The entire industry of SEO has been built upon data collection and testing. All the tools and metrics we use are largely founded upon bits and parts of other tools and metrics - this providing us a simple metric to work with.
This makes sense as computing infrastructure is costly so these services have to integrate with each other.
At any moment though, a company could switch off a metric and poof, we'd have to search for a replacement.
This is the exact reason why I've always tooted the horn loudly on conducting manual keyword research when you can. No matter what tools you use, you have to dedicate time digging in to see the pages you're competing with.
But we have to start somewhere right?
First thing every keyword researcher does is filter their seed results into search volume & keyword competitiveness.
Then right after, we'll start to analyze our competitors and see how they hold up. Most of the time we're going to start seeing familiar faces in our niche.
Now is the time to organize these folks and start extracting keywords they are ranking well for. We especially want to find the new ones...
So how do we do this?
Well, we organize by some sort of domain score. The most popular is Moz's Domain Authority metric.
Today I want us to question that effectiveness today.
Let's start running a test...
Do You Know What Domain Authority REALLY Is?
It's a proprietary metric created by Moz to help us understand a sites "ranking strength.
This is how Moz themselves describe the metric:
"Domain Authority is based on data from the Mozscape web index and includes link counts, MozRank and MozTrust scores, and dozens of other factors (more than 40 in total). The actual Domain Authority calculation itself uses a machine learning model to predictively find a "best fit" algorithm that most closely correlates Mozscape data with rankings across thousands of actual search results that we use as standards to scale against.
While specific metrics like MozRank can answer questions of raw link popularity (i.e. link equity), and link counts can show the raw quantities of linking pages/sites, Domain Authority (and Page Authority) metrics address the higher level question of a domain (or page's) "rankability."
In other words, it's a 40 internal metrics that they rehash to give us a dependable rankability score of 1 to 100.
Two things we need to keep in mind though:
- 1It's a logarithmic scale. Meaning that going from 20-30 is easier than going from 70-80. Much like how it gets harder and harder to reach the 1 position with each spot you climb.
- 2It's used for comparison with other sites. As time goes on, your pall at the same DA as you could move up based on every recalculation of the metric.
How Effective is Domain Authority in "Rankability"?
This is the main question we're trying to answer. If DA is a score that is supposed to tell us how we're doing compared to others, then the conclusion is...
If the DA of our site is higher than average of our keyword competitors, then we should be able to drive through the search results easier.
If a site has a DA of 70 then we should be able to write an article and have it rank easily if the sites on the first page have less than 70 on average - right?
The Test We Setup
We decided to publish 4 "best" type of articles on 4 different sites to see which one would race further to the first page, after 1 month.
Each site had various DAs that ranged from 10 to 33 and all the articles had the same keyword difficulty (according to Ahrefs) & almost the same search volume (150-250.)
We chose Ahrefs because it scored the best in our last keyword difficulty test.
Each article had nearly the same word count, equivalent keyword mentions, plus we included the keyword in one of the subheaders for best practice.
Everything was structured to be nearly identical except for the Domain Authorities.
We published the articles on January 17, 2018, and tracked the performance of the articles until Feb 19th.
Of course, there a few things we can't control...
What To Expect From Our Little Test
Our thoughts were that a higher DA site would be able to throw its weight around.
As you can see in the picture above, we couldn't find a keyword for site 4 that had 150 search volume, so we had to settle for a higher volume keyword (and naturally, a more competitive one.)
But the show must go on right?
To provide a more granular view, we also included KWFinder volume and keyword difficulty (by the way, we're big fans of KWFinder and you can find our discount and full review here.)
It was also quite difficult to find keywords with comparable average DAs. Meaning that we took all the DAs for the 1st page results (of each keyword) and divided them by 10. See below:
In a perfect test, we would see equal average DAs along with everything else being equal.
One important factor we have to think about is the time frame. The longer a time frame, the more opportunity your competition has to make moves. A perfect test would mean our competitors did nothing in that month long time frame.
Before we pressed the publish button on these articles, here's what we predicted by only analyzing DA:
- 1The site with the lowest DA is going to have the hardest time breaking onto the first page because the average DA is highest.
- 2The site with a DA of 18 should have the easiest time breaking towards the first page because the spread between the site and the average DAs is the smallest. From 18 to 26.5.
The Unexpected Result
In an interesting turn of events...
The site with the lowest DA performed the best.
Site 1 with a DA of 10 ended up on the 3rd page. While the rest stayed on the 5th and 8th page.
This is completely different than what we had envisioned...
What Happened With Our SEO Test?
With the results gone sideways, a few thoughts popped up:
1) Does this mean DA isn't reliable?
2) What other variables did we not consider?
3) Are there any takeaways from this?
I'd label this test as inconclusive for now because our samples were so low. A proper test in statistics should have at least 30 samples. We only had 4 and the ranges in our current site inventory weren't wide enough.
One thing I do think we should take notice of is the keyword that we scored highest for also had the lowest keyword difficulty, according to KWFinder.
Which might mean keyword difficulty is a bigger determinant for your ability to rank for a keyword...
But let's go in 1 step deeper and run a manual review of this keyword.
Analyzing Our First Place Keyword
First of all, every keyword difficulty tool is different but DA is already included in the keyword difficulty score for KWFinder.
So we know there's something more here for us...
Let's see what we can grab by analyzing the SERPs for Keyword 1. Maybe this will help us see why it received the best ranking:
The yellow results/pages represent unknowns - we do not know whether we can be can beat those results because they are either too difficult or just don't seem right for this keyword. Normally "best" type of keywords result in comparison articles, whereas this keywords first results goes directly to 1 product on an ecommerce site.
The pink is what we want to see more of - those are positions we know we can inch towards and have a better chance of taking the place of.
I would rather see more pink concentrated at the top of the SERPs because that would be an even better keyword.
SIDENOTE: Normally, I do not go for keywords that have more than 3 unknowns on the 1st page. This result has 5, however we moved forward with this keyword because the "info" articles barely fall in the unknown category. We know they are closer to resembling our "best" type of articles, rather than an eCom result - but they're still not exactly the same as our affiliate content. So they're an unknown.
Where We Stand Now
We aren't any more in love with DA than we were previously. DA is still a widely used metric, but I don't think it's as reliable anymore. Moz just simply has gone down hill in the recent years.
Their tools are quite slow and bulky. Plus, their crawlers and number crunching just aren't as active anymore because they grew too fast as a company. Downsizing in many departments.
But, that doesn't mean it's useless. I'd say 1/7 SEOs still use DA to some degree.
It is free after all. Of course, any free tool should be used to attract customers. Unfortunately, I don't think that the DA metric & Mozbar (chrome plugin) are doing a great job of.
For now, I'd say to use DA lightly and cautiously since it's a free tool. But place more emphasis on keyword tools - especially KWFinder.
Given that they've tweaked DA into their keyword difficulty, it seems to be a better judge of "rankability" for now.
I'd be extremely interested in seeing if KWFinder would add in their own "DA" styled metric but they're probably busy enough as it is. (#featurerequest!)
In any case, I hope you enjoyed this small test and I'm curious to hear about your own thoughts.
Let us know in the comments if you have any ideas for us to test!