It's been a long standing assertion amongst the A-list members of the SEO community that PageRank does not equate with ranking: Too many low PR sites hitting top spots in the SERPs, outdoing their high PR competitors, it's really hard to come to any different conclusion. However, up until now hard statistical data has been hard to find. Yesterday we were approached by Stéphane Labert of HTML4SEO, a French SEO auditing software company. He pointed us to a statistical study they had recently conducted, determining PR evolution in the Google SERPs over the past 6 months and analyzing the correlation between PR as published by Google - this is apparently Toolbar PR (TPR) and web sites' SERP rankings. You can download it here (PDF format): Google PageRank SERP Statistics Warning: This is pretty technical stuff and not for the mathematically faint of heart! Some Remarks So what do we think of this study after having analyzed it in some depth? Well, for one it would have been nice if the paper had indicated how many sites and SERPs were analyzed for the purpose of this study in the first place ("thousands of SEO audits" does seem rather vague) and how exactly the PR values referenced were determined originally. All PR related statements are exclusively tied to PR values as published by Google. This is an important distinction inasmuch as TBR doesn't accurately reflect "real" PR (RPR) as actually used by Google in its internal calculus. For instance, PageRank is not limited to a 10 digit scale of 1-10 whereas the toolbar doesn't reflect decimal fractions.
The Rank and the Pagerank are strongly correlated. A good pagerank improves ranking undoubtly.
This is clearly borne out by the stats graphs that cover May 2007 through October 2007. However, let's not forget that there's another, equally plausible conclusion we may draw from this, especially in view of the fact that it is Google who publish these PR values: Google may want to see a correlation between PR and SERP rank being established or, at the very least, does not seem interested in preventing this from happening. Using a linear interpolation of Rank/Pagerank seems somewhat problematic for the Top Ranks because the aberration between linear interpolation and the gradient is appr. PR1 as published PR is logarithmic. (At least, that's the going assumption.) What's more, employing linear interpolation leads to values reduced by appr. PR1 on B Y-intercept. However, the statements derives from this linear interpolation would very likely be confirmed if we used non linear interpolation instead:
1) The Rank and the Pagerank are strongly correlated. A good pagerank improves ranking undoubtly. 2) The B Y-intercept decreases. In other words, it appears that the requirements in Pagerank to reach the top of the Google SERP decrease month after month.
This would imply that the loss of PR experienced by many sites in the past weeks won't impact rankings negatively as apprehended by many webmasters and SEOs. Another interesting gradient within this context is the one titled "Evolution in 2007 of the Y-intercept of the Rank/Pagerank Linear expression." It demonstrates that the PR value required for a #1 ranking is actually declining regularly month by month. This may be caused by a consistent and regular algorithmic attenuation of PageRank and/or by the increase of indexed pages expanding the starting basis and, hence, reducing overall PR. Finally, HTML4SEO have announced that more's to come:
The study is restricted to the Pagerank, although we could extend it to the backlinks and the HTML content pages listed in the Google SERP. We will soon publish other studies on these 2 other major SEO criteria.
We for our part are certainly looking forward to it!
Discuss Add this link to... Bury
The Rank and the Pagerank are strongly correlated. A good pagerank improves ranking undoubtly.
This is clearly borne out by the stats graphs that cover May 2007 through October 2007. However, let's not forget that there's another, equally plausible conclusion we may draw from this, especially in view of the fact that it is Google who publish these PR values: Google may want to see a correlation between PR and SERP rank being established or, at the very least, does not seem interested in preventing this from happening. Using a linear interpolation of Rank/Pagerank seems somewhat problematic for the Top Ranks because the aberration between linear interpolation and the gradient is appr. PR1 as published PR is logarithmic. (At least, that's the going assumption.) What's more, employing linear interpolation leads to values reduced by appr. PR1 on B Y-intercept. However, the statements derives from this linear interpolation would very likely be confirmed if we used non linear interpolation instead:
1) The Rank and the Pagerank are strongly correlated. A good pagerank improves ranking undoubtly. 2) The B Y-intercept decreases. In other words, it appears that the requirements in Pagerank to reach the top of the Google SERP decrease month after month.
This would imply that the loss of PR experienced by many sites in the past weeks won't impact rankings negatively as apprehended by many webmasters and SEOs. Another interesting gradient within this context is the one titled "Evolution in 2007 of the Y-intercept of the Rank/Pagerank Linear expression." It demonstrates that the PR value required for a #1 ranking is actually declining regularly month by month. This may be caused by a consistent and regular algorithmic attenuation of PageRank and/or by the increase of indexed pages expanding the starting basis and, hence, reducing overall PR. Finally, HTML4SEO have announced that more's to come:
The study is restricted to the Pagerank, although we could extend it to the backlinks and the HTML content pages listed in the Google SERP. We will soon publish other studies on these 2 other major SEO criteria.
We for our part are certainly looking forward to it!
Discuss Add this link to... Bury
LinksMarker is a Web 2.0 content management system
Comments