Modeling Team Competition on On-Demand Service Platforms

Abstract

Activating self-scheduling workers on on-demand platforms when services are most needed is challenging because of a lack of cooperation between workers. To align workers' interests with the platform's profit-driven goals, various ride-sharing and food delivery platforms have recently embraced team competition. The platform declares payment schemes to allocate individual workers' payoffs and stimulate productivity gains, and workers collaborate with team members to compete for customers with other teams. This payment scheme design problem is modeled as a single-leader multi-follower game. The lower-level equilibrium analysis employs quasi-variational inequalities to capture intra-team coordination and inter-team competition. The upper-level optimal payment schemes are computed by a novel algorithm that integrates Bayesian optimization, duality, and a penalty method. The benefits of team competition are manifold. The team competition with a well-designed payment scheme can direct work schedules toward the profitable market equilibrium. When workers have an inaccurate perception of the market, team competition benefits both the platform and workers. Teams can strategically mitigate the negative externalities caused by individual workers' over-competition. By establishing a model framework for studying hierarchy team-based structures, this work helps operators gain a deeper understanding of mixed decentralized and centralized control in on-demand services.