Management Science and Information Systems' Seminar

Topic:Online Allocation Rules in Display Advertising

Speaker:Yinyu Ye

Affiliation:Stanford University

Time:Friday, 12 Dec. 16:00pm

Location:Room K01 Guanghua Building 2

Abstract：Efficient allocation of impressions to advertisers in online display advertising has a significant impact on advertisers’ utility and the browsing experience of users. A suboptimal allocation, in one hand, degrades user experience by assigning unrelated ads to the user, and, on the other hand, it reduces effectiveness of the ad for the advertiser, i.e., loss of revenue. The problem becomes particularly challenging in the presence of advertisers with limited budgets. Practical constraints, such as frequency cap, delay in observing revenue, and unknown rate of arrival, add to the complex interaction among advertisers in the optimal impression assignment. In this talk, we develop online allocation rules of display advertising with budgeted advertisers. That is, upon arrival of each impression, cost and revenue vectors are revealed and the impression should be assigned to an advertiser almost immediately. Without any assumption on the distribution/arrival of impressions, we propose a framework to capture the risk of the ad network on each possible allocation, so that impressions can be allocated to advertisers such that the risk is minimized. In practice, this translates to starting with an initial estimate of “shadow” prices and updating them according to the belief of the ad network toward the future demand and remaining budgets. We apply our algorithms to a real data set and empirically show that Kullback-Leibler divergence risk measure has the best performance in terms of revenue and balanced budget delivery. Finally, we adapt the proposed algorithms to handle practical issues such as delay, frequency cap, and unknown number of impressions.

Your participation is warmly welcomed!