Today, predictive digital ad targeting typically relies on detailed user profiles. As consumers deserve and expect more internet privacy, we aim to develop methods to effectively target advertising in a way that respects consumers’ wishes. In particular, we have developed ID-Free targeting, a method to target advertising based on predicted user behavior, without reference to any profile, web browsing history, or any other identifier of the user being targeted. With the restriction that historical user behavior is not available, the only accessible view of the user is the information about the digital ad opportunity itself, such as the URL and the time of day. In this work, we focus on the problem of making brand-specific targeting decisions based on the URL where the ad will be shown, that is, predicting which ad inventories are most likely to lead to a conversion for a particular brand. In order to address this problem, we learn rich URL embedding features for ad inventory URLs. We share the design of our online learning algorithm, which is a practical implementation for training the URL embeddings with the large, dynamic "corpus" that is the internet. We show that this approach to feature representation gives clear performance improvements over a naive representation, and improvements are even more pronounced when brand-specific training data are scarce. The resulting model enables a fully ID-Free delivery of precise ad targeting to users with completely unknown identity and historical behavior.

A growing proportion of digital advertising slots is purchased through real time bidding auctions, which enables advertisers to impose highly specific criteria on which devices and opportunities to target. Employing sophisticated targeting criteria reliably increases the performance of an ad campaign, however too strict criteria will limit its scale. This raises the need to estimate the number of anticipated ad impressions at a given campaign performance level, thus enabling advertisers to tune the campaign's budget to an optimal performance-scale trade-off. In this paper, we provide a way to estimate campaign impressions given the campaign criteria. There are several challenges to this problem. First, the criteria contain logic to include and exclude combinations of audience segments, making the space of possible criteria exponentially large. Furthermore, it is difficult to validate predictions, because we wish to predict the number of impressions available without budget constraints, a situation we can rarely observe in practice. In our approach, we first treat the audience segment inclusion/exclusion criteria separately as a data compression problem, where we use MinHash "sketches" to estimate audience size. We then model the number of available impressions with a regularized linear regression in log space, using multiplier features motivated by the assumption that some components of the additional campaign criteria are conditionally independent. We construct a validation set by projecting observed RTB data (under real budget constraints) to get impression availability without budget constraints. Using this approach, our average prediction is a factor of 2.2 from the true impression availability, and the deployed product responds to user requests in well under a second, meeting both accuracy and latency requirements for decision making in the execution of advertising campaigns.

Internet display advertising is a critical revenue source for publishers and online content providers, and is supported by massive amounts of user and publisher data. Targeting display ads can be improved substantially with machine learning methods, but building many models on massive data becomes prohibitively expensive computationally. This paper presents a combination of strategies, deployed by the online advertising firm Dstillery, for learning many models from extremely high-dimensional data efficiently and without human intervention. This combination includes: (i)~A method for simple-yet-effective transfer learning where a model learned from data that is relatively abundant and cheap is taken as a prior for Bayesian logistic regression trained with stochastic gradient descent (SGD) from the more expensive target data. (ii)~A new update rule for automatic learning rate adaptation, to support learning from sparse, high-dimensional data, as well as the integration with adaptive regularization. We present an experimental analysis across 100 different ad campaigns, showing that the transfer learning indeed improves performance across a large number of them, especially at the start of the campaigns. The combined "hands-free" method needs no fiddling with the SGD learning rate, and we show that it is just as effective as using expensive grid search to set the regularization parameter for each campaign.