Evaluating Quality Score of New Web Advertisements

Abstract

Web advertisement is the one of the most profitable source of incomes for the search engine giants like Google, Bing, Yahoo and Baidu. According to Google, revenue obtained from web advertisement in 2012 is close to $43billion which is 20% more than that of 2011[1]. Top notch search engine uses Generalized Second Price auction to calculate rank of ad and the bid amount to be paid by user. Rank of ad tells about the position of ad to be displayed on search engine result page. Probability of being clicked of an ad is heavily relied on the position at which it is displayed in search engine page and position of ad involves determining click through rate of ad. It can be calculated empirically for ads that have been displaying repeatedly. But in case of new ad it is very complex and challenging task as historical information of ad is not present. By using bid amount solely ad rank can’t be calculated, as it will decrease the quality of ad, which is core aspect of web advertisement in search engine perspective. Quality of ad drastically affects the revenue of search engine which also results in un-satisfaction of user. In this thesis work we have designed a strategy to determine CTR of new ad by using various features of ad and landing page.