Optimizing search engine revenue in sponsored search

ABSTRACT

Displaying sponsored ads alongside the search results is a key monetization strategy for search engine companies. Since users are more likely to click ads that are relevant to their query, it is crucial for search engine to deliver the right ads for the query and the order in which they are displayed. There are several works investigating on how to learn a ranking function to maximize the number of ad clicks. In this paper, we address a new revenue optimization

problem and aim to answer the question: how to construct a ranking model that can deliver high quality ads to the user as well as maximize search engine revenue? We introduce two novel methods from di fferent machine learning perspectives, and both of them take the revenue component into careful considerations. The algorithms are built upon the click-through log data with real ad clicks and impressions. The extensively experimental results verify the proposed

algorithm that can produce more revenue than other methods

as well as avoid losing relevance accuracy. To provide deep insight into the importance of each feature to search engine revenue, we extract twelve basic features from four categories. The experimental study provides a feature ranking list according to the revenue benefit of each feature.

REFERENCES

• 1: R. Baeza--Yates and B. Ribeiro--Neto. Modern Information
• 2: Retrieval. Addison Wesley, 1999.
• 3: A.Z. Broder, P. Ciccolo, M. Fontoura, E. Gabrilovich,
• 4: V. Josifovski, and L. Riedel. Search advertising using web
• 5: relevance feedback. In Proceeding of the 17th ACM
• 6: conference on Information and knowledge management
• 7: (CIKM--08), 2008.
• 8: C. Burges, T. Shaked, E. Renshaw, A. Lazier, M. Deeds,
• 9: N. Hamilton, and G. Hullernder. Learning to rank using
• 10: gradient descent. In Proceedings of the 22nd International
• 11: Conference on Machine Learning (ICML) (ICML--05), 2005.
• 12: D. Chakrabarti, D. Agarwal, and Vanja Josifovski.
• 13: Contextual advertising by combining relevance with click
• 14: feedback. In Proceeding of the 17th international conference
• 15: on World Wide Web (WWW--08), 2008.
• 16: M. Ciaramita, V. Murdock, and V. Plachouras. Online
• 17: learning from click data for sponsored search. In
• 18: Proceedings of the 17th international conference on World
• 19: Wide Web (WWW--08), 2008.
• 20: K. Dembczynski, W. Kotlowski, and D. Weiss. Predicting
• 21: ads' click--through rate with decision rules. In Workshop on
• 22: Targeting and Ranking in Online Advertising 2008, 2008.
• 23: J. Feng, H. Bhargava, and D. Pennock. Implementing
• 24: sponsored search in web search engines: Computational
• 25: evaluation of alternative mechanisms. INFORMS Journal on
• 26: Computing, 19:137---148, 2005.
• 27: Laura A. Granka , Thorsten Joachims , Geri Gay, Eye-tracking analysis of user behavior in WWW search, Proceedings of the 27th annual international ACM SIGIR conference on Research and development in information retrieval, July 25-29, 2004, Sheffield, United Kingdom [doi>10.1145/1008992.1009079]
• 28: R. Herbrich, T. Graepel, and K. Obermayer. Large margin
• 29: rank boundaries for ordinal regression. In Advances in Large
• 30: Margin Classifiers, 2000.
• 31: Thorsten Joachims, Optimizing search engines using clickthrough data, Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining, July 23-26, 2002, Edmonton, Alberta, Canada [doi>10.1145/775047.775067]
• 32: T. Joachims, L. Granka, B. Pang, H. Hembrooke, and
• 33: Thorsten Joachims , Laura Granka , Bing Pan , Helene Hembrooke , Geri Gay, Accurately interpreting clickthrough data as implicit feedback, Proceedings of the 28th annual international ACM SIGIR conference on Research and development in information retrieval, August 15-19, 2005, Salvador, Brazil [doi>10.1145/1076034.1076063]
• 34: Anísio Lacerda , Marco Cristo , Marcos André Gonçalves , Weiguo Fan , Nivio Ziviani , Berthier Ribeiro-Neto, Learning to advertise, Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval, August 06-11, 2006, Seattle, Washington, USA [doi>10.1145/1148170.1148265]
• 35: T. Liu, T. Qin, J. Xu, W. Xiong, and Hang Li. Letor:
• 36: Benchmark dataset for research on learning to rank for
• 37: information retrieval. In LR4IR 2007 in conjunction with
• 38: SIGIR 2007, 2007.
• 39: F. Radlinski, A. Broder, P. Ciccolo, E. Gabrilovich,
• 40: V. Josifovski, and Lance Riedel. Optimizing relevance and
• 41: revenue in ad search: a query substitution approach. In
• 42: Proceedings of the 31st annual international ACM SIGIR
• 43: conference on Research and development in information
• 44: retrieval (SIGIR--08), 2008.
• 45: M. Regelson and D.C. Fain. Predicting click--through rate using keyword clusters. In Proceedings of the 2nd Workshop on Sponsored Search Auctions, 2006.
• 46: M. Richardson, E. Dominowska, and R. Ragno. Predicting
• 47: clicks: estimating the click--through rate for new ads. In
• 48: Proceedings of the 16th international conference on World
• 49: Wide Web (WWW--07), 2007.
• 50: S. E. Robertson. Overview of the okapi projects. Journal of Documentation, 53:3---7, 1997.
• 51: V. N. Vapnik. Statistical Learning Theory. John Wiley &
• 52: Sons, New York, 1998.
• 53: C. Zhai and J. La erty. A study of smoothing methods for
• 54: language models applied to ad hoc information retrieval. In
• 55: Proceedings of the 24th annual international ACM SIGIR
• 56: conference on Research and development in information
• 57: retrieval (SIGIR--01), 2001.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning

General Terms:
Algorithms, Design, Economics, Experimentation

Keywords:
machine learning, ranking, revenue optimization, sponsored search