Wikipedia I’m in Ur Index Rulin’ Ur Serpz
July 5th, 2007 by Michael Gray in Case StudyIf you're new here, you may want to subscribe to my RSS feed. Read my top posts or learn more about Michael Gray. Want more frequent updates follow me on Twitter. Thanks for visiting!
Go ahead type in any single letter of the alphabet into a google search box. Think it highlights an algorithmic bias that wikipedia is listed top 10 for every one. Sure no one really searches for the “a” but when you play on the fringe you learn things about how an algo works. Still not convinced try the numbers 0-10 … still think I’m not onto something … try it with [aa] and go to [zz] the only one wikipedia doesn’t rank for are [ee] and [tt]. Want more proof try [00] or [11] or [22] or just make up some 3 digit number like [897] or [396] … sure the wiki should absolutely be ranking for all those terms … right
Sphere It
July 5th, 2007 at 4:33 am
well, [aa] doesn’t give a wiki page, but AA is an actual acronym. You make a great point though, the only thing I ever did with google-coop was make a search that eliminates wiki and ebay. By removing those two I’ve increased the quality of most searches by 40% on the front page by dropping 4 (2 regular, 2 indented) results I’d never click on.
July 5th, 2007 at 4:36 am
Warum Wikipedia zu Recht für alle Zahlen mit drei Ziffern in den Top 10 ist…
Einige Blogger haben sich darüber aufgeregt, dass Wikipedia für fast alle Buchstaben (b,c,d), Buchstabenkombinationen (ab, bc, cd) und Zahlen (1, 22, 333) in den Google Top 10 ist. Ich sage: Zu recht!
Schauen wir uns ein paar Beispiele an:
…
July 5th, 2007 at 4:37 am
It’s pretty disheartening, isn’t it? As well as Wikipedia, I’ve got a bee in my bonnet about other massive sites dominating the SERPS - YouTube is pretty much guaranteed top ranking for music videos, etc.
All you need are a few market dominating websites like YouTube and Wikipedia competing with you for keywords and the chances of making the top 10 position is ever less likely….
July 5th, 2007 at 5:09 am
That’s Google’s try to emulate Yahoo Mindset. Personalized search scans the surfer’s brain for a context. If that fails, the query engine adds extreme coordinates to the search query’s context, that is Wikipedia for “informational” and eBay for “shopping”. Probably that’s a hard coded test/debug statement a Googler forgot to comment out, and now that the algo’s source is back in its safe nobody cares. Seriously, I don’t think it’s possible to find an explanation or excuse which is sillier than the actual search results.
July 5th, 2007 at 8:21 am
for [aa] I have a wikipedia listing for alcoholics anonymous at #9
July 5th, 2007 at 8:33 am
That is hilarious, scary and disheartening all at the same time….
July 5th, 2007 at 8:50 am
Though to be fair, Wikipedia does have one of the more comprehensive pages on the letter w available online
July 5th, 2007 at 9:18 am
Sure, that makes sense. If I’m looking for ’sex position 69′, I want boring Wikipedia information. When I desperately need the phone number of the 911 emergency service (can’t remember that, so I have to Google it), I want Wikipedia topping the list. And when I search for Se7en, I want to see text in stead of moving images…
July 5th, 2007 at 10:19 am
it’s sad really, might as well just start searching at wiki directly at least there will be less ads
July 5th, 2007 at 11:53 am
That’s interesting; this only works on seven or eight of the 26 letters on google.co.uk - perhaps Google thinks we Brits aren’t as partial to the old Wiki? I’m curious to know how some of the others have got to the top spots, especially http://www.sainsburys.co.uk and (inexplicably) http://www.kcl.ac.uk for the letter “S”…
July 5th, 2007 at 3:05 pm
it’s very interesting - I think it’s sheer volume of content on Wikip, and links to wikip pages. Could be a bias as well.
It is interesting to note though Michael how when we play on the edges, we do find crazy things.
July 5th, 2007 at 6:10 pm
wow! i’d have thought you’d rank somewhere for some letters due to all of the times the alphabet has sponsored your blog posts.
(though i can’t find any anymore, i do remember seeing that several times here before)
July 5th, 2007 at 8:41 pm
google is fukn full of shit, no hand jobs my ass LOL
July 5th, 2007 at 10:39 pm
Wiki you own us all
July 5th, 2007 at 10:52 pm
It has nothing to do with bias - ebay and wiki have three things going for them: 1) tons of inbound links where the anchor text contains high value terms of every possible combination (wiki has 437,824 inbound links to its site from other sites according to Yahoo! and ebay has 33,430,598), 2) they have lots of internal links and 3) they have tons of pages with very deep content on every imaginable subject. So lots of inbound links + lots of internal links + lots of content = a winning combination of SEO attributes that no one else on the planet can beat. Certainly Google did not set out to make these sites number one or pre-eminent in their fields. Google is only out to make Google number 1.
July 6th, 2007 at 4:57 am
I tried something out which is also quite interesting, although I’m not sure what it really represents.
I tried -1945 1946 and I got some interesting results, notably with no sign of wiki at all.
Then I tried -b c and google returned nothing.
Try c -b and I got similar results but with wiki coming in at number 7.
Still I tried google -msn and got nothing from wikipedia.. but I did come across quite a funky google page here http://joeanderson.co.uk/google/, I quite liked it!
July 6th, 2007 at 8:22 am
I have to agree with Harry on this one. There is no bias, Wikipedia just has tons of inbound links, great internal linking and is an authority site.
Not that PR is the end all be all, but every one of the letter pages at Wiki that I saw had a PR of 6. How many other sites have a page devoted to a single letter with that high a PR? Of course its gonna be in the top 10, its not a competitive search. No one is optimizing for b, g and r.