Interaction Costs = f(Complexity)
For the last 10 or so years "interaction costs" have been on my mind. At first ("Business Intelligence: Potential Impact in Railway Freight Transportation," June 11, 1999) the phrase was confined to the interface between a railroad and its customer. Subsequently, however, I see as the costs of interaction between all nodes in the context of interest.
My hypothesis is that as the complexity of a structure increases the complexity of the interactions, and hence the associated costs, also increases. So much so, in fact, that growing interaction costs can cause the structure to collapse.
In "Simplicity: The Next Big Thing," Rosabeth Moss Kanter seems to be on the same line of thinking. She discusses the need for simplicity. This brings to mind the question of what happens when you remove complexity from a structure. Does it simply -- poof -- vanish? Or, as I suspect, is it displaced somewhere else. Much like the old practice in logistics of pushing inventory upstream, an illusion of progress.
Entropy is a measure of disorder in a system. Few, I suspect, would dispute the assertion that the entropy in the system of the world (thank you, Neil Stephenson) continues to grow as the complexity of the system grows.
We may simplify in a local region of the system, but, like moving inventory upstream, we're not really improving the overall performance.
When we simplify we need to understand the knock-on effects.
Reader Comments