Replication Code for Using Grouped Data to Estimate Revenue Heterogeneity in Online Advertising Markets

In this paper, we showed that a publisher of online ads could estimate informative models, such as revenue models, from a position of limited visibility into the auction process conducted by intermediary ad platforms, even though these platforms provide only grouped data to the publisher. We validate our results using singleton data available to the publishers on auction data from unique customers in specific location-times. We also address the high-dimensional nature of the data and allow for time-varying individual heterogeneity through interactive effects. Our empirical results show that ad networks impact revenue heterogeneity.