When A Platform Competes with Third-Party Sellers in Networked Markets: A Revenue Management Perspective

Abstract

We consider a platform marketplace with both third-party and first-party sellers. The platform determines prices for first-party sellers and charges commissions to third-party sellers and buyers for transactions. Compatibility between heterogeneous types of sellers and buyers is captured by a bipartite network. Given the platform's prices and commissions, buyers/third-party sellers maximize their own payoffs from demanding/supplying products, and market-clearing conditions are satisfied. By characterizing the network structure with Bonacich centrality, which features complement and substitution in agents' utilities, we identify that as the Bonacich centrality of first-party sellers increases in the network, the platform should raise the optimal prices for these first-party sellers. In contrast, as the Bonacich centrality of other market participants increases, the platform should reduce the commissions. This highlights a fundamental difference in the roles of prices and commissions as revenue management tools in a two-sided market. In addition, when expanding first-party sellers, we establish that the widely-adopted greedy-type policy in practice that targets the agent type with the highest marginal value can perform arbitrarily poorly. On the theoretical side, this contributes to the literature on applying the greedy policy to submodular/supermodular function minimization problems. To achieve good performance, the platform needs to incorporate network structure in the design of its expansion strategy. Lastly, motivated by the fairness consideration between the platform and its market participants, we develop an efficient $(1-\epsilon)$-approximation algorithm to obtain a price-commission profile under which a fair allocation of surplus between the platform and its market participants is guaranteed in equilibrium trades. In the end, using a real-world dataset, we apply inverse optimization to estimate problem parameters and provide additional managerial insight into the platform’s revenue management strategy in a business case.  

Keywords: Revenue Management, Two-sided Market