For generations, managers have been taught that process innovation, especially where information technology is concerned, supports and is wholly subservient to an agreed strategy. After six months or so of strategy definition, top managers toss process design responsibility over to their technical people, who embark on an exhaustive design-and-build cycle. During as much as a year of systems analysis—the methodology on which most software and process redesign is still based—the technical department is meant to come up with ideas for improved, computerized processes. There follow two to three years of implementation, by which time the creative ideas that have struggled forth from such a lengthy procedure are likely to be outdated.

The new science of business process reengineering too often follows a similar pattern. Crossfunctional teams may come up with new ideas, but as these often lack state-of-the-art technical input they tend to be incremental, and can still take years to reach implementation.

Product-led corporations would not dream of researching innovation in this way. A pharmaceutical company needing a stream of new drugs, for example, would be crazy to expect a technology department or part-time team of line managers, guided by complex and monolithic methodologies, suddenly to come up...