Why can’t Google differentiate editorial from adverising?

Just encountered an interesting and rather troubling situation as I dig around and try to find out more information about Sprint’s Broadband Mobile service and just how compatible it is with a Mac PowerBook. I did a Google search for sprint broadband “mac os x” and the #3 match is:

Google Search Results Error

Problem is, if you actually go to the referenced page there’s no mention of Sprint Broadband anywhere. Possibly the result was in its “top news” box but I’m guessing, from the words that appear in this search result, that it was a sponsor or other advertising block.
Hmmm…. this calls into doubt some of the esteem with which I hold Google’s search results.

After all, if Google can’t differentiate between an advert and the editorial on a Web page, then aren’t most of its results sullied or polluted, at least at some level?
And so, sure enough, a search for “connect at blazing speeds with the sprint mobile broadband card” produces an impressive 121,000 matches!
Sites with matching text include InfoWorld, Mobile Tech Today, CIO Today and Top Tech News. Now, I suspect that these are all IDG publications and Sprint purchased a run-of-site sponsorship advert that included the specific words I used, so that’s not really such a big deal unto itself.
What troubles me is the false hit I got on my original search. Because Google wasn’t able to differentiate between advertising / sponsorship text and editorial text, it couldn’t recognize that a match of editorial=mac and advertising=sprint was not, in fact, a good match for my search of mac + sprint.
Maybe I am expecting too much, but generally Google astonishes me with its smart search capabilities, so it is surprising to see such a gaffe.
By the way, it’s not limited to Google, either. The same search on Yahoo produces:

Yahoo Search Results Error

This is worse because it’s actually the #1 match. At least Google put it as match #3!
And, credit where it’s due, a search for sprint broadband “mac os x” on MSN Search Live Search (Who came up with this terrible name? Blech!) doesn’t have a single match for InfoWorld on the first pace but instead points to legit info sites like EVDOinfo.com and EVDOforums.com. Nicely done, Microsoft!
Thoughts? Reactions? Tempest in a teapot?

4 comments on “Why can’t Google differentiate editorial from adverising?

  1. You’re absolutely right! I’ve been running into this for years it seems, but it also seems to have gotten much worse within the last few months.
    That’s probably attributable to the increase in text ads recently (or maybe that’s just my perception). I feel it’s probably reality because of applications and browser plugins/extensions like AdBlock that have gotten rid of flash and image ads for many users.
    Google really needs to do something about this. I can think of at least a half dozen searches I’ve done in the last month alone that have been littered with useless results because of text in widespread advertisements similar to the one you mentioned.
    I’m sure it’s probably something they’re working on already, but it’s not an easy problem to solve with high accuracy. How do you tell what’s editorial content and what is advertising? I can think of a number of ways, none of which would ever be 100% accurate.

  2. Great article.
    We run EVDOinfo.com and EVDOforums.com and sometimes we wonder the same thing.
    One issue, that you didn’t hit on, is SPAM. Not email spam, but those that perform Search Engine Spam, ways to “trick” Google.

  3. I imagine that it was just this problem that led Google to intoduce it’s and end markers. However the markers are just for their ads, and not a guide to GoogleBot when indexing the page.
    Should a marker be introduced for GoogleBot (and the rest) it will have a short term benefit because it relies on the webmaster to a) know about the marker; b) be able to integrate it onto the page; and c) to be happy to clutter every page with markers for each of the important search engines.
    Lets hope W3C come up with a standard tag to indicate the boundry between content and navigation (page structure).
    It’ll also blow all those duplicate content sites out of the water – after all they rely on the page structure to keep the articles safe.

Leave a Reply

Your email address will not be published. Required fields are marked *