Abstract

This article develops a framework for driving innovation under highly ambiguous conditions. An analysis of the most novel medicines of the past 20 years shows that a very large group of small companies created more breakthroughs, at considerably less overall cost, than a much smaller group of very large companies. This article’s findings present the first large-scale empirical validation of the theoretical literature predicting the superiority of decentralized parallel searches in ambiguous environments. Accordingly, companies that attempt to “de-risk” the innovation portfolio by narrowing their search efforts to minimize failures run the risk of filtering out the next breakthrough.