Competition in Online Comparison Shopping Services

In our model, two profit-maximizing sellers sell a homogeneous good to Bayesian, riskneutral buyers in an online comparison shopping service. Buyers use a reputation system to update their beliefs about sellers. Buyers purchase from the seller that maximizes the buyer’s expected utility from the purchase. We find that the seller’s profit depends on the distribution of buyer beliefs. A degenerate distribution of beliefs implies either Bertrand competition or a monopolistic market. A non-degenerate distribution implies that both sellers can be profitable, if their reputations differ from each other. The seller with a higher reputation score receives a greater profit. If sellers are similar in every respect, the Bertrand equilibrium obtains. We test the theory with data from Pricegrabber using OLS and quantile regression. Controlling for different seller types, the evidence indicates that higher reputation scores may support price premiums.