Estimating advertisability of tail queries for sponsored search

  • Authors:

  • Sandeep Pandey;Kunal Punera;Marcus Fontoura;Vanja Josifovski

  • Affiliations:

  • Yahoo! Research, Sunnyvale, CA, USA;Yahoo! Research, Sunnyvale, CA, USA;Yahoo! Research, Sunnyvale, CA, USA;Yahoo! Research, Sunnyvale, CA, USA

  • Venue:
  • Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Sponsored search is one of the major sources of revenue for search engines on the World Wide Web. It has been observed that while showing ads for every query maximizes short-term revenue, irrelevant ads lead to poor user experience and less revenue in the long-term. Hence, it is in search engines' interest to place ads only for queries that are likely to attract ad-clicks. Many algorithms for estimating query advertisability exist in literature, but most of these methods have been proposed for and tested on the frequent or "head" queries. Since query frequencies on search engine are known to be distributed as a power-law, this leaves a huge fraction of the queries uncovered. In this paper we focus on the more challenging problem of estimating query advertisability for infrequent or "tail" queries. These require fundamentally different methods than head queries: for e.g., tail queries are almost all unique and require the estimation method to be online and inexpensive. We show that previously proposed methods do not apply to tail queries, and when modified for our scenario they do not work well. Further, we give a simple, yet effective, approach, which estimates query advertisability using only the words present in the queries. We evaluate our approach on a real-world dataset consisting of search engine queries and user clicks. Our results show that our simple approach outperforms a more complex one based on regularized regression.