Users represent an often untapped source of knowledge which companies can capitalize on during different stages of the innovation process. However, identifying helpful users for innovation projects is far from trivial as these individuals are often hidden within considerably larger populations. We contribute to open and user innovation research by empirically investigating the efficiency of pyramiding and screening, two methods used to identify valuable users. Analyzing a sample of 942 children in 42 school classes, we show that pyramiding, a search process based on personal references from user to user, is significantly more efficient in identifying rare individuals than screening, even though many references are not based on close personal relations. Pyramiding's relative efficiency advantage increases with the size of the population being searched. Finally, we explore how searchers can further increase pyramiding search efficiency by using information to select promising starting points or prematurely abandoning unpromising search chains.
