Learning New Auction Format by Bidders in Internet Display Ad Auctions Shumpei Goke, Gabriel Y. Weintraub, Ralph Mastromonaco, and Sam Seljan * This version: October 19, 2021 Abstract We study actual bidding behavior when a new auction format gets introduced into the mar- ketplace. More specifically, we investigate this question using a novel data set on internet display ad auctions that exploits a staggered adoption by different publishers (sellers) of first- price auctions (FPAs), in place for the traditional second-price auctions (SPAs). Event study regression estimates indicate a significant jump, immediately after the auction format change, in revenue per sold impression (price) of the treated publishers relative to that of control pub- lishers, ranging from 35% to 75% of pre-treatment price levels of the treated group. Further, we observe that in later auction format changes the lift in price relative to SPAs dissipates over time, reminiscent of the celebrated revenue equivalence theorem. We take this as evidence of initially insufficient bid shading after the format change rather than an immediate shift to a new Bayesian Nash equilibrium. Prices then went down as bidders learned to shade their bids. We also show that bidders sophistication impacted their response to the auction format change. Our work constitutes one of the first field studies on bidders’ responses to auction for- mat changes, providing an important complement to theoretical model predictions. As such, it provides valuable information to auction designers when considering the implementation of different formats. * Goke: Department of Economics, Stanford University (shgoke@stanford.edu). Weintraub: Stanford GSB (gwein- tra@stanford.edu). Mastromonaco: Shopify (ralph.mastro@gmail.com). Seljan: Xandr (sam.seljan@xandr.com). The authors thank Liran Einav, Brad Larsen, and participants at INFORMS Annual Meeting, Manufacturing and Service Operations Management Conference, INFORMS Revenue Management and Pricing Section Conference, and Stan- ford IO Student Workshop for helpful discussions. The authors thank AppNexus/Xandr for sharing data. However, the opinions expressed in this paper belong to the authors and do not necessarily reflect Xandr’s or Shopify’s views. Goke thanks Stanford Graduate Fellowship (William R. Hewlett Fellowship) for financial support. 1 arXiv:2110.13814v1 [econ.GN] 20 Oct 2021