Cost-sensitive Learning for Utility Optimization in Online Advertising Auctions

TL;DR

This paper introduces a cost-sensitive learning approach for online advertising auctions that weights prediction errors by business value, improving profit-related metrics.

Contribution

It proposes a novel cost weighting scheme for log loss in conversion modeling, bridging the gap between model training and business-aware evaluation.

Findings

Significant offline performance improvements

Enhanced online bidding efficiency

Better alignment with advertiser profit metrics

Abstract

One of the most challenging problems in computational advertising is the prediction of click-through and conversion rates for bidding in online advertising auctions. An unaddressed problem in previous approaches is the existence of highly non-uniform misprediction costs. While for model evaluation these costs have been taken into account through recently proposed business-aware offline metrics -- such as the Utility metric which measures the impact on advertiser profit -- this is not the case when training the models themselves. In this paper, to bridge the gap, we formally analyze the relationship between optimizing the Utility metric and the log loss, which is considered as one of the state-of-the-art approaches in conversion modeling. Our analysis motivates the idea of weighting the log loss with the business value of the predicted outcome. We present and analyze a new cost weighting scheme and show that significant gains in offline and online performance can be achieved.