Motivated by Online Ad allocation with advertisers that pursue multiple objectives, we introduce and study a problem of Ad Allocation with Secondary Metrics. For instance, advertisers pay per click which is the primary metric the platforms optimize for, but require the average cost of a conversion - the secondary metric - to be below some threshold. This is an explicit option for Facebook advertisers. Further, even when this is not an explicit option and the advertisers can only configure pay per click campaigns, sales teams often negotiate with advertisers in terms of expected conversion thresholds and this becomes the implicit secondary metric for the ad platform or the sales teams. We study this problem under both market and advertiser perspectives. ; (Market Perspective) We adopt the per-impression auctioning approach used in the industry and propose modified sorting and allocation rules for the auction that explicitly take into account the secondary metric performance. We run the algorithm in an industrial setting on live traffic in a large ad network1. We find a significant impact on the realized secondary metrics without compromising primary metrics. For instance, the linear scoring function gives 30% lift on secondary metric and was selected as a default allocation algorithm in the ad network. ; (Advertiser Perspective) We present an efficient dynamic programming algorithm that calculates the best response strategy for each advertiser. We implement and test this solution on offline impression data and compute strategies for advertisers. We find that for a fraction of advertisers, we could not find non empty allocation. We cross check this result with the online results, and find that 92% of these ads did not meet their target conversions. This suggests that these ads may have unrealistic expectations.

Over the past decade, advertising has emerged as the primary source of revenue for many web sites and apps. In this paper we report a first-of-its-kind study that seeks to broadly understand the features, mechanisms and dynamics of display advertising on the web - i.e., the Adscape. Our study takes the perspective of users who are the targets of display ads shown on web sites. We develop a scalable crawling capability that enables us to gather the details of display ads including creatives and landing pages. Our crawling strategy is focused on maximizing the number of unique ads harvested. Of critical importance to our study is the recognition that a user's profile (i.e., browser profile and cookies) can have a significant impact on which ads are shown. We deploy our crawler over a variety of websites and profiles and this yields over 175K distinct display ads. We find that while targeting is widely used, there remain many instances in which delivered ads do not depend on user profile; further, ads vary more over user profiles than over websites. We also assess the population of advertisers seen and identify over 3.7K distinct entities from a variety of business segments. Finally, we find that when targeting is used, the specific types of ads delivered generally correspond with the details of user profiles, and also on users' patterns of visit.

Data centers run many services that impact millions of users daily. In reality, the latency of each service varies from one request to another. Existing tools allow to monitor services for performance glitches or service disruptions, but typically they do not help understanding the variations in latency.