The practice of paid placement advertising—where advertisers pay a fee to appear alongside particular Web search results—is now one of the largest and fastest growing source of revenue for Web search engines. This article studies the allocation of placement slots. The payperclick standard for paid placement makes the problem challenging because it is not optimal to simply allocate slots to the highest bidders. We model several allocation strategies, including stylized versions of those used by Overture and Google, the two biggest brokers in paid placement, along with two alternative designs, and compare their performance via computational experiments. All mechanisms perform better when providers’ willingness to pay for paid placement is positively correlated with their true relevance. Ranking providers based on the product of clickthrough rate and bid price fares well broadly, while ranking purely by bid price performs well when the content providers’ relevance is positively correlated with their bid. Ranking by bid improves greatly with editorial filtering. Search engine’s placement revenue decreases when users’ attention is significantly lower for lowerranked listings, emphasizing the need to develop better user interfaces and control features. Due to the tradeoff between direct revenue increases and indirect revenue losses (due to consumer defection), the search engine must carefully choose the total number of paid slots. We also study how the rank allocations for each mechanism change over time as the search engine obtains more information about clickthroughs at each rank. We propose a rank revision strategy that weights clicks on lower ranked items more than clicks on higher ranked items. This method is shown to converge to the optimal (maximum revenue) ordering faster and more consistently than other methods.
[1]
P. Klemperer.
Auction Theory: A Guide to the Literature
,
1999
.
[2]
Barnes.
Discussion of the Paper
,
1961,
Public health papers and reports.
[3]
C. Lee Giles,et al.
Discovering Relevant Scientific Literature on the Web
,
2000,
IEEE Intell. Syst..
[4]
David Heckerman,et al.
Empirical Analysis of Predictive Algorithms for Collaborative Filtering
,
1998,
UAI.
[5]
Hemant K. Bhargava,et al.
Optimal Design of Contingency Pricing in IT-Intensive Commerce
,
2002,
ICIS.
[6]
D. Hoffman,et al.
How to acquire customers on the Web.
,
2000,
Harvard business review.
[7]
W. Campbell,et al.
THE UNIVERSITY OF TEXAS AT DALLAS
,
2004
.
[8]
Yu Jeffrey Hu.
Performance-Based Pricing Models in Online Advertising
,
2004
.
[9]
C. Lee Giles,et al.
Accessibility of information on the Web
,
2000,
INTL.