We study actual bidding behavior when a new auction format gets introduced into the marketplace. More specifically, we investigate this question using a novel dataset on internet display advertising auctions that exploits a staggered adoption by different publishers (sellers) of first-price auctions (FPAs), instead of the traditional second-price auctions (SPAs). Event study regression estimates indicate that, immediately after the auction format change, the revenue per sold impression (price) jumped considerably for the treated publishers relative to the control publishers, ranging from 35% to 75% of the pre-treatment price level of the treatment group. Further, we observe that in later auction format changes the increase in the price levels under FPAs relative to price levels under SPAs dissipates over time, reminiscent of the celebrated revenue equivalence theorem. A possible interpretation of these facts is initially insufficient bid shading after the format change rather than an immediate shift to a new Bayesian Nash equilibrium. The gradual decrease in prices can be interpreted as the result of bidders' learning to shade their bids. We also present suggestive evidence that bidders' sophistication may have impacted their response to the auction format change. Our work constitutes one of the first field studies on bidders' responses to auction format 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.

In online advertising, an important quality control step is to audit advertising images ("creatives") before they appear on publishers' webpages. This ensures that advertisements only appear on webpages where the ad is appropriate. Assigning the correct sensitive categories to each creative - such as alcohol, tobacco, etc. - is one of the most important aspects to get correct. If a sensitive creative is displayed on the wrong webpage, it can ruin the user's experience, the publisher's reputation, and may have legal implications. To protect against this, humans audit every creative before it is displayed through our ad exchange; this process is costly and time consuming. This paper explains how we automated sensitive category detection. To detect whether a creative has any sensitive content, we use a pre-trained deep convolutional neural network (Xception [1]) to process the creative image and merge this with the historical distribution of sensitive categories associated with the creative's landing page (the webpage that loads when the ad is clicked, which may also contain sensitive content). This representation is then passed into a series of fully connected layers to make a prediction of whether a creative belongs to a sensitive category. We show in offline testing that this model achieves slightly better than human performance (model accuracy 99.92%; human accuracy 99.88%) on a large fraction of creatives (61%) while making 3.5 times fewer mistakes in certain categories for which mistakes are especially costly. These results changed somewhat when deploying this model at scale in production, where a small modification resulted in classifying fewer creatives than estimated offline, with approximately the same accuracy (52% classified with 99.87% accuracy).

Online advertising is a multi-billion dollar industry largely responsible for keeping most online content free and content creators ("publishers") in business. In one aspect of advertising sales, impressions are auctioned off in second price auctions on an auction-by-auction basis through what is known as real-time bidding (RTB). An important mechanism through which publishers can influence how much revenue they earn is reserve pricing in RTB auctions. The optimal reserve price problem is well studied in both applied and academic literatures. However, few solutions are suited to RTB, where billions of auctions for ad space on millions of different sites and Internet users are conducted each day among bidders with heterogenous valuations. In particular, existing solutions are not robust to violations of assumptions common in auction theory and do not scale to processing terabytes of data each hour, a high dimensional feature space, and a fast changing demand landscape. In this paper, we describe a scalable, online, real-time, incrementally updated reserve price optimizer for RTB that is currently implemented as part of the AppNexus Publisher Suite. Our solution applies an online learning approach, maximizing a custom cost function suited to reserve price optimization. We demonstrate the scalability and feasibility with the results from the reserve price optimizer deployed in a production environment. In the production deployed optimizer, the average revenue lift was 34.4% with 95% confidence intervals (33.2%, 35.6%) from more than 8 billion auctions over 46 days, a substantial increase over non-optimized and often manually set rule based reserve prices.

In this paper, we report the first empirical study and live test of the reserve price optimisation problem in the context of Real-Time Bidding (RTB) display advertising from an operational environment. A reserve price is the minimum that the auctioneer would accept from bidders in auctions, and in a second price auction it could potentially uplift the auctioneer's revenue by charging winners the reserve price instead of the second highest bids. As such it has been used for sponsored search and been well studied in that context. However, comparing with sponsored search and contextual advertising, this problem in the RTB context is less understood yet more critical for publishers because 1) bidders have to submit a bid for each individual impression, which mostly is associated with user data that is subject to change over time. This, coupled with practical constraints such as the budget, campaigns' life time, etc. makes the theoretical result from optimal auction theory not necessarily applicable and a further empirical study is required to confirm its optimality from the real-world system; 2) in RTB an advertiser is facing nearly unlimited supply and the auction is almost done in "last second", which encourages spending less on the high cost ad placements. This could imply the loss of bid volume over time if a correct reserve price is not in place. In this paper we empirically examine several commonly adopted algorithms for setting up a reserve price. We report our results of a large scale online experiment in a production platform. The results suggest the our proposed game theory based OneShot algorithm performed the best and the superiority is significant in most cases.