Star-studded or equalitarianism: how does the distribution of creative stars affect exploration-exploitation balance?

Recent research has focused on creative stars who create disproportionate new insights. Despite this attention, organizations have to deal with the puzzle of how to distribute the limited number of creative stars to strike a balance between exploration and exploitation. This study presents a multi-agent simulation model and compares the performance of organizations with varying distributions of creative stars. We find that if the organization consists of subgroups that are minimally connected, moderately to highly concentrated creative stars can reap the benefits of joint exploration. As the confusion increases, the centralized distribution of creative stars becomes more advantageous. When facing particularly high confusion, organizations with moderate dispersion of creative stars benefit more from cross-group links. High centralization levels will eventually be overshadowed by the increased difficulty of leveraging knowledge. In the face of a low degree of knowledge localization and high confusion in exploration, the impact of degree of concentrated distribution of stars on learning performance takes an inverted-U shape when stars act the linking-pin role.