Time-Based Crowdsourcing Contests
We study a contest wherein an organizer elicits solution(s) to a problem from a large population of agents at the earliest time. Each agent's solution time is random and it depends on agent’s effort and expertise. We call an agent whose ex-post solution output contributes to the organizer’s objective a contributor, and consider a general case in which the organizer minimizes the total solution time of any number of contributors. We establish that when agents’ solution times are less (resp., highly) uncertain, it is optimal for the organizer to screen and allow entry to only a small group (resp., a larger group) of highest-expertise agents. Furthermore, it is optimal for the organizer to compensate each agent based on the agent’s own solution time and on whether the organizer can observe agents’ efforts.
