I was recently searching for “sprint broadband” and “Mac OS X” and noticed a very interesting problem: the matches I got from both Google and Yahoo were incorrect because the matching pages included “Mac OS X” in the editorial and were also matching on “sprint broadband”, but ephemerally: they were advertising or sponsor information and no longer appear on the page.
There are a couple of reasons this happens, notably the inevitable delay between when a page is indexed in a search engine and when that cached or analyzed copy is updated due to changes on the site, but the bigger problem is that neither Google nor Yahoo can differentiate between the editorial on a Web page and the advertising.
In my book, that’s a problem, and a search on the matching phrase “connect at blazing speeds with the sprint mobile broadband card.” you’ll doubtless be surprised just how many matches there are!
At least Microsoft caught the problem (or just lucked out and didn’t see the Sprint advertising on the page): their Live Search did not show the same erroneous results from Infoworld.com as one of the top three matches.
As i say in my main blog post about this subject — Why can’t Google differentiate editorial from advertising? — the problem is just a matter of expectations more than anything. I rely on Google to offer up good matches to searches and this was a failure on their algorithm, one that’s repeated at Yahoo.
What do you think? Is it critical for search engines to differentiate between editorial and advertising when analyzing and scoring the content of a page?