DP11223 Incentive Compatible Advertising on a Social Network
A platform that operates a social network allows firms to post display ads to network members. Each member is interested in exactly one type of product. The network structure is correlated with the profile of members' privately known preferences over product types. The platform's policy consists of a display rule (which specifies the stationary probability with which each product is shown to each network member, as a function of the network structure) and an advertising fee (which the platform charges from firms as a function of their reported type). We provide conditions for the existence of an incentive-compatible policy that maximizes and fully extracts firms' surplus. This objective is easier to attain when the network is more informative of members' preferences, consumers are more attentive to advertising and their frequency of repeated purchases is higher, and advertisers are less informed of the network structure. We provide a more detailed characterization when the network is generated according to the "stochastic block model", thus linking our model to the "community detection" problem in Network Science.
