It seems obvious that the major web search engines have solved the problem of finding exact matches to user queries and separating the wheat from the chaff. Search for Twitter ( Google, Yahoo, Bing ), iPhone ( Google, Yahoo, Bing ) or even Attivio ( Google, Yahoo, Bing ) and you'll see largely the same results and variations - all of good quality. (Both Yahoo! and Bing blend results for the latter query with incorrect spelling suggestions, but one presumes they may be doing that deliberately to differentiate themselves, even if in this case it makes the results considerably worse.)

But what happens when the user is vague and doesn't know exactly what they are looking for? Or is interested in a topic that is more ambiguous, statistically speaking? For example if they are trying to find out about the "one-electron universe" hypothesis, but mistakenly remember it as the "one electron theory"...

Click any of the thumbnails below to view the respective query results (new window)

Google	Yahoo	Bing
Result #1 is a relevant discussion, #2-3 are relevant scientific abstracts. Result #4 is excellent - the Wikipedia article on one-electron universe. This is the best, highest-ranking result overall for this query. Notice the term 'theory' does not appear on the result page. If you change the query to one-electron universe, the Wikipedia article is #1 and results 2+ are excellent. (Ironically, #7 is a Yahoo! Answers page).	All results are general information about electron theory and not particularly relevant. If you change the query to one-electron universe the Wikipedia article is returned as #1 and a few other good hits appear amongst general articles. Overall page design is virtually identical to Google.	Results contain general information about electrons and theories except #6-7 which matches the query but is not relevant. Overall the least relevant results. If you change the query to one-electron universe, the Wikipedia article is returned as #1 but results 2-9 remain overly general and not relevant. Result #10 is good. The overall page design is slightly flashier, but there is also more wasted space, and having to mouse-over the results to see a thumbnail denies the user the opportunity to visually differentiate results. The front-page of Bing is cool however.

Google is clearly better on this query and, in my view, on this general class of queries. Their use of a (presumably large) Language Model, along with use of the relationships between text and pages, simply gives tremendous advantage.

Yahoo does a good job of keeping the results clean, even if they are not directly relevant, and does visibly better on the more exact query. Bing has visual appeal but has more work to do, at least on this type of query.

Overall an interesting example of the subtleties of search. Feel free to This e-mail address is being protected from spambots. You need JavaScript enabled to view it your favorite web queries and analysis, or post a comment.

TrackBack URI for this entry

Comments (0)