Personalized pricing, competition and welfare

Data-driven AI pricing algorithms in on-line markets collect consumer information and use it in their pricing technologies. In the simplest symmetric Hotellingís model such technologies reduce prices and proÖts. We extend Hotellingís model with vertically di§erentiated products, cost asymmetries and arbitrary adjustment costs. We provide a characterization of competition in personalized pricing: Sellers compete in o§ering consumer surplus, personalized prices are constrained monopoly prices and social welfare is maximal. For linear adjustment costs, adopting personalized pricing technology is a dominant strategy for both sellers. We derive conditions under which the most effi cient seller increases her proÖt through personalized pricing. While aggregate consumer surplus increases, consumers with high switching costs may be hurt. Finally, we discuss several extensions of our approach such as oligopoly.