Fiddling with Results?

SEW points out that a search on the new MSFT beta engine for more evil than satan returns Google as #1, and explains this history – apparently Google did the same to MSFT back in 1999. I find this disturbing in all sorts of ways. It shows that engineers at…

evilsatanSEW points out that a search on the new MSFT beta engine for more evil than satan returns Google as #1, and explains this history – apparently Google did the same to MSFT back in 1999.

I find this disturbing in all sorts of ways. It shows that engineers at any search company not only *can* fiddle with results, but at times when they find it convenient or funny (or perhaps in other instances, good for their business in some way) they *will.* Not a good precedent, in my humble opinion.

UPDATE: Google called me to inform me that they did not fiddle back in 1999, that the “evil” query back then returned MSFT on its own…not surprisingly…also MSFT posts on its blog that the initial results that brought Google to #1 for “more evil than satan” were a MSFTbomb, ie, not an inside job. Well, at least, not an official inside job…

14 thoughts on “Fiddling with Results?”

  1. I know Google didn’t fiddle back then, but I’m almost certain MSN did this time. The Google incident happened because people were linking to Microsoft using the word “evil”. But even if people were doing this to Google, MSN wouldn’t list them first because their engine depends on keyword density far more than link text…

  2. Not that I’m complaining… but… recently I wrote one of the first reviews for Microsoft’s new much heralded Media Center Edition 2005 software. For much of the last few months my review, which was mostly critical of Microsoft’s decision not to include HDTV capability in their new release as well as the slow as molasses speed of large mp3 libraries, has been in the top three search results for Google under the search “MCE 2005”. Most of the time it’s been the number one entry. At present it is number five for the search MCE 2005 at Google.

    At the new MSN beta on the other hand, the article does not appear to be listed in the top 500 search results. MSN search does list a much more positive piece that I wrote on MCE 2005 which basically just aggregated all of the media coverage after the launch, but the review, which I would assume Microsoft would see as critical of a product that they are reportedly spending $20 billion on is nowhere to be found on the MSN Search Site.

    In fact even with a “Thomas Hawk” MCE 2005 search my review of their product does not show up (although a link to the review from PVRblog to my article does show up as the 16th item with the above search.

    Of course a “Thomas Hawk” MCE 2005 over at Google turns up not only the review but countless other prominent sites that linked the review (Slashdot, Engadget, PVRblog, ehome Upgrade, etc.).

    Now I assume that because Slashdot, Engadget, et. al linked the review in the first place I get high placement on a Google search for “MCE 2005”. I’m not so much of a conspiracy theorist to yet assume that Microsoft would remove a review that they felt was critical of one of their important products. I mean come on. Who am I? How eggotistical do I have to be think they really care about my puny little review. On the other hand you never know… but then again… I suppose Google could also be as guilty of placing a review negative to a Microsoft product higher than it might be otherwise.

    Just some initial observations on the technology.

  3. I guess that the habit of linking to Google’s search results using the keywords as anchor text is the reason for that, not deliberate fiddling.

  4. Yes, didn’t mean to suggest that Google actually rigged the results way back then. In fact, their explanation at that time was the same as they do today when various link bombs go off — the search is “reflecting the web.” It’s also not clear if MSN has actually rigged its results or not.

  5. It’s funny that downloadfirefox dot com and google dot com don’t seem to fit in with all the other “evil” related search results on msft’s new engine. I put together a couple other interesting queries with screenshots for your viewing pleasure. Obviously (hopefully it’s obvious) I took a sarcastic approach on my blog but it’s all in fun. I mean they’re in beta right now anyway, right?

  6. It is somewhat disturbing — you wonder when these guys will grow up. But certainly one is not surprised when these companies engage in forms of hyper-competitive behavior. Just try getting a GMail invite through to someone’s Yahoo Mail address. What? It *somehow* got filtered? πŸ™‚

  7. BTW, just tried the same query but using MSN’s personalization options as follows:

    more evil than satan {frsh=16} {popl=100} {mtch=0}

    The Google home page is nowhere to be found, likely indicating that the result showing Google at #1 is indeed hand-tweaked.

  8. No, it’s not hand-tweaked. I’ve been doing some searches that return my site when searching on Google.

    A search for “canadian voip” returns Google as the #1 hit. I know that I and others have linked to Google searches, especially with title tags of “Google Search for Canadian VoIP”, so that may help explain it.

    There are a handful of other searches where Google is either the #1 hit or somewhere in the top 10. Ouch…

  9. Tweak the results ranking a bit further:

    more evil than satan {popl=100} {mtch=50} {frsh=50}

    and Search Engine Watch is up to number 2 and Microsoft to number 3. Earlier on today, Microsoft was number 1 for these settings!

    The SEW page has its cache dated today, so no two day lag between crawling and updated the index – not for some pages anyway!

  10. Aw, come on – let the search guys have their moments of fun. πŸ™‚

    I’m sure that Google, MSN, and Yahoo! staff are like pro-wrestlers – in public, there’s posturing a-plenty, mean snarling, and direct competition; in private, all sipping Martini’s together and discussing the latest football scores. πŸ˜‰

Leave a Reply

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