PageRank vs. MozRank
Please Note: LinkScape and MozRank have been updated since this publishing. The graphs and data below are no longer accurate, albeit still interesting IMHO.With the announcing and opening of the fantastic SEOMoz MozRank API for Linkscape, I thought it would be worth taking a bit of time to compare MozRank to PageRank using existing and fresh PageRank data. Essentially, we wanted to determine if there were any patterns of dispersion between the MozRank and PageRank for random pages across the internet. Using the API, we were able to collect data on 1000 various pages. We selected only pages above PR0 to discourage the likelihood of identifying pages with no MozRank data. We also, as expected, have a skew in the representation of higher PR data, as they are just not as common. *Be sure to read Nick from SEOMoz’s Comments Below
Using exiting PR data, we were able to see a clear pattern in the differences between PR and MR as PageRank increases. As PageRank increases, MozRank is more likely to under-represent the actual PageRank (PageRank minus MozRank = The Blue Dots). Once a page or site reaches a PR5, MozRank almost always under-estimates it’s value, which would be represented by a blue dot below the 0 line.
In order to make sure that our tests were not being hindered by aged PageRank data, we ran a similar test with only 100 unique URLs but grabbed the PageRank on-the-fly. As you can see in the representation to the right, a similar pattern emerged. Low PageRank pages were more likely to be over-represented in MozRank data.
The more interesting item to note is that it appears MozRank and PageRank are most similar at the PR3/MR3 range. If you see a PR3, you are likely to see between a MR2 and 4.
I ask you to take this post with a grain of salt, though. I am looking at a very small data-source with very rudimentary statistics. Perhaps someone more talented than I could do a followup. Kudos, once again, to the SEOMoz team for this great resource.
8 Comments
Trackbacks/Pingbacks
- Linkscape Index Update - With Focus on Quality | ITegrity SEO Web Development Education - [...] look at the last index’s mozRank comparison to PageRank (now out of date) check out The Google Cache’s mozRank…
- Linkscape Index Update - With Focus on Quality | Webmasters - [...] look at the last index’s mozRank comparison to PageRank (now out of date) check out The Google Cache’s mozRank…
- SEO News » Linkscape Index Update - With Focus on Quality - [...] look at the last index’s mozRank comparison to PageRank (now out of date) check out The Google Cache’s mozRank…
- Why I Couldn't Resist Adding SEO Site Tools to Chrome - [...] select SEO Site Tools from a drop-down list. The icon in the drop-down list also displays either PageRank, mozRank…
I’m impressed how quickly you produced this! The chart looks rather simular to some I’ve seem Nick make here at the SEOmoz office.
Nice post. And good feedback. This is a familiar graph… mR as you see it is actually scaled in a somewhat arbitrary way (at the moment). You could get a better match by doing a linear transform. This won’t affect the relative comparisons between pages (the order of mozRank).
Then again, PageRank is probably mixing up a lot of effects beyond pure link popularity (e.g. domain authority, domain age, penalties, …)
Try adding noise (e.g. =PR+RAND()-0.5) to get a prettier graph (less bunching up due to lack of PR precision). That illustrates the correlation better 🙂
oh yeah, you should join our google group if you’re interested in more SEOmoz API dev stuff:
http://groups.google.com/group/seomoz-api
I don’t know about Mozrank. I’ve seen highly trusted domains (government sites) with high Pagerank and very low Mozrank and some commercial sites with low pagerank 4 and a very high Mozrank (even a score of 8.3, beating Wikipedia.org, yeah! ).
So I take both measures with a huge bottle of scientific salt.
Author Note: Google’s Toolbar PR is not an accurate measurement of the actual PageRank. It is qualified by anti-spam measures. For example, a page that sells links may have its Toolbar PR decreased but its MozRank will still remain high. However, this is rare enough that across the board we see a fairly reasonable correlation between the two.
It’s always refreshing to read your posts – thanks again
Thanx for the info! Its not getting easier, is it… 🙂
Yeah it’s difficult to identify on what basis the mozrank is decided and how the site are rated, as it is different from the expected results!
So what do think more relevant ? If PR is not updating at its normal interval ( once in 3 or 4 months), the toolbar PR may be very wrong to make any conclusion about the website. So as a marketer can I trust Moz Rank to determine the current value of the website ?