I have been pondering the ideal size of a community of late – be it a company (successful companies are communities too), a circle of friends or the user base for a wiki or a forum. Of course, I am not the first to ponder the question, nor will I be the last.

Paul Graham wrote an essay “You weren’t meant to have a boss” which is really about large versus small companies, and it raises some good questions – if a little controversially.

However, it is Robin Dunbar (now teaching at Oxford) who produced the most famous research, back in 1993. His work was popularised in Malcolm Gladwell’s book ‘The Tipping Point’ – to such extent that many in the social media space talk about “Dunbar’s number“.

Dunbar took research on non-human primate social groups and used some (fairly finger in the air – by his own admission) statistical methods to extrapolate this to humans. Based on our brain’s larger neocortex size, he predicted 150 as the mean size limit for a human’s meaningful social network.

This has been widely used as a sound-bite, even featuring in The Wall Street Journal (article written by Carl Bialik). In a later paper (2003), Dunbar talks about a number range of 100-300 as the number of people in our social world (defined as the people we might turn to in severe stress, or at least approach at the airport if we needed help).

Now that we have social software, we can study people’s social graphs in ways that were very difficult previously. That said, some of the research is a little esoteric, for example Christopher Allen has an interesting post with links to research on the playon blog, looking at data from groups on World of Warcraft (for the non-gamers out there, this is a massively multiplayer on-line role playing game). We have to remember that on-line games and social networking sites like Facebook don’t directly relate to real life relationships – much as the average Facebook addict might find that hard to accept.

Researchers also have the concept of subgraphs – essentially tighter ‘cliques’ that exist within the social graph, as clusters of more tightly meshed relationships, that is individuals who have more mutual friends. In his book ‘Evolutionary Psychology’, Dunbar talks about circles of intimacy – different rings of friendship, with different levels of intimacy. We can map that to our own lives, where we usually have a smaller group of people that we are closer to.

Although Dunbar doesn’t use social networks, his view is that they might help our brains push past this limit. However, on-line networking doesn’t replace the social grooming required to maintain relationships. We still need to meet ‘IRL’ (in real life). He isn’t sold on the idea that social networks make his number outdated. Language may provide a cheaper form of social grooming – it certainly beats picking nits out of your friend’s hair – but it isn’t clear if communication technology provides even greater short cuts. The research, Dunbar says, “made us realize people don’t know what these wretched things called relationships are — and that helps explain why we’re so bad at them”.

Continued in Part II >>