Monday, 16 February 2009

search interaction is short

*warning - read the comments below before you read the article discussed here*

I came across an interesting article which is, to some extent, both a challenge for interactive information retrieval, and a blow to idea that search should be like a conversation (rather than guessing a searcher's intentions). One of their notable findings is that the average search session is 2.9 interactions long. Nice to see that its not considering search session length in terms of time (a common metric, but not always applicable during information seeking), but instead in the number of interactions. This is something in the vein of my own research.

This finding really only allows 1) an initial search, 2) an interactive refinement and/or a scroll, and 3) a selection. This also assumes, since the 2.9 is less than 3, that one of these is optional. and its unlikely to be the searching or the selecting. I want to go over the paper in some more detail, but its certainly interesting.


Daniel Tunkelang said...

Access-deprived readers like me might appreciate this pointer to the full article.

I found some very intriguing statements in the paper:

"Most Web queries belong to the informational level, and transactional and navigational information are a relative small
proportion within all the queries."

"This result shows that people generally just consider the information shown on the first SERP and usually spend some
time to browse many of links listed on the first SERP."

"approximately 40% of the time, searchers entered queries but did not click on any links in the SERP."

Regarding the first statement, I wonder if a search that leads to a Wikipedia page is considered "informational". I think those should be considered navigational, or at least that they should be distinguished from informational queries that do not map neatly to a canonical page on the web.

Regarding the last two, I think it's significant that 40% of searchers don't find what they're looking for on a given search. That's pretty high, no? And in my mind it argues strongly for better support interactive IR.

Finally, maybe I misunderstand the terms, but what does the average query length of 2.9 have to do with the number of interactions? It's the number of terms contained in the average query. Wouldn't you be more interested in session length?

Max L. Wilson said...

It is, as you say, a challenge for us all. 40% sucks. and people aren't giving it the time to interactions. Often a criticism of more complicated search interfaces is that people want it clear and simple. Even with only the basic clear and simple interactive query refinement on the major search engines, the quote in the abstract says that people aren't doing it very much.

As for the 2.9, though, You are right, despite the abstract suggesting there are on average 2.9 interactions a session, the usage of 2.9 within the paper is in terms of words per query. They do, however, discuss interactions in the paper, but very briefly.

A typo, or missing fact from within the paper. im not sure sadly. it would be interesting to know what the average number of interactions of a web search engine is! darn it.

GeneG said...

I read the article after I saw your blog post. Sigh. I hope people don't pay good money for it.

Here's what I wrote

Max L. Wilson said...

i've added a warning to the blog post (rather than remove it) to read the comments, since it ends up being disappointing.