I was thinking again about the discussion here last week about fuzzy concepts – in particular that of networks. In that post, I was trying to get at the value of the network concept for economic analysis. Neil Kay made an interesting point in the comments – which was that you could substitute the word ‘map’ for ‘network’ all through the discussion and that might add some clarity. His point is that we need to define the boundaries of the network, along with the purposes for which a network is constructed.

In the main, I think that this is correct. But on the other hand, I’m not sure that ‘everything is a network’ actually is a strawman. The way I tend to think about networks is as an economic substrate, or as a basic analytic unit. In neo-classical economics, the economic substrate is formed by utility-maximising members of homo economicus, which follows from the basic axioms upon which orthodox economics is built. If we take an evolutionary approach to economics, as in The General Theory of Economic Evolution by Kurt Dopfer & Jason Potts, their set of axioms leads to an economic substrate of economic actors and the connections between them. In other words, a network.

From this angle, everything is a network, but that ends up being a trivial observation. This is where Neil’s comments come into play – the point is to figure out how networks differ, and how they serve different purposes. This is what I was getting at when I talked about the three forms of organisation – markets, hierarchies and networks:

I think it would be more useful to define the different forms of organisation as networks, and then look at the three questions of description, classification and evolution. From this perspective, the three forms have different network structures through time, which leads to differences in classification.

In light of this, the innovation network mapped in the figure above is interesting for a couple of reasons. It shows a project team from one of our research partners. The colours of the nodes indicate people working from different offices, and the size of the nodes indicates the person’s rank within the organisation (bigger = higher ranking). A lot of people talk about network structures as being inherently ‘good’ or ‘bad’, but I think it’s more useful to take Neil’s approach and make this judgement based on how effectively the structure serves the purpose for which it is designed.

The structure above has a clear split between people in the two geographic groups. Furthermore, most of the connections between the two sites goes through upper managers. From an information flow standpoint, this is a ‘bad’ network structure. However, when I was talking about this with the firm last week, we realised that for the purposes of this project, this may actually be exactly the right structure. Keep in mind that maintaining network ties is always expensive. If the two geographic groups are working on different parts of the project, there might not be a need for extensive knowledge sharing between them.

So I can take the evolutionary approach to this first by describing the network (a geographically distributed innovation network for a project-based firm, and so on), then by classifying it (still working on that bit – but clearly it is different from other types of innovation networks that have been researched to date), and finally I can track how it evolves. The last point is where the real action is – if this project continues to completion we’ll have a chance to map the evolution of the problem-solving network within the project over its life. If we can do this, then I think that will be a real contribution to network research.