• SEO-Blog
  • Startseite
  • RSS-Feed, complete Feed
  • About the Blog, Blogcharts
  • Archive: 2003, 2004, 2005, 2006, 2007, 2008, 2009
• Weitere Postings
  • Diversification as a road to success
  • Zanox goes to Springer
  • Google on the trail of the VeriSign sitefinders
  • Ranking-factors: H1 to H6 in detail
  • Small site, small problems – big site, big problems
  • A glimpse into the heart of Google
  • Robots.txt deluxe – the Google-supported extensions
  • Supplemental Index – second class websites?
  • Supplemental Index – how did I get here?
  • Revised Google-Webmaster-guidelines
•  

Recognizing linksales – theoretical danger or already in practice?

The last few month, the subject matter of linksale or alternatively link-renting has made quite a ruckus. Google, whose very foundations of ranking-criteria are struck, has been announcing, for a while, that sites that are selling links will be detected and dealt with accordingly. Matt Cutts has even invoked users to report these sites, so that they could test a algorithm which is supposedly being developed.

Now that DaveN has blogged that both linkbuyers and linksellers are supposedly already affected, I should go ahead and test how hard it is to realize the detection of bought links. The following tests are all based on the assumption that most linkselling sites will follow a certain formula: strong sites will sell links to sites are not subject-relevant but who have a high commercial relevance. That Google is able to detect the subject-relevance of a site should, at the latest, be known since AdSense. It should also not be hard to figure out a selection of strongly commercial keywords thanks to the AdWords-database. Seeing that, at the moment, I did not have something like that just lying around, I had to pick both of these criteria – hopefully useful ones - by hand.

Sites from the linux-community are well liked all along by linkbuyers thanks to their age and the naturally grown links. That is why I examined about 1.000 De-domains with “Linux” in the domain name. Of these domains, I examined automated, outgoing links from the startpage for the large commercial keywords: finance-, travel, internet-access and such.

The results of this short, about 15 minute, examination are interesting. Even though only a little time and effort were invested, the results are already quite persuasive. And even though I will not publish any URLs, which should hopefully be understandable, it is still the “well known” linkbuyers that show up. The detection is surprisingly quick and reliable.

Now, if we consider that the resources at Google are nearly infinite, then we can say that linkbuyers should rethink their future strategies. In the long run, subject-irrelevant-links in footers or those in the navigation that are lovelessly put one below another will be automatically detected and depreciated. Links that have a subject relevance and which are set from within the content should also stay hard to detect in the future.