There has always been a polarity of views between those who see Communities of Practice as something that should be allowed to flourish naturally, springing up as a bottom-up initiative in response to user demand, and those who see communities as more powerful when they are aligned with the business strategy, and structured to provide a valuable resource to their members.
This is the "emergent or structured" debate.
I have always been in the "structured" camp, and while recognising communities of practice as an organic structure, would rather see them organised into a productive garden, than allowed to run to jungle. This view seems to be supported by recent research
Now, as part of a Linked-In discussion on "Communities of practice, limited or unlimited", we have a useful piece of data from Arjan van Unnik, who was head of Knowledge Management at Shell.
Arjan says that in 2006, Shell relaxed it's policy of "managed structures CoPs", and threw CoPs open to anyone who wanted one.
The net result was that, rather than increasing, CoP use decreased by 32% as a result of the fragmentation that followed.
Arjan says that he then had to "implement corrective actions and increase the management of the CoP portfolio" in order to get things back on track.