Google Love Affair with Wikipedia
May 2nd, 2007 by Michael Gray in GoogleIf 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!
Much has been said about Google algorithmic favoritism for Wikipedia. I was working on a project this weekend about celebrities, and wikipedia’s dominance in the SERP’s was nothing less than astounding.
Rather than be accused of rigging the test I’m going to grab all of the actors/actresses listed in the top actors section of the Yahoo Buzz Report. The list has been de-duped and alpha sorted, results with Wikipedia in the top ten are indicated with an *, results with Wikipedia in the top 3 are indicated with an !:
[Aishwarya Rai] !
[Alec Baldwin] !
[Angelina Jolie] !
[Angie Harmon] !
[Carmen Electra] !
[Drew Barrymore] !
[Elizabeth Hurley] !
[Eva Longoria] !
[Goldie Hawn] !
[Halle Berry] !
[Hilary Duff] !
[Hugh Grant] !
[Jaimee Foxworth] !
[Jamie-lynn Discala] !
[Jennifer Aniston] !
[Jennifer Lopez] !
[Jessica Alba] !
[Jessica Biel] !
[Kelly Ripa] !
[Kerry Washington] !
[Kristanna Loken] *
[Linda Hamilton] !
[Lindsay Lohan] !
[Lucy Liu] !
[Pamela Anderson] *
[Richard Gere] !
[Rosie Perez] !
[Sarah Silverman] !
[Scarlett Johansson] !
[Shawnee Smith] !
[Shilpa Shetty] !
[Tobey Maguire] !
[Tori Spelling] !
[Yasmine Bleeth] !
So out of 35 of the most popular actors searched on the web last week Wikipedia has top 10 listings on all 35. More impressive is that they have a top 3 listing on 33 out of 35 listings. Most were number 2 with the odd #1 or #3 thrown in. If you bothered to click through you might have noticed that IMDB was #1 for most of the listings, and you’re wondering why I’m not complaining about that. Let’s cruise on over to the top overall searches for the same time period. Again list has been de-duped and alpha sorted * for Wikipedia top ten ! for Wikipedia top 3. To make it even more apparent I’ll remove any of the people listed in the top actors list as well:
[Akon] !
[American Idol] !
[Avril Lavigne] !
[Beyonce Knowles] !
[Britney Spears] !
[Chris Daughtry] !
[Christie Brinkley] !
[Dancing Bush]
[Democratic Presidential Debate 2007]
[Department Of Defense] !
[Elvis On American Idol]
[Hi-5]
[Internal Revenue Service] *
[Kelly Clarkson] !
[Lost] !
[Mark Twain] !
[Mike Penner]
[Milka Duno] !
[Naruto] !
[NBA] *
[New England Journal Of Medicine] *
[Prime Numbers] !
[Prince Harry] !
[Pussycat Dolls] !
[Regis Philbin] !
[Rocky Mountain News]
[Rosie O'Donnell] !
[RuneScape] !
[Sea Otters] !
[Spider-Man 3] *
[T-Pain] !
[WWE] !
Out of 32 possible terms Wikipedia had top 10 listings on 27. Out of 27 possible terms Wikipedia had top 3 listings for 22. These aren’t ordinary terms either, they are high volume search terms. Ok some people are saying that’s Yahoo list not Google’s, so why didn’t you use the Google Zeitgeist? Well as has already been demonstrated Google lies on the Google Zeitgeist so let’s leave the Pinocchio of search sitting in the corner. Next we’ll look at the terms that didn’t have Wikipedia listings int he SERP’s:
[Dancing Bush] - that’s news, people were really looking for DancingBush.com, not enough time for it to be in Wikipedia yet and be authority scored yet.
[Democratic Presidential Debate 2007] - that’s a news item again.
[Elvis On American Idol] - It was Elvis theme week on American Idol and Celene Dion sang with Elvis, so again news.
[Hi-5] - This one surprised me HI-5 is a transplant Australian kids band and has a Wikipedia page so go figure.
[Mike Penner] - Mike is sports writer who announced he’s having a sex change operation, so again news.
[Rocky Mountain News] - I have no idea why people are looking for this, nothing jumped out at me, drop something in the comments if you know.
What conclusions can we draw, wikipedia’s, not a good source for recent news, so you won’t find those types of items there. What else changed, IMDB.com dropped out of all of the searches except [Spider-Man 3] which is a movie. That’s exactly how it’s supposed to work, IMDB is a topical authority, they dominate in their area of expertise and not in any other area.
What’s the problem with with Wikipedia, well in a word wikiality (how funny is it that Wikipedia has a #2 listing for [wikiality] further proving my point). Anyone with a computer can edit Wikipedia, and we rely on the least common denominator of mob intelligence to determine accuracy. That wasn’t entirely true, you can actually win an editing war if you are willing to expend large amounts of time defending your fiefdom of personal edits, I just have better things to do with my time.
Why am I complaining, isn’t Wikipedia free and non commercial? Have you visited techcrunch and learned that Wikipedia has removed the no follow from links to their new commercial project wikia. Starting to see a problem yet? Who do you think is going to be the number one traffic referrer for Wikia … Wikipedia. Kinda gives them an unfair advantage getting all that free traffic from Google doesn’t it? What’s really going on is Google is willingly allowing Wikipedia to be a doorway page to funnel traffic to Wikia, not a bad set up to be on the other side of the cash register for eh?
How do we fix this problem, Google has to admit that Wikipedia really isn’t an authority for anything. Sure some of the information there is good, but I can’t tell you how many facts in the past 3 weeks I found that were omitted, misinterpreted, or just plain wrong. What you think I should do my part and fix them? Know what I’m not on Wikipedia or wikias payroll so why should I help them out? Adding no follow and taking away the value of the links removed the invisible currency and payment that existed. So now I have to guess and determine if I am going to get enough traffic from them before somebody who doesn’t have to qualify their level of expertise comes along and decides they don’t like my edit for whatever reason and just removes it. Wow sure sounds like a website that should be dominating the SERP’s don’t you think?
Sphere It
May 2nd, 2007 at 10:01 am
With the recent digg-nation revolt fresh on my mind and your discussion here about wikipedia it got my scratching my head…
If digg’s loyal users can revolt…surely the small(er) army of wikipedia editors (when it continues to dawn on them that their volunteer work for a “non-profit” continues to be leveraged for wikia) can too!
What will that riot look like? With editors (the police) involved in that conflagration it would get ugly fast. (I think Wikipedia has the power to quickly lock down any/all stories with the click of a radio button but…for awhile I can only imagine what could happen to the millions of wikipedia entries.)
1. African Elephant? Tripling in numbers every month!
2. Global Warming? True…no false…True and we’re all going to die tomorrow.
3. Best Actor in 2005? Will Ferrel in Office Space
Back to your point on search serps…Yahoo! is just as guilty I believe in providing Wikipedia with top listings. Everyone’s algo seems to see the content/traffic/black-box stuff as highly relevant to many searches. There’s not much a search engine can do … set up the search system and business models be damned. If your business gets bruised by it…you certainly have to re-examine the algo but it’s tricky.
I think an analogy might be to have the advertising group from the LA Times knock on the door of the editorial staff asking them to change policy on a story because it’s hurting ad sales. You want to see a riot? That would surely do it.
/marc
May 2nd, 2007 at 10:43 am
Rocky Mountain News - that’s the newspaper where the main editor of Kotaku.com writes for, so I’m guessing maybe he mentioned them and that caused a spike. Actually, I think his gig working there is just about over or maybe just finished, so that could be it.
May 2nd, 2007 at 1:02 pm
Great post. In Google’s defense, Yahoo also shows similar love towards Wikipedia (I spot checked a few and they all had wiki in the top 10).
I’m not confident the engines will demote wikipedia rankings, as they should in many cases. If one engine does find a way to stop using Wikipedia as a crutch, then the others might follow. As Marc mentions, an editor revolt or increased awarness that Wikipedia is feeding Wikia’s pockets could also hurt Wikipedia, but it would take a lot to bring down this Goliath (they rank #2 for that term, by the way). People would need to stop linking to wiki, editing wiki, and, most importantly, clicking on wikipedia in the SERPs to bring it down.
May 2nd, 2007 at 3:00 pm
Google does not have a love affair with Wiki, it has love for any site that allows “votes” and human interaction to be displayed.
SEOmoz is a great example of a blog that will do well in future algorithms with it’s inclusive format and voting system. Want to compete? Be in the same place by building the ultimate interactive site for your niche.
What are some examples of sites that make money using this same format? Look at Amazon.com for example. People review stuff, it is not the original product that ranks, it is the combination of other factors and product review is just one of them.
Life is good!
May 2nd, 2007 at 3:32 pm
If digg’s loyal users can revolt…surely the small(er) army of wikipedia editors (when it continues to dawn on them that their volunteer work for a “non-profit” continues to be leveraged for wikia) can too!
I run an 11-year-old web site about the rock band U2. It’s been the authority site since it launched, and only fell to #2 in the rankings in 2001 when the band’s official site opened.
About 1-2 months ago, the U2 Wikipedia page finally took over the #2 spot on Google and bumped me down to #3. The editors were patting themselves on the back in the “talk” section.
That’s just one anecdotal piece of evidence, but I wouldn’t hold my breath waiting for such a revolt.
May 2nd, 2007 at 5:03 pm
Problem is, the TechCrunch story was basically a troll and you didn’t bother reading the comments, which pointed out:
* Wikia is not “the commercial spinoff” of Wikipedia. They’re completely separate organisations.
* Wikia links only don’t get linkjuice if they’re done via the interwiki map, which includes lots of outside wikis.
* Wikimedia projects are for the readers; we actively don’t care about third parties (SEOs, Googlemancers) trying to curry favour with fourth parties (Google), and never mind our readers or editors, as it would be a disservice to our projects to do so. SEOs are not our constituency, however much they complain (and boy do they complain).
Any further questions, just ask.
May 2nd, 2007 at 5:06 pm
By the way:
“If digg’s loyal users can revolt…”
You do realise, of course, that Wikipedia has been dealing with idiots putting the Forbidden Number into every input box on the web all day today. Distributed spam is still spam, and we deal with it accordingly.
ps: if you can really convince people to put rel=nofollow onto every link to Wikipedia, I fervently ask you to do so. It isn’t actually our ambition to be as absolutely popular as possible - we’re here to write an encyclopedia. Being Top 10 only distracts from that, as we get spamming idiots from Digg and SEOs thinking we’re a platform for them. We’re here to write an encyclopedia.
May 2nd, 2007 at 5:08 pm
“Wikia links only don’t get linkjuice” - that should be “only get linkjuice”.
The policy for this is currently under discussion. I’m personally pushing that, if our Google page rank is as gosh darned superpowered as SEOs keep saying, then we should give it preferentially to Free Content sites, as defined by http://freedomdefined.org/ .
May 2nd, 2007 at 7:31 pm
“if you can really convince people to put rel=nofollow onto every link to Wikipedia, I fervently ask you to do so. It isn’t actually our ambition to be as absolutely popular as possible - we’re here to write an encyclopedia”
David,
We both know that it is unrealistic to believe all site owners change their links to nofollows, so how about wikipedia simply add a noindex to every page?
I agree that wikipedia is focused on building out a strong encyclopedia and webmasters are more at fault for wikis search success (from linking to it), but I noticed that you are really an SEO yourself. On your own wikipedia bio, this is what bothers me:
“I am the first Google, Yahoo!, and Live Search hit on my own name, recently outranking the Dutch painter.”
May 3rd, 2007 at 7:56 am
“I noticed that you are really an SEO yourself.”
Lmao. If watching SERP ranking for your own name makes you an SEO than I guess everyone and their grandmother is an SEO. Puleeeez.
“so how about wikipedia simply add a noindex to every page?”
Yeah, all the whiners out there being outranked by spam/Wikipedia/whatever else would love that now, wouldn’t they?
Fact is most commercial sites with their panties in a twist over getting outranked by Wikipedia offer even less value than a Wikipedia stub page. If you’re outranked by Wikipedia, you deserve to be outranked.
@Graywolf,
It’s more like a fours*me sex *rgy than a love affair. You forgot to check Yahoo and Live - Wikipedia rules the SERPS on those engines too. And if you say “well yeah, but those engines drive less traffic” I’ll tell you so what? Those SERPs prove Google isn’t treating Wikipedia any worse or better than Yahoo or Live.com.
You want to know the real hidden agenda for Google here? Google wants to get rid of commercial sites on the organic front page. Keep organic serps informational so that when people search to buy something, they go to Adwords instead.
See? By favoring info sites like imdb and wikipedia, Google is increasing Adwords CTR and Ad spend (when commercial / affiliate sites can’t rank organicly they go to Adwords - more $$$ for Google).
May 3rd, 2007 at 8:36 am
“Have you visited techcrunch and learned that Wikipedia has removed the no follow from links to their new commercial project wikia.”
Beep.
First, Wikipedia cannot have “projects”. Wikipedia is not a company. You’re probably talking about the Wikimedia Foundation.
Second, the Wikimedia Foundation does not run Wikia, and actually has no financial nor managerial links to Wikia (except that Wikia donates a little money to the Foundation, which is frankly not important compared to the general Foundation budget).
Third, Wikia is not singled out for a favor. Actually, Wikipedia has a whole “interwiki table” which gives links without the “nofollow” attribute to hundreds of sites, including Wikitravel and other sites unrelated to Wikipedia and Wikia.
Fourth, Wikia’s sites seldom show up high in search engines. This is because they are really on niche markets (encyclopedia of Star Wars or similar).
Thus, please get your facts right and stop whining.
May 3rd, 2007 at 10:19 am
Nice article and congrats on getting the data. We have all seen Wikipedia dominate the SERPs though and I feel that the main point of the reasons why Google lists Wikipedia at the top is being forgotten.
The way I see it is this:
You may dislike Wikipedia for it’s unreliable information but the average searcher LOVES Wikipedia. It’s easily accessible, relatively unbiased concise information. People want to read this. Google is simply giving searchers what they want, it snowballs in magnitude once Wikipedia is getting all those top searches.
Traffic to a website has to be one of the main factors Google uses in it’s SERPs these days. More traffic to Wikipedia, more relevance to Google SERPs. Bring up a site with more information on all these top searches that millions of people read and Google will start adding that to the top of the SERPs instead.
We don’t need to tell Google they are doing a bad job but rather the person who wants to rely on the information they read in Wikipedia.
May 3rd, 2007 at 6:22 pm
same goes for automotive searches. Try doing a make / model (Ford Mustang) type search and Wikipedia usually comes up #1 or #2.
May 8th, 2007 at 2:57 am
Jimmy Wales is practically an angel.
The wikipedia search function is so awesome its hard not to see Jimmy Wales new search project being a huge success.
May 8th, 2007 at 3:00 am
Even though I too am a libertarian only a complete douche would name their kid after an Ayn Rand character.
September 24th, 2007 at 12:44 am
Very, very old post, I know. But for any future readers here, I’d just like to point out that the reason wikipedia results are so high is that THEY ALL LINK TO EACH OTHER. Of course Wikipedia results are going to be high. Wikipedia pages have ranks, and when popular ranks point to each other you can get an exponential increase in their ranking. It’s like, a whitelisted link farm. That’s why the results are so high.