Models of malicious behavior in sponsored search

  • Authors:
  • Garud lyengar;David Phillips;Cliff Stein

  • Affiliations:
  • Columbia University, New York, NY;College of William & Mary, Williamsburg, VA;Columbia University, New York, NY

  • Venue:
  • SpringSim '07 Proceedings of the 2007 spring simulation multiconference - Volume 3
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Search engines such as Google, Yahoo, and MSN now auction off search terms to potential advertisers. The potential advertisers place their bids on each search term of interest, as well as specifying a daily budget. Each search on this term displays an advertisement that is linked to the advertiser's website, and the advertiser pays the search engine every time the link is activated. When an advertiser's budget is reached, the search engine stops displaying their ad. This kind of advertising is extremely popular -- the combined revenue of Yahoo and Google in 2005 was estimated at over 4.5 billion dollars. We develop small models which still have the property that malicious behavior such as bid-jamming still occurs as a rational best-response strategy. Such malicious behavior occurs frequently in practice. We are able to derive bidding strategies which are the best-responses when the budget of the bidder is low relative to her competitors, as well as strategies which protect against bid-jamming.