Skip to main content

Google Senior Trends Analyst disputes claim that shared web hosting hurts SEO performance

SEO analytics
(Image credit: Pixabay)

John Mueller, a Google Webmaster Trends Analyst has weighed in on a debate over how the company ranks websites on search result pages. 

Contrary to the outcome of recent analysis, Mueller says that a website’s SEO performance shouldn’t be negatively impacted by the presence of so-called bad neighbors on a shared server.

The study Mueller was responding to states that, when it comes to Google SEO scores, shared web hosting servers have “a detrimental effect on the organic performance and rankings of the websites.” But posting on Twitter, Mueller argued that the methodology of the research in question is fundamentally flawed. 

He said that the effect noted in the report by Reboot Online wouldn’t hold for normal Google searches and, more importantly, that he knows of no Google algorithms which take the IP addresses of other sites on the same server into account when ranking a page. 

Mueller backed up his argument with the example of Blogger, a hosting service where sites share “all the same infrastructure [and] the same IP addresses” but some rank terribly on Google while others do great.

Google SEO

The original study was set up to test the theory that “with cheaper shared hosting solutions usually attracting many lower-quality websites (like spam and PBN websites)”, Google’s ranking algorithms would be biased against sites run on shared servers. 

To test this idea, the authors created twenty test websites targeting a made-up, nonsense keyword. Half of these sites were hosted on AWS (Amazon Web Services) and half were setup with cheap shared hosting platforms. After just over three months of tracking how each performed on Google rankings, the Reboot Online team concluded that “websites hosted on a shared IP address ranked less strongly than those hosted on a dedicated one.”

Although Mueller says that he loves it “when folks experiment like this”, he also suggests that this data isn’t particularly useful. 

As he points out, a problem with using a nonsense keyword to evaluate Google’s algorithms is that, for Google to function properly, the platform needs to understand the context and meaning of a search term. For obvious reasons, that can’t happen if a keyword is picked specifically because it’s not used by real people.

Via: SearchEngineJournal