Christopher Allen has posted a thoughtful and detailed account of the way in which Dunbar’s number influences maximum and optimum group sizes in our offline and online social interaction. I had a long, interesting chat with Christopher Allen on a San Diego sidewalk a few weeks ago and I know he is very interested in the use of social software tools and ideas to strengthen civil society and help non-profit agencies operate better. This post shows that he is thinking through these issues with an attention to detail and analytical precision that befits somebody of his experience
R.I.M. Dunbar is an anthropologist whose paper “Co-Evolution Of Neocortex Size, Group Size And Language In Humans” proposed that a single person can maintain a collaborative relationship with a maximum of about 150 people at any one time. This idea has since been misinterpreted to suggest that 150 is some kind of optimum group size, which it is not
Allen quotes various sources that suggest in fact there are multiple plateaus of group size, notably around 12 for small groups and 40-50 for larger groups, and in fact groups as large as 150 are relatively rare because they are hard to manage. Dunbar’s original research seems also to relate to fairly specific types of group where people are bound together in mutual dependence for survival or other reasons that are a lot stronger as social drivers than the bonds that link most business collaborations or online communities, and quotes Dunbar as saying

“The group size predicted for modern humans by equation (1) would require as much as 42% of the total time budget to be devoted to social grooming … My suggestion, then, is that language evolved as a “cheap” form of social grooming, so enabling the ancestral humans to maintain the cohesion of the unusually large groups demanded by the particular conditions they faced at the time.”

This is an interesting thought in an online social context where the problems of information overload and maintaining contact with lots of different people are taking us further and further away from the kind of social grooming and interaction that our brains are built to support. It reminds me again of Matt Webb’s glancing project
It also very much chimes with our experience of running an organisation at various sizes from 2-55. We were probably most productive at about 5, 12 and 25 – from 40-50 is definitely a tricky period where the overheads and friction caused by introducing a layer of middle management for this first time are not yet offset by organisational efficiencies or productivity gains. It is no coincidence that the second richest seam of amusing people stories we carry with us emanate from this period – half-day meetings to discuss whether celebrating Christmas within the company was a good idea or not and the existence of an entire team dedicated to gathering knowledge for some unspecified future purpose (mea culpa on both counts). The richest seam, as some will know, was the denouement that followed, pigeons et al. Lesson learned: whilst our communitarian organisational culture in those days was highly enjoyable and quite effective, it was particularly vulnerable to the effects of bad hiring decisions made under pressure
Funnily enough, we are thinking about these issues again now because the temptation of growth has reared it head again. Our core team if currently eight people who are all talented, autonomous and work together with little friction or wastage, and we always planned to go no further than twelve before we consolidate value, but projects are coming thick and fast and we may have to get there sooner than we anticipated. There are several things we could try to maintain a flat peer structure, from arbritarily increasing our prices to reduce demand to a level that a team of 12 can manage to simply dividing amoeba-like and creating a cluster of small teams. It is an ongoing discussion and I hope we have enough time to work it out
How does this relate to social software design? Well, there is the obvious conclusion about different tools and design for different groups (5, 12, 25, 40+), but I think there is more to think about here, such as how to track and respond to patterns of group growth and activity in online systems. Food for thought…..