Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We study users' responses to sponsored-search advertising using consumer-level data from Microsoft Live. We document that users click ads in a nonsequential order and that the click through rates depend on the identity of competing ads. We estimate a dynamic model of utility-maximizing users that rationalizes these two facts and find that 51 percent more clicks would occur if ads faced no competition. We demonstrate that optimal matching of advertisements to positions raises welfare by 27 percent, and that individual-level targeting raises welfare by 69 percent. Revealing the quality of the advertiser prior to clicking on a sponsored link raises welfare by 1.6 percent. (JEL D12, L86, M37)