Taming the Cluster Concept: Mapping the Technology Landscape through a Small World Network Configuration

Over recent decades a discourse has become established which emphasises the over-riding importance of close proximity for harnessing knowledge spillover, and therefore for innovativeness. The most obvious manifestation of this discourse is the concept of the ‘cluster’ associated with the work of Porter (1998) which is now ubiquitous in policy circles. However, many authors have now begun to question this discourse of close proximity. Those that question it often highlight the problems associated with over-embeddedness resulting in ‘lock-in’ in order to show that agglomeration and close proximity is not purely a good thing. Bathelt et al (2004) have commented that, despite the now widespread acknowledgement that agglomeration and clustering can also be harmful as well as beneficial, ‘there still seems to be a missing link in the conceptualisation of the connection between internal and external cluster relations.’ This paper demonstrates that Watts’ and Strogatz’s (1998) ‘Small World’ network can fill this missing link. Beyond this, and crucially, it is demonstrated by use of a multi-agent based model that because of the inherent properties of a ‘Small World’ network firms adopting such a network configuration can collectively arrive at more good innovations at lower cost by ‘mapping the technology landscape’ (Fleming and Sorenson, 2003) and thus shrinking the innovation problem space.