Competitive Multi-Product Pricing with Demand Learning and Substitution Effects

Abstract

Many firms are selling different types of products. Typically sales applications are characterized by competitive settings, limited information and substitution effects. The demand intensities of single types of products are affected by the own products as well as the products of competitors. Due to the complexity of such markets, smart pricing strategies are hard to derive. We analyze stochastic dynamic multi-product pricing models under competition for the sale of durable goods. In a first step, a data-driven approach is used to measure substitution effects and to estimate sales probabilities in competitive markets. In a second step, we use a dynamic model to compute powerful heuristic feedback pricing strategies, which are even applicable if the number of competitors’ offers is large and their pricing strategies are unknown. Moreover, our approach allows taking additional features, such as customer ratings or shipping times into account. Adaptive estimations are used to update the estimation of sales probabilities and to further improve the strategy.