Research Of Ranking Rules In Keyword Auction

As the rapid growing and application of search engine technology, keywordadvertisement has become the principal revenue source of search engine companies such asGoogle, Baidu, etc. Keyword auction is the sale mechanism of keyword advertisement,which is different from the traditional sale mechanism CPM (Cost Per Mile). Keywordauction is the extension of classic auction mechanisms, and its mechanism has changed fromGFP (Generalized First Pricing) to GSP (Generalized Second Pricing). As truth telling is notthe dominant strategy in GSP, many researchers are interested in the research of advertisers’bidding strategy, equilibrium analysis and mechanism design in keyword auction.Based on the former researches, I focus on the research of ranking rules in keywordauction. From the angle of search engine companies, I use Monte-Carlo simulation tocompare the search engine revenue, user experience and efficiency in different ranking rulesbased on the distributions derived from the empirical data. I also take the long-terminfluence of ranking rules into consideration and classify the simulation into two categories,namely short-term and long-term. In the short-term situation or long-term situation with onlythe influence on advertiser number, search engine revenue arrives its maximum value whenranking according to bid. And in the long-term situation with only the influence on naturalCTR (Click Through Rate) or both the influence on advertiser number and natural CTR,search engine revenue increases as the weight of quality score increases. We can concludefrom the simulation results that the way advertiser number influences on search enginerevenue is different from that of natural CTR, and my research outcome can give searchengine companies some instructions when they make decisions about ranking rules.