Testing & Experimenting  - 
Authorship Markup: No Correlation to Ranking Ability

In conjunction with the MozCon conference on right now in Seattle. +Moz has released its annual ranking factors study: http://moz.com/blog/ranking-factors-2013

The graph below illustrates that they found close to zero correlation between presence of any schema markup on a page, including author and publisher markup, and ability of the page to rank in search. In fact, author markup actually showed a negative correlation (although the amount of variance is probably within the margin of error).

Now if author markup showed even high correlation with increased rankings, that alone still wouldn't be proof that Authorship was being used as a direct ranking signal ("correlation does not equal causation"). But...if Authorship were a ranking signal at this point, we would expect to see some correlation.

Luis Galarza's profile photoKeith Goulding's profile photoEunice Coughlin's profile photoKevin Wiles's profile photo
Uh oh! Negative correlation is kinda scary. +Mark Traphagen what do you make of this? No need for a knee jerk reaction, right? Wait and watch? 
Great stuff +Mark Traphagen thanks for sharing... Will be very interesting to see how things develop moving forward
I guess we'll have to wait until next year to see that one in action. I think people hyped up AuthorRank a bit too early. It takes Google engineers months (if not years) to implement a new ranking signal.
I don't agree with this I added Rel=Author to my site and 3 days later I had gone from 2nd to 1st for my main term. I think it is having an impact!
Whether influencing ranking or not, authorship certainly improves clicktrough rates.
+Vikram Kinkar don't freak out about that. The graph here exaggerates the deviations from zero because they are all within a very small range. The amount of deviation is so small as to be virtually meaningless. Think of it as similar to the margin of error in public polls.If a variation is within the margin of error, you shouldn't use it as a point of comparison to other units also within the margin of error.
+Kevin Wiles using one instance of personal, circumstantial and possibly coincidental "evidence" against a major study with hundreds of thousands of data points is not a wise stance. The number of factors that affect any small change in rankings from day to day is huge, and we don't even know about many of them. Correlation does not equal causation, especially in a one-off "test."
+Malcolm Simmonds some of us have been saying "only a matter of time" for nearly two years now! LOL

While I think it is likely that Google will some day use Authorship data more actively in the mix of its search ranking algo (provided they can work out all the remaining problems with it as a reliable and meaningful signal), i don't think anything in Google is inevitable. The one thing you must understand about Google is that it's run by engineers and scientists. They base their decisions to implement or kill things based strictly on what the data show.
I think that if more clicks happen because of authorship (what +Erik Hooijer said) than higher ranking would naturally follow. Eventually Google engineers would make that connection. 
I see, +Mark Traphagen, engineers and scientists, eh? That's an interesting perspective from which to judge their actions - such as ditching Google Reader, perhaps. :)
+Eunice Coughlin a lot of SEOs suspect that a rise in CTR could be a ranking factor. BUT, Google is also careful to not allow any one feature it adds to search (such as a rich snippet result) to artificially distort results in a disproportionate way. So as +AJ Kohn has remarked, it is just as possible that they normalize for increased CTR of rich snippets.
+Mark Traphagen I see your point. But wouldn't they see authorship + rising CTR as performance indicators along with other data they collect? 
Certainly. But my suspicion is that right now Authorship is not widely adopted enough, and there are still too many problems with assessing the meaining of the signal, for it to be yet incorporated as a major factor. I believe they are committed to making it work though. Just may take longer than we wanted ;-)
Agreeing with +Mark Traphagen here. I guess patience is key and time will tell whether Authorship will become a ranking signal.
If you see improved ranking as a result of implementing authorship, I think it's because it made that page more easily crawlable, not because there is a direct correlation to rankings. Again, correlation vs. causation is important, but if authorship markup helps a page get indexed faster or more frequently, it logically follows that would have a corresponding positive increase on ranking. Just my opinion.
Not only that +Jenny Halasz, but Authorship could be contributing indirectly in other ways as well, For example, if people see you consistently in their search results providing the kind of content they're looking for, they may become more regular visitors to your site, and also be more inclined to engage with and share your content. All of these could be positive signals to Google as well.
Thanks for sharing this +Mark Traphagen I think it just goes to show the power that Google have that so many people are doing so many studies and tests on this matter, its like we're all trying to guess Google's next move...without success just yet!
In the Moz rankings opinion survey, I gave author rank a score of 1, out of 10, because I had no choice but to give it a score. If there was a "not presently in use" box I could have checked, I would have.

Remember that these ranking scores are like mirrors on car doors, not necessarily accurately reflecting the actual size and shape of the objects seen within them.

Given some recent statements (in May) from Matt Cutts about an author's authority and expertise becoming one of the signals that might boost the rankings of pages in different topics, I do believe that it's possible that we might see some aspect of author rank being used in the not distant future.

Again, the ranking reports from Moz are based in part on the opinions of a number of SEOS, and not necessarily SEOs who have an expertise on authorship markup, or author rank itself. 

The questions were created by Moz, and not these SEOS either, so for example, there was a question about whether or not registering a domain name for more than one year might be a positive or negative signal for the ranking of a page within the survey (an SEO myth that Matt Cutts has personally and publicly busted multiple times since 2005).
I can believe 'Moz' or I can trust my own eyes. On several queries I ran yesterday 9 out of 10 entries had authorship.
Thanks for that info +Bill Slawski - somehow I misread the post; I thought the SEO survey was something separate, and that the stats were based on actual statistical studies. My bad! So all this is really telling us is that most of these SEO's "don't think" Authorship is a ranking signal yet?
+Darren McLaughlin Authorship results appearing is not the same thing as them having a direct effect on search rankings. In certain search verticals, Authorship adoption is high, so of course you're going to see a lot of Authorship results.
It makes little practical difference to me +Mark Traphagen.  Whether the people who have Authorship are doing the other things necessary to rank or not, they still have it and appear high in many searches.   Therefore, to me, it's a good thing to have that probably helps your overall visibility.
I think the Authorship idea is a sound one.  I clicked on your profile and learned that you were an expert in - oddly enough - Google Authorship.  That adds weight to what you're saying and influences me in a real world way.

A system that verifies identity and accessibility has to be a good thing!  I personally am feeling that a real transition from the anonymous scammy search to more trusted verified social is here.  
Yes, +Darren McLaughlin we're on our way to that, and it's where Google wants to go. But Google has gained its success by being untiringly patient. They wait for the data. They only implement a change or feature when the data shows it will bring the desired results. I believe they are still at the early stages of being able to assess author authority (and even what that means), but I do believe they are very committed to it.
I'm not surprised by the data. The idea of authorrank has always been on the assumption of widespread uptake that we haven't seen as of yet.

I think its useful to see Authorship markup and a similar system to RDF triples in that it can identify and match basic subjects, predicates and objects together. As such its only as strong as the data it holds which at the moment is too low to offer any real benefit to search quality. 
Add a comment...