Measurement and analysis of OSN ad auctions

Advertising is ubiquitous on the Web; numerous ad networks serve billions of ads daily via keyword or search term auctions. Recently, online social networks (OSNs) such as Facebook have created site-specific ad services that differ from traditional ad networks by letting advertisers bid on users rather than keywords. With Facebook's annual ad revenue exceeding $4 billion, OSN-based ad services are emerging to be a significant fraction of the online ad market. In contrast to other online ad markets (e.g., Google's ad market), there has been little academic study of OSN ad services, and OSNs have released very little data about their advertising markets; as a result, researchers currently lack the tools to measure and understand these markets. In this paper, our goal is to bring visibility to OSN ad markets, focusing on Facebook. We demonstrate that the (undocumented) feature that suggests bids to advertisers is most likely calculated via sampling recent winning bids. We then show how this feature can be used to explore the relative value of different user demographics and the overall stability of the advertising market. Through the exploration of suggested bid data for different demographics, we find dramatic differences in prices paid across different user interests and locations. Finally, we show that the ad market shows long-term variability, suggesting that OSN ad services have yet to mature.