When Craig Bradford of Distilled reached out and requested if we would prefer to run some SEO experiments on Moz utilizing DistilledODN, our reply was an instantaneous “Yes please!”
If you are not aware of DistilledODN, it is a subtle platform that means that you can do a lot of cool issues within the SEO house:
- Make virtually any change to your web site by means of the ODN dashboard. Since the ODN is a cloud platform that sits in entrance of your web site (like a CDN) it does not matter how your web site is constructed or what CMS it makes use of. You can change a single web page — or extra possible — total sections.
- The ODN means that you can A/B cut up take a look at these adjustments and each measure and predict their impression on natural visitors. They even have a function known as full-funnel testing permitting you to measure impression on each SEO and CRO on the identical time.
When you discover one thing that works, you see a optimistic end result like this:
SEO experimentation is nice, however virtually no person does it proper as a result of it is unattainable to regulate for different elements. Yes, you up to date your title tags, however did Google roll out an replace right now? Sure, you sped up your web site, however did a bunch of spam simply hyperlink to you?
A/B cut up testing solves this downside by making use of your adjustments to solely a portion of your pages — usually 50% — and measuring the distinction between the 2 teams. Fortunately, the ODN can deploy these adjustments near-instantly, as much as 1000’s of pages at a time.
It then crunches the numbers and tells you what is working, or not.
Testing Google’s UGC hyperlink attribute
For our first take a look at, we determined to deal with one thing easy and quick. Craig instructed taking a look at Google’s new hyperlink attributes, and we had been off!
To summarize: Google lately launched new hyperlink attributes for site owners/SEOs to label hyperlinks. Those attributes are:
- rel=”sponsored” – For paid and sponsored hyperlinks
- rel=”ugc” – For hyperlinks in user-generated content material (UGC)
- rel=”nofollow” – Remains a catch-all for all adopted hyperlinks
On the Moz weblog, all feedback hyperlinks are presently marked “nofollow” — following years of SEO greatest practices. Google has said that utilizing the brand new attributes will not offer you a rankings increase. That stated, we wished to check for ourselves if altering these hyperlinks to “ugc” would impression the rankings/visitors of our weblog pages.
To be clear: We are not testing if the pages we hyperlink to vary rankings, however as an alternative the supply web page that hosts the hyperlink — on this case, the weblog pages with feedback.
Here’s an instance of a remark the ODN modified.
After we set the take a look at working, 50% of weblog posts had feedback with “ugc” hyperlinks, whereas 50% saved their unique “nofollow” attributes.
We anticipated a “null” take a look at — which means we would not see a big impression.
In truth, that is precisely what occurred.
If we detected a big change, the likelihood cone on the backside proper would have pointed extra dramatically up or down.
In truth, at a 95% confidence interval, the take a look at predicted visitors would both fall 26,000 visits/month or achieve 9,300 visits/month.
Hence, a null end result.
This validates Google’s statements that utilizing the “ugc” attribute will not offer you a rating increase.
What ought to Moz take a look at subsequent?
While “null” assessments aren’t as enjoyable as a optimistic end result, now we have plenty of cool A/B SEO testing forward of us.
The good thing is we will now take a look at out adjustments with the ODN, and after we discover one which works, go that to our builders to make the adjustments completely. This cuts down on useless growth work and stops the guessing recreation.
We have a Trello board arrange for take a look at concepts, and we would love so as to add some group concepts to the combo. The ODN is presently working on the Moz Blog and Q&A, so something in these web site sections is honest recreation.
We’re additionally taking a look at experiments the place we use Moz information to tell these choices. For instance, a Moz Pro crawl recognized that the Moz Blog titles presently use H2 tags as an alternative of H1. Google lately indicated this possible should not impression rankings, however would not it’s good to check?
What wild/intelligent/ridiculous/apparent SEO issues ought to we take a look at? With every good take a look at, we’ll publish the outcomes. Leave your concepts within the feedback beneath.
Big due to the Distilled Team, together with Will Critchlow and Tom Anthony, for embarking on this journey with us.
And if you would like to be taught extra about DistilledODN and SEO cut up testing typically, this submit is very advisable.