Large scale query log analysis of re-finding

Abstract

Although Web search engines are targeted towards helping people find new information, people regularly use them to re-find Web pages they have seen before. Researchers have noted the existence of this phenomenon, but relatively little is understood about how re-finding behavior differs from the finding of new information. This paper dives deeply into the differences via analysis of three large-scale data sources: 1) query logs (queries, clicks, result impressions), 2) Web browsing logs (URL visits), and 3) a daily Web crawl (page content). It appears that people learn valuable information about the pages they find that helps them re-find what they are looking for later; compared to the initial finding query, re-finding queries are typically shorter, and rank the re-found URL higher. While many instances of refinding probably serve as a type of bookmark for a known URL, others seem to represent the resumption of a previous task; results clicked at the end of a session are more likely than those at the beginning to be re-found during a later session, while re-finding is more likely to happen at the beginning of a session than at the end. Additionally, we observe differences in cross-session and intra-session re-finding that may indicate different types of refinding tasks. Our findings suggest there is a rich opportunity for search engines to take advantage of re-finding behavior as a means to improve the search experience.

Exploring Query Patterns in Email Search

“…The mean character length of queries was 5.87 and many were only partial words. Although in line with our previous analyses [6], this is much shorter than reported elsewhere in the literature with 1.6 words being reported for desktop search [4] and 12.1 characters for web page re-finding [16]. It is also shorter than the 1.4 words reported for lab-based studies of email re-finding [7].…”

“…More recent work has looked at querying behaviour over significantly longer time periods; examining temporal aspects such as query repetition and how pages are re-found [13,15,16,3]. These analyses reveal that query re-use behaviour is extremely common.…”

“…For example 33% of all SE queries submitted have been identically submitted previously by the same user and for 39% of all queries the user returns to a Web page that he has found before via a separate search [15]. SE log analysis has shown that queries submitted with the intention of re-finding are typically shorter than those for new content and the clicked on pages for such queries tend to rank higher in the results list [16].…”

“…the number of messages clicked per query is much larger. Email query chains had on average 5.17 message clicks compared to the 1 or less page clicks associated with SE queries [16]. Secondly, email collections are of course much smaller than the web, which naturally increases the chance of serendipitous overlap.…”

Workload Modeling for Computer Systems Performance Evaluation

“…• Users occasionally search for the same thing again and again over extended periods of time [680,703], possibly using search as an alternative to keeping a bookmark.…”