How many needles in a haystack? The search for user innovations in manual and algorithmic-augmented systems

Abstract

Innovation scholars and managers with interest in user innovation often search for user innovators and their innovations. Despite its importance, the search process itself has received little attention in the innovation management literature. This paper fills that gap and contributes to innovation theory in three ways. Firstly, we introduce a distinction between the one-off and continuous search, usually done by the researchers and managers of platforms for sharing user innovations, respectively. Secondly, and given the recent surge of interest in machine learning for innovation management, we explore the levels of automation of that search on a scale from entirely manual to fully automated, and different configurations of humans and algorithms. Thirdly, we propose and empirically test a simple quantitative model for the expected number of user innovations or innovators that integrates the human and algorithmic components. We discuss the theoretical, methodological, and managerial implications of this study.