ABSTRACT
When placing online display ads, advertisers have an enormous set of sites to choose from. In this paper, we offer evidence that advertisers try many sites before settling upon a favored set, and explore the economic consequences of this behavior. Using data provided by a display ad network, we estimate a model of advertiser demand for web sites in which advertisers learn about the effectiveness of advertising at different sites. Preliminary results suggest that advertisers spend considerable resources advertising on sites they eventually abandon—in part because their prior beliefs about the efficacy of advertising are too high. A counterfactual scenario of better informing advertisers about their expected outcomes leads to higher advertiser welfare, but losses to sites and the ad network.

Santiago Balseiro, Carl F. Mela, and Jason M.T. Roos