2007 Tour de France Winner - Wikipedia
July 6th, 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!
The Tour de France is scheduled to begin tomorrow, so I was doing some research for another site and was really amazed at the favoritism Google shows Wikipedia. So I decided to break it down and show bad it really is. Below is a list of all the stops on this years Tour de France, I’ve gone through and removed the accent characters because most people are never going to bother typing them in. Lastly to the right of each term you find the # where wikipedia is listed.
[Tour De France] #3
[London] #2
[Canterbury] #4 and #8
[Dunkirk] #1 and #2
[Ghent] #1
[Waregem] #1
[Compiegne] #1
[Villers-Cotterets] #1
[Joigny] #2
[Chablis] #1 and #2
[Autun] #1
[Semur-en-Auxois] #6
[Bourg-en-Bresse] #2
[Le-Grand-Bornand] #6 (french wikipedia)
[Tignes] #4
[Val-d'-Isere] #4
[Briancon] #1
[Tallard] #7
[Marseille] #1 and #2
[Montpellier] #1
[Castres] #1 and #2
[Albi] #3
[Mazamet]
[Plateau-de-Beille] #6
[Foix] #2 and #3
[Loudenvielle] #3
[Pau] #1
[Orthez] #1 and #2
[Gourette] (no listing in top 10)
[Castelsarrasin] #1
[Cahors] #1
[Angouleme] #1
[Cognac] #1 and #2
[Marcoussis] #6
[Paris] #3
Out of 40 Tour de France related search phrases wikipedia has top 10 listing for 39 of them, wow Wikipedia must be a great source for Tour de France info and English, Belgian and French cities.
At SMX I heard Matt get up and say it’s really only SEO’s who have a problem with the wiki, and for regular users it answers their questions. While I will agree that some of the pages do have a lot of information (I have no idea if it’s right or not) I will say this is really a prime example of the “authority knob” set to high, and out right domain trust abuse.
So on what data am I basing that statement? Lets look at the wikipedia pages for Loudenvielle, Joigny and Semur-en-Auxois. What do we have 2 completely empty pages and one wikipedia stub. What conclusion can we draw from that, basically if you have a trusted domain like wikipedia, you can put just about anything, or in this case the lack of anything on a nearly blank page and it will rank.
So C’mon Matt I double dog dare you to tell me how those three pages are helping anybody … other than wikipedia … in fact I triple dog dare you … with a sundae and cherry on top.
To head you off at the pass, blah blah blah yes those 36 other pages do contribute value (accurate or not) but that’s not the question. The question is why does Google rank 3 useless pages from wikipedia, in top 10 spots, for terms that people are likely to be looking for in the next coming weeks?
Secondly don’t be a complete dingnut and say “gee Gray you could make the internet a better place by adding to those wikipedia pages”, cause that’s just lame, with all the traffic wikipedia gets they don’t need anymore help from me.
Lastly I strongly urge everyone on a wordpress platform to use the wikipedia nofollow plugin. Plain and simple any idiot in front of computer (including me), can sit down and edit any article on wikipedia, no further explanation should be needed.
(hint for there’s an SEO lesson in this post as well)
Tags: tour+de+france, tourdefrance
Sphere It
July 6th, 2007 at 6:30 am
Hi…
when I’m searching for “Gourette” here in France, I have Wikipedia FR result in 9th position… Looks to me you made de strike on it.
July 6th, 2007 at 11:00 am
If everyone just nofollowed all links to the Wik we’d be fine. I’m not sold that there is some special tweak in the algo to favor Wiki, but it’s not out of the question either.
The problem is Wikepedia is a SAFE place for people to link to and Google has webmasters so damned paranoid about where they link that people don’t like to give them out to sites that could be the least bit competitive. Wikipedia doesn’t sell anything so…
But they DO compete for search results and that’s why we should nofollow when referencing the site. Either that, or find somewhere to link.
Need to link out to a good site to beef up the outbounds on your thin affiliate pages? Try a simple ‘your keyword site:edu’ or ‘your keyword site:gov’ and there are plenty of NON-Wikepedia pages to which you can link.
July 6th, 2007 at 1:17 pm
I’ve thought that for the past year. I’ll Google something and pull up the Wiki page because it is at the top of the SERP. The page is sometimes totally useless.
It would be nice to use one of your backlink tools and figure out how many sites actually link back to these useless pages. If G is supposed to use links as a factor for how well a page ranks in the SERPs, then each of these pages should have a lot of links. Considering that some pages are only stubs, that is highly unlikely.
July 6th, 2007 at 1:30 pm
Oh yeah - one more comment. If you find that no one is linking to these pages then that proves that G has modified the SERP algo to artificially boost Wiki to the top of the SERPs. Thus, the “Use the NoFollow plug-in” campaign that SEOs promote is useless b/c G isn’t using links in their ranking calcs anyway. G is going to put Wiki at the top of the SERPs no matter how useless the page is or how few links their count is. However, if G wants its SERPS to just send everyone to Wiki for every term, then maybe we should stop using G as our search engine and just go to Wiki for everything?
July 6th, 2007 at 1:47 pm
If every time I go to Google the results direct me to Wikipedia, how long will it be before I drop the middle man and go directly to Wikipedia?
July 6th, 2007 at 1:52 pm
Man, I must really be avoiding work today b/c I have one LAST comment.
This has me thinking about how much does G really trust their algo to rank empty pages so high? You know that the algo calculates links and page content into the SERPs. But when a page has no links and no content, then how does the algo still shoot it to the top of the SERPs? Does Wiki get a free pass to the top of the SERPs for every page on its site? Also, I thought that the time people spend on a page before clicking the Back button can have an impact on the SERPs. These pages obviously have zero “Time on page” values because you will immediately click the Back button right when the empty page is loaded.
I suggest that G makes a minor tweak to their SERP algo. Assuming that Wiki will continue to get a free pass, then common sense should still get to play a little part in the results.
SerpRank := GetRanking();
if (URL == ‘Wiki’) {
//Free pass to top for all Wiki pages
SerpRank := 3;
//Common sense adjustment
if (Links == 0 || Content == null){
SerpRank := 10,000;
}
}
July 8th, 2007 at 11:32 pm
I didn’t know there existed a no-follow plugin just for Wikipedia. Thats funny. Issuing top ranking for blank pages indicate Google is not perfect.
July 10th, 2007 at 4:19 am
Hi Michael,
We’ve just written a search plugin for firefox that lets you search Google without wikipedia. Thought you might like it!
Perhaps we can start a movement to get the public to use it and then Google might take notice.
Will
July 10th, 2007 at 12:05 pm
Of the three articles you cited from wikipedia, most of the following sites (after the article listing) are even more useless. On the second SERP most seem to be commercial sites. I think we can presume that most people will specify if they are looking for accommodation. Thus, it’s far more likely they are looking for information.
The only problem seems to be that there is not enough quality information on these places on the internet.
Or perhaps there is, only the information buried under all those commercial sites. If that is the case, I don’t think we will be seeing SEO’s complaining about that.
July 13th, 2007 at 3:24 pm
>> I’ve gone through and removed the accent characters because most people are never going to bother typing them in.
Erm, yeah. Well, maybe most SEOs might not type them in if their first language isn’t one with accents, but, you do know that the accents aren’t just there to be pretty squiggly lines don’t you.
For example in Greek we have one word παιδάκι which means ‘child’, and you add a dieresis and hey presto, παïδάκι means ‘cutlet’
Well, it would if most browsers actually supported the iota character with a dieresis
July 25th, 2007 at 3:23 pm
Again a very good article Graywolf.
I am very happy that not everyone is kissing Google and MC’s ass.
The wikipedia pages or everywhere and to find other good resources you sometimes have to dig very deep or just use another engine if your lucky living in a land where more then 1 SE has up to date results.(Ask and Yahoo suck overhere)