What Is Semrush's Authority Score?

Authority Score is calculated by Semrush, the SEO software company founded in 2008 and now listed on the New York Stock Exchange (NYSE). The figure draws on several data sources the company owns outright, none of them Google's: a link index built by Semrush's own crawler, SemrushBot; a keyword database Semrush uses to estimate every domain's organic traffic; and that same link index, analyzed a second time to flag spam patterns.

The keyword database is what separates Semrush from a pure link vendor. According to its official documentation, the organic traffic that feeds into AS is estimated "based on a website's total keyword positions and the click potential of every position," not from real analytics data supplied by the domain owner. It's an estimate built on Semrush's own rank tracking, carrying the same margin of error as any third-party traffic estimate.

Semrush recalculates Authority Score for every domain once every two weeks, per its own knowledge base. A recent shift in a domain's link profile or organic positions can take up to fourteen days to show up in the public number.

None of the three data sources comes from Google. The search engine doesn't share its own link index or internal ranking data with third parties, so any tool offering an authority metric, Semrush included, necessarily builds it on its own crawling infrastructure. That's true of Moz's Domain Authority and Ahrefs' Domain Rating just as much as it is of Semrush.

Semrush documents in reasonable detail what goes into the calculation, though not exactly how it's combined. According to its official knowledge base, AS is built from eight weighted signals, grouped into three categories.

The first category is Link Power, a single signal "calculated by the amount of follow referring domains, along with the authority of those referring domains." The second is estimated organic traffic, also a single signal, drawn from the keyword database described above. The third, and largest, category is what Semrush calls Natural Profile: six signals that function as spam indicators. Among them, no organic rankings in Google's top 100 results, an unnaturally high percentage of dofollow referring domains, an imbalance between link volume and organic traffic, too many referring domains sharing the same IP address or IP network, and the presence of another domain with a nearly identical backlink profile.

Semrush describes combining these eight signals as a complex algorithm backed by artificial intelligence and machine learning. That's where the transparency ends: the company publishes what it measures, not how much each piece weighs. No official documentation shows a formula along the lines of "40% links + 40% traffic + 20% spam," no coefficient table, no explanation of how the six spam signals get weighed against each other before points come off the final score.

That opacity around weighting isn't unique to Semrush: Moz doesn't publish the exact coefficients of its machine learning model either. What sets Semrush apart is the number of distinct signal types it says it combines, eight across three categories, against a single signal type for Ahrefs.

Semrush also notes that AS runs on a relative scale, not an absolute one: a domain's result also depends on how every other domain in the index is moving. If a site's link profile and traffic stay exactly the same but its competitors improve, its AS can drop even though nothing changed on the site itself.

The costliest misreading is treating Authority Score as a Google figure. It isn't. Google has repeatedly denied keeping an internal domain authority score: in March 2018, John Mueller, then Google's Search Advocate, answered in a Reddit AMA that Google has "nothing like a website authority score," as recorded in Ahrefs' support documentation. Verification note: the original Reddit thread was blocked to automated retrieval, so this quote rests on that secondary source, not on the primary one. No third-party metric, whether from Moz, Ahrefs, or Semrush, feeds into Google's ranking algorithm. Semrush itself never claims otherwise anywhere in its own documentation.

The second misreading is treating AS as a pure link metric, equivalent to Domain Rating or Domain Authority. It isn't, and this is the core methodological point of this page: of the eight signals Semrush combines, only one directly measures inbound link strength. The other seven capture estimated organic traffic and link-profile health. Treating an Authority Score of 45 as equal to a Domain Rating of 45 is a basic error, because AS already folds in information DR doesn't consider at all.

The third misreading is reading AS itself as a traffic measurement. It isn't: estimated organic traffic is one of eight weighted signals, not the final output. A domain can carry high organic traffic and a mediocre Authority Score if its link profile is weak or trips one of the six spam signals. For a domain's actual traffic, the right source is an analytics tool with access to the site's own data, not AS.

Semrush is explicit about the score's intended use: "Authority Score is best used for domain comparison, and not for determining good/bad on an absolute scale," according to its official knowledge base. It works for ranking a list of domains against each other, for instance when prioritizing link building opportunities, as long as the comparison uses the same tool at similar points in time.

Because more signal types feed into it than a pure link metric, AS in theory offers a rounder picture of a domain's overall health: how many links exist, whether they translate into real traffic, and whether the profile shows signs of manipulation. That's the advantage over Domain Rating, which by Ahrefs' own description is "purely link-based" and doesn't filter for quality beyond the link graph itself.

The trade-off of combining more signals is losing traceability. When a domain's Authority Score rises or falls, Semrush doesn't publicly break down which of the eight signals caused the move: it could be a loss of referring domains, a dip in estimated traffic, or a triggered spam indicator. With a pure link metric like Domain Rating, the range of possible causes is narrower, so it's easier to trace why the number moved.

Semrush also flags a different, less technical limit: "the impact of a link also strongly depends on your website's niche," per the same documentation. An AS of 40 can represent a solid profile in a small, low-competition niche and a mediocre one in a link-saturated market like finance or travel. The company explicitly advises against using AS as the sole criterion for evaluating a domain's performance.

Authority Score gets confused constantly with Moz's Domain Authority and Ahrefs' Domain Rating. All three run on a 0-to-100 scale, all three get presented as a domain's "authority," and all three often show up side by side in the same agency report. The confusion is understandable; the equivalence isn't real.

Each tool crawls its own link index, which alone produces different numbers for the same domain. But the bigger reason is that all three measure different things by design. Domain Rating is, by Ahrefs' own description, "purely link-based": it distributes a kind of "link juice" between domains and doesn't factor in traffic or its own spam signals. Domain Authority draws almost entirely on link data (root domains, anchor text distribution, its own spam score) processed through a machine learning model, also without organic traffic in the formula. Authority Score is the only one of the three that folds in estimated organic traffic as a full-fledged signal alongside links and spam.

That's the real reason the same domain shows three different numbers, and why the gaps don't always point the same way: a site with plenty of links but weak traffic can carry a high DR and a lower AS; a site with fewer links but solid traffic and a clean profile can land the other way around. None of the three numbers is wrong, each answers a slightly different question, with its own data and its own formula. Averaging them into some "average authority" makes no methodological sense: that would mean blending two pure link measurements with one that also factors in traffic.

Semrush didn't start as an authority tool either. Founders Oleg Shchegolev and Dmitry Melnikov first built Seodigger in 2006, a tool for their own use that checked Google rankings and the keywords driving traffic to a site. In 2007 it became SEOquake, a free browser extension offering a quick look at keyword and domain data. On August 3, 2008, Semrush officially launched as its own platform, starting with just two tools built around spotting keyword gaps.

With those two tools and a handful of friends as early customers, Semrush built out the rest of today's suite piece by piece: competitor analysis, site auditing, rank tracking. The company grew fast, passing one million registered users by 2012 and going public in 2021.

That origin in keywords and rankings, rather than links, matters for understanding today's Authority Score: unlike Moz or Ahrefs, which both started as link-analysis tools, Semrush began as a keyword and competitive-research tool. That same in-house ranking database is what now feeds the estimated organic traffic component inside AS.

Semrush officially launched its AI Toolkit on March 4, 2025, according to the press release on its own site. Eugene Levin, Semrush's President, put it this way: "AI-driven search is not going anywhere and is changing how brands are discovered. AI Toolkit is the first step on equipping businesses with the intelligence they need to understand where they stand, how they compare, and how to turn AI-driven search into a real growth opportunity." In the same statement, he added: "Understanding how AI portrays a business is no longer optional."

The toolkit measures brand mentions, sentiment, and share of voice across ChatGPT, Perplexity, Gemini, Copilot, and Google AI Overviews, built from three pieces: prompt tracking, brand performance, and a visibility overview. Through 2025 and into early 2026, Semrush folded the AI Toolkit into a broader bundle called Semrush One, while continuing to offer it as a standalone product.

Per Levin, most companies had no way of knowing how they showed up in AI-generated results before this launch, a shift he frames as redefining how brands get discovered and positioned before a user ever visits the site itself.

Of the three metrics covered in this series, Authority Score comes closest to a full picture of a domain's health, precisely because it inherited Semrush's DNA as a keyword tool: estimated organic traffic wasn't bolted on as an afterthought, it comes from the same ranking database the company has been building since SEOquake. That gives Semrush a real edge in connecting links to actual traffic, something neither Moz nor Ahrefs offers inside their own authority metric.

The cost of combining more signals is traceability: when a domain's AS rises or falls, Semrush doesn't disclose which of the eight signals caused it, so tracing the actual cause takes more work than with a pure link metric like DR. The two-week update cycle also means a recent change can take up to fourteen days to show up, and the algorithm's exact weighting, which Semrush itself describes as backed by AI and machine learning, stays undisclosed too.

Like DA and DR, a high AS is no substitute for actually checking whether a domain's content is trustworthy, a point Semrush itself concedes when it warns against using AS as the sole criterion.