Assortment games with effort in online marketplaces

Abstract

This thesis studies an assortment game of duopolistic retailers competing in as sortment and advertising effort decisions in an online marketplace. We suggest ways to maximize players’ payoffs and search for an answer to the question if the marketplace can provide a value-added assortment and advertising service to its sellers. In this re gard, we consider two different systems a) when retailers have overlapping assortments b) when retailers have access to an exclusive product. We experiment each case under a centralized regime, in which retailers are centrally managed by the marketplace, and a decentralized regime, in which retailers compete in assortment and ad effort decisions. We assume that prices are determined by the marketplace, ad opportunity is given only to one retailer and applies to all products in the retailer’s offered assortment, customers are rational utility maximizers and their choice process is governed by the multinomial Logit framework. We augment the utility function of customers by an advertising effort mechanism. In our case, products’ displayed positions at the marketplace are linked to retailers’ advertising effort. We first characterize this problem as a non-linear mathematical optimization problem for a single retailer. Then using a computational framework, we show that the advertising offering causes marketplace to increase profits in both decentralized and centralized settings. A retailer has a substantial advantage if he wins the advertising opportunity. Although ad effort makes the other retailer out of the game, our results show that centralized system always protects the retailer, who otherwise loses much more in the decentralized system. In addition, offering an attractive exclusive product gives a competitive power to the retailer, who lost the ad effort game. We also observe that in all cases, competition causes system to use more ad effort. We observe that competition always causes