Optimal Fee Structures of Crowdsourcing Platforms

Crowdsourcing platforms specialize in hosting open contests and usually charge a percentage of the prizes as service fees. While prior research has studied the design of contests and the behavior of contestants, the strategy of a crowdsourcing platform has remained largely unexplored. We develop a game-theoretic model of crowdsourcing services and find the optimal fee structure of a platform. We prove for the case of a single contest that the service fees should be an increasing concave function of task prizes and show that this also holds true for the case of multiple contests. We further find that for a platform with many users and tasks, there is an optimal ratio of the number of contestants and contests. Our research is one of the first to focus on the strategies of crowdsourcing platforms and our results have interesting managerial implications. We show that the linear fee schedule widely used in practice is not optimal and that a platform is better off lowering the fee rate for contests with high prizes. It is also in the best interests of a platform to develop both sides of the crowdsourcing market proportionally and keep the ratio of contestants and contests at the optimal level.