MANAGING THE UNCERTAINTIES OF INNOVATION - EXTENDING THOMPSON (1967)

Thompson's two domain model, with its central core of rational activities separated from a mantle of boundary-spanning activities, tells us how organizations cope with externally generated uncertainties. Innovation is an internal source of uncertainty that this model does not adequately cover. Following Burns and Stalker, we suggest that the rational core and the organic boundary-spanning are incompatible domains of activity with different modes of governance. Many writers overlook the problems of coupling these domains and simply argue that ''upstream'' innovation activities can be managed organically while ''downstream'' implementation activities can be mechanistic. Sometimes they see a mixed mode of governance being applied between the project's initiation and completion points. We argue that far from there being a gradual transition from organic to mechanistic, there is a dialectical tension between these modes which is best managed by ''nesting'' the organic within a host matrix of bureaucratic activity. This recalls Bower's model of the resource allocation process in a stratified organization. Thompson's model, apparently focused on externally generated uncertainties, is thereby extended to deal with the internally generated uncertainties of the innovation process.