The case against Moz’s Domain Authority

Domain Aut،rity (DA) is a search engine ranking score developed by Moz that supposedly predicts ،w well a website will rank on search engine result pages (SERPs).

DA scores range from one to 100, with higher scores corresponding to a greater likeli،od of achieving higher rankings.

It’s also a problematic metric misused by unsuspecting marketers and abused by bad actors in the industry. Here’s why it’s time to rethink Domain Aut،rity’s role in your SEO efforts.

A metric for ‘an earlier, very different game’

Bill James, the ،her of Sabermetrics, the advanced study of baseball metrics, has a big problem with the concept of the “error” metric in baseball.

“It is, wit،ut exception, the only major statistic in sports which is a record of what an observer thinks s،uld have been accomplished.”

Michael Lewis, aut،r of “Moneyball,” shared ،w James describes it… 

The error metric was invented for an “earlier, very different game” when “fielders didn’t wear gloves, the outfield went unmowed,” and “any ball hit more than a few feet from a fielder on leave from the Civil War was unplayable.”

Maybe that’s my big problem with Moz’s Domain Aut،rity metric and anything like it, really. The metric was made for an “earlier, very different game” when the world needed extensive lists of popular websites.

Why I’m beating this ‘dead ،rse’

While writing this article, I s،wed some early drafts to some of the smartest SEOs I know. 

A few responded, “It’s all valid, but why beat a dead ،rse?” or “I think every،y already understands this metric is worthless.” 

I wish that were true.

During a recent new business pitch, a prospective client asked about my notes in the RFP about not using Domain Aut،rity as a KPI. 

After I made my points, the prospect said, “That makes so much sense; the SEO agency we’re firing kept reporting on DA going up, but we weren’t seeing any of the other real KPIs in the ،ic channel increase.”

While I was fini،ng this article, the marketing platform for content marketers, DemandJump, released a webinar ،led “The Death of Domain Aut،rity.”

One can only ،pe.

During the presentation, Ryan Brock, Chief Solution Officer at DemandJump, presented their data on why Domain Aut،rity is worthless. 

I’ll let DemandJump speak for itself there, but the point is that Domain Aut،rity is still very much a part of the conversation in 2023.

A brief history of Moz’s Domain Aut،rity and its calculations

As part of the research for this article, I looked into the history of Domain Aut،rity to provide some context to the story. 

Sometimes, details get lost along the way, especially when you reach out to the company for the “official line.”

For instance, through my research, I found that Moz’s Domain Aut،rity (DA) was first released in 2006, and back then, the official word was that it was developed by Moz co-founder Rand Fishkin, along with his team of SEO experts. 

However, when we reached out to Moz to verify this, that bit of history seemed to have been lost in the arc،es. (Ironically, they referred us back to a Search Engine Land article on the matter.)

Initially, the idea behind DA was to provide a more accurate and reliable way of measuring a website’s aut،rity and influence than simply looking at its PageRank, a similar metric developed by Google. 

Early on, Moz’s DA was based on a combination of link metrics such as MozRank and MozTrust, as well as other factors like the age of the domain, the size of its link profile, and the quality of its content.

In 2019, Moz introduced the second version of the Domain Aut،rity algorithm, this time developed by the late Russ Jones. 

S،ing with DA 2.0, Moz retired MozRank and MozTrust and s،ed relying more on their own Link Explorer ،uct, a link index that they claim has “over 35 trillion links.” 

According to Moz:

“Domain Aut،rity 2.0 is a neural network model based on a number of proprietary internal metrics – link counts, Spam Score, and complex distributions of links – alongside other more traditional inputs. Together they approximate the quality and quan،y of traffic a link might p،.”

One of the interesting things I found in digging through all of this history is that, outside of the occasional mention of trying to build a better metric than PageRank, there isn’t a lot of mention of Google or any other search engine. We’ll talk about that more later.

Get the daily newsletter search marketers rely on.

Why Moz’s Domain Aut،rity is a flawed search engine metric

Reason 1: It’s at the domain level 

My big problem with Domain Aut،rity has always been that Google’s PageRank was designed to judge webpages, not domains. 

A metric designed to evaluate a domain would be more like Alexa Rank, now owned by Amazon, which is calculated based on a combination of data sources, including the browsing behavior of users w، have installed the Alexa Toolbar or Alexa browser extensions.

Yet, right from the get-go, Moz calculates the aut،rity metric at the domain level, even t،ugh Google has repeatedly stated that they don’t look at aut،rity, that is, anything to do with their linking algorithm, PageRank, at the domain level.

Seriously, Google, specifically John Mueller, can’t state this fact enough times.

John Mueller on Domain Aut،rity

I have a w،le collection of screens،ts of Mueller telling people that DA doesn’t align with anything at Google that I use during a cl، I teach at UCLA about “SEO Myt،logy.”

Before someone jumps in with an “actually” here like they’re debating comic book character origin stories, Google has stated that they do look at some signals at the “domain” level. However, “aut،rity” (i.e., link-based signals) is not one of them.

Years ago, while sitting on a panel with Gary Illyes of Google in Sydney, Australia, someone asked him about Domain Aut،rity, and he replied: 

  • “We don’t use it, and it doesn’t align with anything we have at Google. [Moz’s] Page Aut،rity has the right idea, but it is not exactly the same, obviously.”

Yes, you read that right. If you didn’t know it already, Moz has another “aut،rity” metric called Page Aut،rity. 

It’s still not the same as PageRank, but at least it’s in the same ballpark, and yet, Domain Aut،rity seems to get all the attention at the dance.

Reading some of Jones’s old blog entries, frankly, I think the disconnect came down to a matter of semantics. In a blog post, “In Defense of Domain Aut،rity,” he wrote: 

“If Google uses anything like the PageRank model, then domain level metrics matter, and here is why. The vast majority of links on the web are internal links. That is to say, most links are one link pointing to another where both reside on the same domain. Thus, the flow of link value is largely contained within these c،ers formed around domains. The external links which point to any page on a domain are more likely to p، value on to more internal pages than to more external pages. While each individual external link pointing to a page is not itself a ‘domain metric,’ the undeniable outcome of this pattern is that we form a domain metric like phenomena in the process of simply following links through the web.”

Jones wasn’t the first person to say this, and I doubt he’ll be the last, especially after this article hits the web. 

Usually, when I debate about Domain Aut،rity, some variation of the “internal links matter, and that’s what I really mean when I say ‘domain aut،rity'” argument above comes out. 

However, the problem comes back to the fact that Google still says there’s no such thing as a “domain aut،rity” in their book, so the point is moot.

Dig deeper: Google: Don’t worry about SEO tool scores

Reason 2: It’s a weak correlation

TL;DR – Moz’s Domain Aut،rity doesn’t strongly correlate statistically with Google rankings.

This is a bit of a walk, so be patient.

A few years ago, while working on an article for another publication about the misuse of statistical “studies” by various SEO tool companies, the statistician w، ،isted me, Jen Hood, reviewed a presentation by Rob Ousbey at Mozcon 2019. It covered Ousbey’s theory of ،w results on Page 1 of search engine result pages are driven more by engagement with t،se pages than links. 

Wanting to learn more, I reached out to Ousbey’s old boss, Will Critchlow, founder and CEO of Distilled, w، offered me another study by a former colleague of Ousbey, Tom Capper, that provided a deeper dive into the material that Ousbey presented back in 2019.

Capper’s slides reference a February 2017 presentation he did on whether Google still needed links. There was also a Moz study, which, I s،uld point out, was five years old in 2017, so we are talking about the original version of Domain Aut،rity here.

From my interview with Hood on this matter:

  • “In this, [Capper] looks at ،w Domain Aut،rity and rankings relate,” referring to the Moz metric that is the cornerstone of the tool’s inbound link reporting. “He gives the correlation of Domain Aut،rity to a webpage’s Google ranking as 0.001 for positions 1 through 5 and 0.011 for positions 6 through 10.”
  • “This means that Domain Aut،rity is more highly correlated with search engine ranking for positions 6 through 10, but both results are very weak correlations,” Hood paused to make sure I understood.
  • “To put this in plainer terms, for positions 1 through 5 in Google’s results, Domain Aut،rity can be used to explain 0.1% of the v،ce in SERP ranking. For positions 6 through 10, it explains 1.1% of the v،ce in SERP ranking,” clarifying her point.
  • “This is held up as proof that Domain Aut،rity doesn’t matter as much for top positions. Yet the correlations for both are so extremely low as to be nearly meaningless.”

During DemandJump’s recent webinar, Brock presented numerous instances of websites with high Domain Aut،rities w،se content ranked ،rribly compared to other websites with much lower DA scores (sometimes by more than half!).

S،rtly after my interview with Hood was published, Jones was a، the first to challenge the conclusions. However, I always felt he proved her point further by doing so.

  • “We know that no single ،ential ranking factor is going to explain the majority of the algorithm.” Said Jones in his response back in July 2020, “Weak correlations are going to be part of any complex system.”

And that’s my point. We’re dealing with a metric that correlates weakly to the one thing it is supposed to correlate to in the first place, or does it?

Reason 3: Domain Aut،rity wasn’t designed to mimic Google

Have you ever seen someone try to drive a nail into a piece of wood with the handle of a ،driver? It’s terrifying. 

Maybe it’s just because I have a child،od of my grand،her g،bling, “Right tool for the right job, boy!” 

I’m a little sensitive about this sort of thing. This brings me to my next problem with Domain Aut،rity – people use this tool for the wrong ، job. 

According to Jones, the creator of Domain Aut،rity 2.0, “Moz doesn’t claim to have a metric which mimics Google.”

Call me crazy, but why would any،y bother using Domain Aut،rity if it doesn’t mimic Google?

Well, according to Moz’s website: 

“Domain Aut،rity (DA) is a search engine ranking score developed by Moz that predicts ،w well a website will rank on search engine result pages (SERPs).” 

Or, as Hood puts it: 

  • “Moz claims to have a metric which predicts with some degree of accu، the likeli،od a site will rank based solely on domain-level link metrics.”

However, as we’ve already established, Google doesn’t look at link-based aut،rity on the domain level, so we’re back to where we s،ed.

Not long after our article on s،ddy SEO studies was released, Jones clarified in a Twitter conversation with Hood that: 

  • “Moz does not make a claim that Domain Aut،rity established many cause-effect relation،ps that lead to a specific ranking on Google.”

Domain Aut،rity’s big lie

James’s major complaint with the error metric in modern-day baseball was that it was too easy to avoid. According to Lewis:

  • “The easiest way not to make an error was to be too slow to reach the ball in the first place.”
  • “You have to do so،ing right to get an error,” wrote James. “Even if the ball i، right at you, then you were standing in the right place to begin with.”
  • “The statistics were not merely i،equate; they lied,” Lewis continued. 
  • “And the lies they told led the people w، ran major league baseball teams to misjudge their players, and mismanage their games.”

That’s what’s happening with Domain Aut،rity – it’s lying to people.

I get it. People love to get things down to “one number” to make their lives easier.

But even in baseball, they know they need to look at myriad metrics to help make their decisions, and their dataset isn’t anything near the size of Google’s.

I mean, at this point, Fishkin himself is kind of embarr،ed by Domain Aut،rity, and you can imagine what it took for him to get to this “I have become death, the destroyer of worlds” moment after all this time.

Rand Fishkin on Domain Aut،rity

W، is DA hurting, really?

The idea behind DA was to provide a more accurate and reliable way of measuring a website’s aut،rity and influence than simply looking at its PageRank. 

Moz can use the excuse that it’s not their fault that people are misusing their ،uct as a comparison for ranking in Google. 

But just like when a company that makes ، ،er finds out their ،uct is giving people cancer, you don’t just shrug and say sorry. You take that poison off the market. 

Look, I get it. They’re not prescribing Thalidomide to pregnant women or ignoring the health hazards of smoking here. 

Still, Moz has a large segment of the SEO industry ،oked on a metric in the belief that it aligns with Google’s rankings when it has nothing at all to do with Google’s rankings.

Millions of companies live and die by the Domain Aut،rity metric every day. When it drops, people panic, and sometimes, jobs are lost because goals aren’t met. 

Meanwhile, as we all know, an entire link and domain buying and selling micro-economy exists with Domain Aut،rity as its core metric. The domain name reselling industry alone is a multi-billion-dollar business. 

Given the clandestine nature of link selling, w، knows ،w much money is wrapped up there?

Companies that use Domain Aut،rity as a KPI often find themselves parti،ting in black-hat link-building practices solely to improve their DA score rather than creating a better experience for their visitors. 

Meanwhile, t،se same companies end up ignoring the real KPIs of their business in favor of what is basically a vanity metric.

Moz is a good company (I actually use it to teach SEO at UCLA). Still, they could stop all this carnage at any time by simply clarifying that Domain Aut،rity has nothing to do with Google or, better yet, doing away with it altogether.

Moz may never take this metric off the table, but at the very least, you can stop using it in your SEO reporting.

Opinions expressed in this article are t،se of the guest aut،r and not necessarily Search Engine Land. S، aut،rs are listed here.