Friday, September 21, 2012

A simple taxonomy for collaborative groups

What sorts of structures do collaborative groups come in?

I sat down this morning to write something about models for collaborative communities, and ended up writing a rather long post on communities, collectives and networks.  Or, more properly, on a few different stable group types I've noticed reappearing, that seem to form a set of 'archetypes' - as a result of group dynamics, and the different sorts of structures that can be stable.  The labels I've used make sense to me, but I haven't used them to be consistent with others - be warned!

This is an entry in my Collaborative Communities wiki - not currently online - which is the start of a project to create a book on Collaborative Communities, and how to work with them.

Group Taxonomy

Different group sizes, shapes, methods of interaction etc. support a range of different group types. There are a few particular, stable archetypes that many groups seem to approximate.

  • Communities
  • Collectives
  • Networks
These represent distinct - overlapping, fuzzy-edged - group types.

This is not an attempt to define or explain these terms - but rather an attempt to describe different group archetypes. Most people will probably disagree with this vocabulary!

In particular, 'communities' is normally used for any sort of group of any size that has internal connections and its own (emergent) behaviour, especially if that has a particular community 'feel'. It is also used to describe groups of greatly varying size based on geographical area (e.g. from a block to a district of a city with millions of people in it), or any group of people with a common identifier (irrespective of whether they are connected or identify with that group - for instance 'gamers'). In short, you can get away with calling any group a 'community'.


For the purposes of this project, communities are considerd to be groups, to which the members of which self-identify as 'belonging', with members connected to one another (primarily via first-order connections), and feel connected to one another, by virtue of being in this group.

Communities under this definition tend to be quite modest, by the standards and possibilities of modern connective capacity. (This is because they are constrained by Dunbar's Number, which is a limit on people's capacity to understand the set of human relationships in a group.) Depending on how permeable their edges are, these groups have a typical, maximum stable size at somewhere between 50 (for highly permeable groups, e.g. coworking communities) and 150 (for closed groups - e.g. villages or tribes).

A stable community can be supported in the most part by informal social norms and relationships. Few rules are needed, and policing and enforcement mechanisms are unecessary. Community maintenance roles can be largely informal.


Once a group extends beyond this stable size, it will tend to either splinter and break into two, or it will require different supporting structures and take on different dynamics.

In the past, with limited connectivity tools, this usually meant bureaucratisation was necessary - the creation of a structured heirarchy, or some sort, with a centralised rule and enforcement framework. The cost of introducing these is high enough that a group will often go through an unstable 'danger' period as it leaps from being a stable community to a sustainable organisation.

Contemporary connection technology means different (more cost effective) mechanisms can be introduced to support the expansion of a group into taking on a new, flatter, leaner form, which we're calling a collective.

A collective has some sort of an underlying point of connection (which might be a purpose, or shared attribute between members), and a sense of membership based on this point of connection. Collectives foster relationships between members, but these need not be primarily first-order connections (i.e. it doesn't matter whether members know the other members), and while there may be a point of connection, there is less affinity between members.

The ability to efficiently and effectively connect with many more members than the Dunbar number does not give us the ability to 'cheat' this number - it just means we can have very large networks. In practice, we will not be able to hold in our minds the set of relationships between the people we are connected to, and the emergent group dynamics will therefore be different.

A (mature) collective will normally have unevenly distributed connection density - with a core that looks a lot like a community, attached to an extended network. (The reason this is not just a community and a network together is that there is no real boundary between the core and the rest of the network - all are part of the collective.)

There are no real bounds on how many members can be part of a collective. Some collectives might be more sustainable at certain numbers (due to economies of scale), for instance 200+, but group dynamics can be stable at virtually any size - with appropriate supports in place.

Collectives require additional supports that communities don't need - though unlike bureaucracies, they don't necessarily need to be centralised. These supports might include:

  • formalised rule systems, including enforcement mechanisms (covering things like social behaviour and collaboration)
  • advanced communication tools (beyond face-to-face contact and direct emailing)
  • formal decision making methods
Some of these things may be applied successfully in smaller communities, but are not normally needed for communities to function.


Networks are groups of people joined by interpersonal networks, but not necessarily held together by any shared agenda or rallying thread.

Everybody (not in isolation) is part of large networks, by virtue of their personal connections, and their connections, and their connections... however being able to utilise these connections to collaborate effectively is relatively new. Modern communications technologies make it possible for networked collaboration to exist in ways that simply weren't feasible before.

Our 'mass' engagement methods are symptomatic of past challenges of coordinating large numbers of people. These methods typically aggregate participants as a large number of individuals with limited capacity to interact. Mass media and democratic voting systems are good examples. Some platforms fostered emergent network behaviour better than others - economic markets, for instance.

Modern technology makes radically different collaboration possible. Wikipedia is the most obvious example - Wikipedia provided a platorm that enabled stygmergic collaborative between millions of individuals, in a radically efficient way, and in only a few years made bureaucratically-managed encyclopedias redundant.

1 comment:

  1. On reflection - I would probably call that second group a TRIBE, not a collective. That would be pretty consistent with Godin and others.