Exploring idea selection in open innovation communities: A stochastic cusp catastrophe model perspective

With the spread of crowdsourcing models, open innovation communities are able to collect ideas quickly, but it is still a major challenge to select high-quality ideas from a large number of user contributions. Adopting a nonlinear and complex perspective, this study employs a stochastic cusp catastrophe model to shed light on idea selection. We utilised a unique dataset from 12 years of LEGO IDEAS to track the idea selection process over a six-month period at weekly intervals. First, we applied machine learning methods to identify three key factors influencing idea selection: “view,” “comment,” and “update.” Then, we performed a coordinate transformation based on the equilibrium surface in the cusp catastrophe model to build an idea selected catastrophe model. This model illustrates that idea selection in open innovation communities exhibits discontinuous catastrophe. Through catastrophe analysis, we categorized ideas into four types, each with distinct managerial value, highlighting the importance of focusing on “promising ideas” near the selection threshold. Furthermore, adjusting control variables along feasible paths can make previously unselected ideas into selected ones, facilitating the identification of high-quality ideas. Our research contributes to the existing literature on idea selection in open innovation communities, and provides practical insights for innovation managers.