Parallel Innovation Contests

We study multiple parallel contests where contest organizers elicit solutions to innovation-related problems from a set of agents. Each agent may participate in multiple contests and exert effort to improve her solution for each contest she enters, but the quality of her solution also depends on an output uncertainty. We first analyze whether an organizer's profit can be improved by discouraging agents from participating in multiple contests. We show, interestingly, that organizers benefit from agents' participation in multiple contests when the agent's output uncertainty is sufficiently large. A managerial insight from this result is that when organizers elicit innovative solutions rather than low-novelty solutions, agents' participation in multiple contests may be beneficial to organizers. We further show that an organizer's profit is unimodal in the number of contests, and the optimal number of contests increases with the agent's output uncertainty. This finding may explain why many organizations run multiple contests in practice, and it prescribes a larger number of contests when organizations seek innovative solutions rather than low-novelty solutions