and chimps
  3. Social network size in humans. Dunbar, R & Hill, R. 2002.
  4. Social Intelligence. Emery, Clayton, Frith. 2008. Oxford University Press.
  5. Understanding primate brain evolution. Dunbar, R. 1997, 2006, Social Intelligence.

Other references.

Inside the Animal Mind, Chris Packham. (BBC Television)

Exit mobile version