Modeling and predicting user behavior in sponsored search

doi:10.1145/1557019.1557135

Paper

Modeling and predicting user behavior in sponsored search

Published Jun 28, 2009 · Josh Attenberg, Sandeep Pandey, Torsten Suel

UNKNOWN SJR score

64

Citations

4

Influential Citations

Full text

Semantic Scholar

Abstract

Implicit user feedback, including click-through and subsequent browsing behavior, is crucial for evaluating and improving the quality of results returned by search engines. Several recent studies [1, 2, 3, 13, 25] have used post-result browsing behavior including the sites visited, the number of clicks, and the dwell time on site in order to improve the ranking of search results. In this paper, we first study user behavior on sponsored search results (i.e., the advertisements displayed by search engines next to the organic results), and compare this behavior to that of organic results. Second, to exploit post-result user behavior for better ranking of sponsored results, we focus on identifying patterns in user behavior and predict expected on-site actions in future instances. In particular, we show how post-result behavior depends on various properties of the queries, advertisement, sites, and users, and build a classifier using properties such as these to predict certain aspects of the user behavior. Additionally, we develop a generative model to mimic trends in observed user activity using a mixture of pareto distributions. We conduct experiments based on billions of real navigation trails collected by a major search engine's browser toolbar.

Study Snapshot

Key takeawayPost-result user behavior can be used to improve sponsored search results by identifying patterns and predicting expected on-site actions.

PopulationOlder adults (50-71 years)

Sample size24

MethodsObservational

OutcomesBody Mass Index projections

ResultsSocial networks mitigate obesity in older groups.

Sign up to use Study Snapshot

Consensus is limited without an account. Create an account or sign in to get more searches and use the Study Snapshot.

Create an account

Paper

Modeling and predicting user behavior in sponsored search

Sign up to use Study Snapshot

References

How does clickthrough data reflect retrieval quality?

Paired comparison tests using clickthrough data provide more accurate and sensitive evaluation results than absolute usage metrics for retrieval quality in our domain.

Search advertising using web relevance feedback

Augmented queries and ads using knowledge-rich features improve the relevance of sponsored search ads and reduce the need for explicit bid phrases.

Computational advertising and recommender systems

Computational advertising aims to find the "best match" between users and advertisements in various contexts, involving massive optimization and search problems.

BrowseRank: letting web users vote for page importance

BrowseRank, a new method for computing page importance, outperforms traditional methods like PageRank and TrustRank in several tasks by using a user browsing graph and continuous-time Markov process.

Optimizing relevance and revenue in ad search: a query substitution approach

Our two-stage approach optimizes ad relevance and revenue in web search, resulting in faster and more accurate online matching while improving expected revenue.

Citations

Extracting Fast and Slow: User-Action Embedding with Inter-temporal Information

This study proposes a unified method that analyzes user actions with inter-temporal information, enabling successful interpretable analysis of human temporal behaviors in real-world data sets.

Kalman Filtering Attention for User Behavior Modeling in CTR Prediction

Kalman Filtering Attention (KFAtt) outperforms existing user behavior models in predicting click-through rates for e-commerce search engines by focusing on relevant behaviors and correcting biases towards frequent behaviors.

CausalFlow: Visual Analytics of Causality in Event Sequences

CausalFlow is a novel visualization approach that integrates automatic causal discovery methods into event sequence visualizations, enabling users to understand event occurrence and identify causal pathways in timeline-based event sequences.

Advertisement recommendation based on personal interests and ad push fairness

The proposed ad recommendation method, considering both ad-push fairness and personal interests, improves ad click-through rate and user satisfaction on news websites.

Personalized Attraction Enhanced Sponsored Search with Multi-task Learning

Personalized Attraction Enhanced Sponsored Search with Multi-task Learning can increase revenue and user experience for E-Commerce platforms, with a 2% increase in revenue per thousand impressions in Taobao.