On February 2nd, Ahrefs decided to redefine how they are calculating Domain Rating (DR) and much to everyone’s surprise, this new calculation resulted in most domains that currently have a DR 30–50 rating will drop to zero because “they don’t deserve the “medium” DR that they have in the first place.”
Many of Ahrefs users were not so receptive to the news, some even offended, but most were simply left confused about how to adapt to the new scoring system we grew to love. After all, Ahrefs is an SEO tool and when a primary metric for success is changed this drastically, we no longer know what “quality” means. We now have to sell our clients and superiors on a new goal and retrain our teams on what DR ranges are worthwhile to pursue backlinks from.
Why Did Ahrefs Do This?
Did you know that in the US in 2013, the top 10% of families held 76% of the wealth, while the bottom 50% of families held 1%? In the same way, the statistical reality Ahrefs is facing is stacked towards a small percentage of domains in their index.
In Ahrefs index, less than 1% of domains made up the DR 20-100 range, while 99.4% of domains made up the DR 0-20 range.
In their blog post announcing the change, they show the difference in DR levels as shown below:
But, when you take that chart data and put it in a format that’s easier to visually understand, you get:
Anything above DR 40 isn’t even visually represented. That’s just how many domains there are on the internet that cannot compare to high DR domains. But, it makes complete sense. Here’s why.
If your domain is in the majority, it means your DR is less than 20. Not because it’s low quality or because you have links from low-quality sources. It just means the vast majority of the internet isn’t made up of massive domains like Google or relatively moderate domains like The Onion. Big surprise.
In Ahrefs’ defense, it doesn’t make sense to say that simple domains with less than 1000 referring domains are a DR 50, while Google, who is 12,000x larger, has a DR100.
So, what’s Ahrefs to do? Expand DR to 1000 so you can feel good about your DR 40, or rework their calculations to give their users an honest depiction of the web?
I support the latter, but it doesn’t come without some frustration that hopefully, Ahrefs can resolve.
Complaints that are not Ahrefs’ problem
Who are you to say I’m a DR 0?!
I knew it was a stigma that people are thin-skinned these days, but damn…
You need to remember, this is a universal change that affects all domains, so it’s all relative. Your competitors backlink profiles are mostly made up of sub-DR 20 links now too, so the playing field is still level, it just means your new target DR is DR 3-10 instead of DR 45 (as an example).
And yes, the reality is that your domain compared to the best on the internet is relatively insignificant. This isn’t a bad thing, it’s just how it works.
The keyword I want to rank for has DR 0 domains ranking above my DR 5 domain.
It wasn’t uncommon to see lower DR’s outranking higher DR’s before either. DR is correlated with keyword rankings because backlinks are too. You can’t ignore the other SEO strategies and expect link building to carry you through entirely (and that’s coming from us, a link building service).
We have links from Forbes, INC, and similar domains, why the DR drop?
It’s a big fat unfortunate lie in the SEO industry that links from massive publications like Forbes, Entrepreneur, and the like, are sure-shot ways to ranking success. When in fact they are usually nofollow and Google is quite aware of the usual suspects that sell or fall victim to forced links.
Aside from that, the reason why your domain dropped despite that stellar link profile is because, in the grand scheme of things, your domain is still tiny compared to the backlink profiles of the high DR domains. Plus, Ahrefs doesn’t count links that are nofollow in their formula.
If you’re asking, “Why would you compare my DR 0 to a DR 100?”, the answer is because Ahrefs has to if they are going to give you an accurate depiction of domain quality across the internet. Everything is relative, so you cannot expect to objectively understand how “good” your domain is unless you’re comparing it against the best and worst that Ahrefs’ index has to offer.
What Ahrefs needs to correct
How do the new DR ranges correlate with the old?
From the research we do as part of our backlink audits, this is how the old DR ranges compare to the new, but we’d really love it if Ahrefs answered this question more thoroughly. We used the same backlink data for each version of the DR update. The original data, 1.0 is what’s been used for the last several years. 2.0 is the initial big switch. 3.0 is the finetuning from the big switch that was taken 2/7/2018. Let’s compare the three!
First, let’s look at where the bulk of the links lie in each version, according to the backlink data we have for the top 8 results for “Active Wear”. (note: none of these are Multifuse clients.)
Notice where the heavy green cells are. This indicates where the bulk of the links for each 8 domains lie for a given 5 point DR increment.
If we simplify this even more, we get the following comparison which shows the 1.0, 2.0, and 3.0 data side by side:
But ultimately, how does DR 1.0 compare to 2.0 or 3.0? What did the 3.0 revision to DR solve?
Eh, the update didn’t solve much, but my test is also only looking at 7000ish links. Ahrefs, please answer this in a blog post.
It seems Ahrefs adjusted their calculations to spread domains out more evenly across the lower DR tiers, but it’s still heavily weighted at the lower end of the DR spectrum. This is likely for the purpose of easing concerns among those that just want to see their site have a higher DR number. It’s a little illogical, but hey, I can’t completely blame Ahrefs for tweaking their calculations for those that just want to see a pretty number. As long as the playing field is still level, all is well.
You can see our method of finding this data here.
How helpful is DR for research when 99% of domains are within 5-10 points of each other?
This is perhaps our biggest issue as SEOs. It’s important that Ahrefs truly leveled the playing field for the sake of accurate data representation instead of simply continuing to give you a “feel good” number to take back to your boss, but now 99.4% of us are left researching domains that don’t have much differentiation in terms of DR.
How do we know if a domain is brand new with nothing to offer, or 10 years old with plenty of backlinks? Obviously, we can look at the backlinks, but before we relied on DR to give us this quick indication.
It’s almost as if DR needs to split into two indexes, separating out the DR 0-20’s into one scoring system and the DR 21-100’s into another. Admittedly, this seems like an awful idea, but it would help with researching the domains that best relate to yours to avoid high DR domains from skewing your results.
Or, have DR go up to 1000 so that there are more point levels to differentiate 99.4% of the domains that are DR 0-20. It would be easier to compare domain value with 0-200 points rather than the current 0-20.
Why does Ahrefs not count nofollowed links?
Not counting nofollowed links seems short-sighted. It’s been shown that nofollowed links are valuable, whether from a link equity standpoint or simply being a part of a natural link profile. They should have some value as part of your formula, albeit less.
Why does Ahrefs only count one link from a domain?
Similar to the nofollow link situation, this also seems short-sighted. The value from links on different pages from the same domain are surely there, but I’m sure there’s diminishing value as that count increases.
Ahrefs’ flagship score needs to look at more than just backlinks.
When you look at why users are paying for Ahrefs, it’s primarily for backlink data and competitive analysis needs. The primary indicator of quality is DR, but it seems that DR (or the competitive equivalents) isn’t fulfilling the needs of most users if it’s exclusively looking at backlink information.
The flagship domain score should also loop in metrics relating to domain legitimacy, like ranked keywords and estimated traffic data, backlinks from identified spammy sources, backlinks using clearly manipulated anchors, and even some indication of whether or not it’s suffering from a Google penalty (or simply being filtered out of the first page).
We know Google looks at more than just link equity when assigning the value to the target, so why would your flagship domain score not consider that? This data is already being indexed by Ahrefs, so Y U NO USE IT?!
What Ahrefs is Getting Right
Product Enhancements & Accurate Data
You cannot progress as a company unless you recognize the faults in your systems. For Domain Rating (and currently Moz, Majestic, and SEMrush), the flaw was showing a domain with 1000 backlinks as DR 40, while a DR 90 domain has several million. That’s not keeping the playing field level, so kudos to Ahrefs for correcting it despite the expected outcry, even if it does cause some minor problems.
The main issue for SEO’s understanding what the new DR equivalents are. I refer to my images above that show this. All of the domains from the past and future will be just as effective for you as before, but with a different score.
You can’t discredit a company for making changes for the sake of accurate data when that’s the reason you use them.
Sheer Index Size & Crawler Activity
No crawler (including Google) will be able to index the entire internet, so he who has the most data is the most valuable source for competitive analysis.
A comparison between Ahrefs and the others would be helpful, but not all disclose similar data points. The ones that do (like Ahrefs and Moz) are helpful but don’t always align with one another. Instead, we decided to test which crawler had the most data for 5 domains of different sizes. Ahrefs was usually the victor because they simply have more data to work with. This makes sense considering they are the second most active crawler on the web (just after Google).
Personally, I’m excited about the new method of calculating DR, but something needs to be done to help differentiate a domain that’s in the DR 0-20 range. Introducing a new score to understand how natural and legitimate a domain is would also round out the product for SEO needs.
Many are still going to say this Ahrefs change doesn’t make sense because they now see their domain as a lower DR score, but in my possible unpopular opinion, if the domain score you’re using isn’t leveling the playing field to give you real perspective of your domain’s authority compared to other domains on the web, they are providing skewed data right from the start.
are your backlinks justified?
Meet Dale to discuss how our powerful link building services can transform your organic website traffic this year.
we want to give you a free backlink
it's on the house
Our free trial will show you how we work, the valuable publishers that we use and our approach to natural, ethical link building.
It almost seems like they don’t understand their own scale. They claim it’s logarithmic, but then say, “the vast majority of sites that currently have DR 30–50 will drop to nothing because, in fact, they don’t deserve to have the ‘medium’ DR that they currently have.” But that doesn’t describe a logarithmic scale at all.
In a logarithmic scale like the Richter Scale a 5.0 earthquake is 10x greater than a 4.0. Using a logarithmic scale, like what was used with the original PageRank, a DR of 87 would actually be 794x more authoritative than website with a DR of 58.
So under their old scale a DR of 30-50 was never considered “medium” and I doubt anyone with a DR or 50 seriously thought they had half the backlinks of Google or Facebook. The team at Ahrefs seems very confused.
Admittedly, I’m no data scientist, but I trust Ahrefs had enough of them in a room to ask these kinds of questions. My thought goes back to just how wide of a gap a current DR 0 is compared to a DR 100. It can still be logarithmic, but since the DR 100 are just so massive, it’s hard to visualize or comprehend.
How are you getting a DR 87 being 794x larger than a DR 58?
On a 1 to 10 logarithmic scale like PageRank, the number shown is actually the power of 10. So a PR of 5.8 represents 10^5.8 and a PR of 8.7 represents 10^8.7, which is 794x bigger: (10^8.7)/(10^5.8)=794. It also means that a PR 10 website has 1 billion times more authority than a PR 1.
I’m just surprised as I think this disproportionately hurts smaller websites. The harder something is to measure, the harder it is to improve and see results from your efforts. Since their target customers are likely SMBs and not Google or Facebook, I’m surprised they would roll out any change that makes it harder for smaller websites to gain a more precise measure of their link-building efforts.
The simplicity that is a true logarithmic scale (like you describe) seems like it may have a lot of mistakes too. I’m right there with you in that there’s things I don’t understand about their calculation, but I’d trust that they thought them through for the sake of an accurate representation. I’m sure they knew full well how the reaction would be, so they must have strong logic in pushing this update through.
The research part does hurt though and is one thing in my blog post I ask for their guidance on.
So bad, My domain DR dropped 🙁
Yes, but so did everyone’s DR. When it’s all relative, your new “lower” score means nothing.
Hi. awesome post explaining things.
But I’m not sure you took into consideration changes that took place 24-48 hours later after DR update.
After first DR update our website dropped from 54 to scary 29 DR, but 24-48 hours later it ended up even higher on 59 (this includes our competition as well).
I guess DR drop after second quick update, isn’t really applied to absolutely everyone – as my example shows, some websites actually proven to be better than they were originally according to new(new) ahrefs DR.
This is a fair point. I’ll work on updating the charts to see what changes on a larger scale!
Hey Maciek, our blog post has been updated. Thank you!
The recent Authority Hacker podcast episode (#95) has Tim Soulo from AHRefs talking abou this very issue.
I’d recommend listening if one is interested in the subject.
Comparing competition of niche sites with authority sites is now pain in ass. Earlier no of domains pointing to url, url rank combined with DR was good way to judge competition. was used to earlier matrices. Been using them from years.