Personalized Ad Delivery When Ads Fatigue: An Approximation Algorithm
We consider a crucial aspect of displaying advertisements on the internet: the individual user. In particular, we consider ad fatigue, where a user tires of an advertisement as it is seen more often. We would like to show advertisements such that, given t
- PDF / 350,735 Bytes
- 6 Pages / 430 x 660 pts Page_size
- 9 Downloads / 178 Views
2
Yahoo!, Inc., 2821 Mission College Blvd., Santa Clara, CA, USA Yahoo! Research, 2821 Mission College Blvd., Santa Clara, CA, USA
Abstract. We consider a crucial aspect of displaying advertisements on the internet: the individual user. In particular, we consider ad fatigue, where a user tires of an advertisement as it is seen more often. We would like to show advertisements such that, given the impact of ad fatigue, the overall efficiency of the system is optimized. We design an approximation algorithm, for the case that we study, that approaches the optimum as the number of unique ads shown, if there is only one available position, increases.
1
Introduction
Internet advertising is a booming business, and already provides a large portion of search engine revenue. Placing ads strategically is key to optimizing the efficiency of these advertising systems. The current measures used in most academic literature to determine the match between advertisement and user is the estimated clickthrough rate, based on the advertisement, the keyword, and the position on the web page [1,11,5,10,2,6,7]. In practice, landing page information [12] and demographic targeting [9] have been used to better match advertisements to users. In this paper, we explore another crucial aspect in determining when and where to place ads: the individual user. There are many ways that an individual user’s experience may influence how an advertisement is received, including previous positive or negative experiences with advertisements and previous exposure to a company name or logo. We concentrate in this study on how previous experiences viewing an advertisement influence a user’s likelihood to click on that advertisement. The setting used is one where ads are embedded in webpages, such as AdSense at Google or Content Match at Yahoo!. An individual user may view the particular site several times in a single day, if the site is a user’s homepage or a frequently visited resource, and ads are less likely to be clicked as they are shown more often. This phenomena is referred to as ad fatigue [4], since the user tires of the ad after viewing it several times. We study the problem of determining which ads to display where and when in order to maximize efficiency over several displays of the same page. Efficiency is considered to be the expected number of clicks times the value of that click to the advertiser. We design an algorithm that achieves close to optimum efficiency. X. Deng and F.C. Graham (Eds.): WINE 2007, LNCS 4858, pp. 535–540, 2007. c Springer-Verlag Berlin Heidelberg 2007
536
2
Z. Abrams and E. Vee
Model
Our model follows that of [8]. We assume the input to the algorithm is: – The max number of times, T , a particular web page will be viewed by a single user during the course of some time period. – A vector f of fatigue rates that contains T values. We use ft to denote the value corresponding to the tth element of the vector f , and ∀t, 0 ≤ ft ≤ 1. – A vector CT R representing the decay in expected clicks due to position on the page, sorted f
Data Loading...
