The transactive memory system has been studied extensively, yet we still know little about how it influences the effectiveness of temporary teams. Additionally, little is known about the boundary conditions of the well-established benefits of transactive memory systems on team performance. Our primary goal in this study is to build and test a theory that investigates the influence of a transactive memory system on the performance of temporary teams while accounting for conditional effects of both task and relationship conflict. On the surface, a transactive memory systems perspective may seem incompatible with temporary teams. Transactive memory systems typically require time or team member familiarity to develop. However, team members on temporary teams often are selected because of their expertise, not team member familiarity, and often must quickly and effectively operate under time and outcome pressures. We present a theory that suggests transactive memory systems should have a meaningful influence on temporary teams, and its effect is accentuated in the presence of task conflict and attenuated in the presence of relationship conflict. We test our theory using a sample of 202 teams participating in the Global Game Jam, the world's largest hackathon devoted to designing and developing games within a 48-h period. In addition to implications for literatures on transactive memory systems and temporary teams, our study adds to a growing literature providing practical advice and insight regarding hackathons, a pervasive source of innovation and idea generation.