Journal · Search

Domain authority score: what it means and what it costs you

It is a vendor's rough proxy for something Google really does keep: a judgement about your site. We checked three searches, and the highest score won only one of them.

Highland Willow · Boone, North Carolina · 31 October 2026

Domain authority is a number a great many business owners look up: check domain authority alone is searched about 9,900 times a month in the US, by Serpstat's count. Look yours up and the next question is obvious: is it good? This post answers that, then tests the assumption underneath it, that the site with the highest score wins. We looked up an authority score for every site on page one of three real searches. The highest score won one of them.

It is a companion to our post on PageRank, which covers the link score Google itself uses and works through what a lost position costs in money. This one is about the vendor scores, and about what Google keeps that they stand in for.

What the score is

Domain Authority is Moz's score, and it has become the generic name for a family of them. Moz describes it as a prediction of a site's ability to rank in search results. Since March 2019 it has come from a machine-learned model built mostly on link data, to which that version added a spam score and link quality patterns. Ahrefs's Domain Rating, Semrush's Authority Score and Serpstat's Domain Rank do a similar job with their own indexes and formulas; Semrush's also uses estimated search traffic. All run up to 100, and all are relative: Ahrefs points out that your score can fall without you losing a single link, because other sites gained some.

Relative also means the score can move when the vendor changes its method. When Moz rebuilt Domain Authority in 2019, its announcement told customers who report the number month to month that they may see a significant change. The ruler moved, whether or not anything about their sites had.

Does Google have a domain authority?

Google says it does not use Domain Authority, and has said so plainly for years. Gary Illyes wrote in 2016 that Google did not really have an overall domain authority. John Mueller has said Google does not use Moz's score in statements from 2016 to 2022, once calling it a metric from an SEO company, though in 2020 he also said Google does have some metrics that are more on a site level.

Two documents add to that.

In 2024 a set of Google's internal search API documentation was committed to a public GitHub repository in March and came to wide notice in May. Among thousands of named attributes was one called siteAuthority. Google confirmed the documents were genuine and cautioned against reading them out of context. They list what exists, not what is used or how much it weighs.

In 2025 the Justice Department published, as trial exhibit PXR0356 in United States v. Google, notes of a call with a long-serving Google search engineer. They describe a page quality score called Q* as incredibly important, largely static, and largely related to the site rather than the query, and they say a competitor who saw Google's logs would have a notion of authority for a given site. What feeds Q* is mostly redacted; the one input the notes name is PageRank.

Wikipedia's account of the leak reads it as contradicting Google's denials. Our reading is narrower, and it is a reading, not a finding. Google says it does not use Moz's number, and nothing here says it does. What the documents describe is Google keeping its own, largely site-level judgement of quality, which nobody outside Google can see. We find it useful to read any vendor's domain score as a rough outside proxy for that judgement, built mostly from links. The vendors do not describe their scores that way: Moz calls Domain Authority a prediction of ranking ability, and Ahrefs calls Domain Rating the strength of a backlink profile.

Does the highest score win? Three searches, measured

We took three searches that touch work we sell: jobber alternative, website audit service and seo roi calculator. For each we used Google US page-one results already on disk from our own market research, as Serpstat recorded them, and Serpstat's Domain Rank for every site on the page. Serpstat is the tool we pay for; Moz's numbers would differ.

How we picked them matters, so here it is. We chose the first two knowing that small sites ranked first in both, which is a reason to distrust them as a test. An independent review of this post found the third on disk, measured on 26 October for a different reason, and it goes the other way. One result in the first search listed the same Facebook post at positions 4 and 10, and one in the third listed the same page at 9 and 10; we counted each once.

In two of the three, first place was outscored by nearly everything below it. For jobber alternative, the top result scored 31, and 7 of the 8 results below it scored higher, among them Reddit at 86 and Facebook at 99. Jobber's own comparison page, the brand the searcher named, scored 54 and came 8th. For website audit service, the top result scored 16, and every one of the nine below it scored higher, including a Fiverr listing at 72 in 8th.

In the third, the highest score won. For seo roi calculator, first place went to an SEO software company scoring 60, the highest on the page; none of the 8 below it scored higher.

So the honest summary is: the highest score sometimes wins and sometimes does not, and three searches cannot say how often. With nine or ten results each, a rank correlation is too noisy to lean on, so we will not. What all three winners share is that they offer exactly what was searched for: an alternative to Jobber, a website audit for sale, an SEO ROI calculator. That fits the page mattering a great deal, alongside the score; it does not measure how much.

What a low score costs you

Less than the market for raising it suggests, and more than nothing.

It costs you on searches where everyone is strong. The score is a fair proxy for how many trusted sites point at you, and on broad, crowded searches where every result has years of links, that gap is real. Two of the first search's nine places went to Reddit and Facebook, which no small business outscores. On searches like that, a low score means fewer places you can take. And if Google's quality judgement is largely about the site rather than the query, as the court exhibit says, it follows your site into every search it appears in; that is our inference from one line, not something the exhibit spells out.

It costs you less on searches where the page decides. Our three searches were won at scores of 31, 16 and 60. The lowest scores anywhere on those first pages were 23, 16 and 13, so a site scoring 25 was above the floor in all three. Three searches cannot say why any site missing from page one was missing. Treat them as a reason to look at your page first, not as proof that the score does not matter.

Chasing the number has costs of its own. Links bought to raise a score break Google's spam policies when they pass ranking credit, and Google says sites that break those policies may rank lower or not appear at all. A score can also drop when the vendor updates its model or its index, so a report that celebrates the number month to month is partly measuring the ruler. Money spent moving a vendor's score is money not spent on the page.

What a position is worth in money is worked through, with a formula you can reuse, in the PageRank post. On its invented round numbers, page two against first place for one search term came to roughly $810 a month.

How to use the score properly

1. Pick the searches first. The five to ten things a customer types before buying from you. A score with no search attached answers nothing.
2. Run the floor test, as a rule of thumb. Look up the score of every site on page one for each search, in whichever tool you use, and find the lowest. If yours is at or above that floor, your score is probably not the main thing keeping you off the page, so look at the page first: does it say, in the searcher's words, the thing they asked for?
3. If you are far below the floor, expect links to be most of the gap. Earn links from places related to what you do: suppliers, customers, trade associations, local press. The PageRank post shows why one trusted link carries further than many from nowhere, and where that stops being true.
4. Read linking sites, not just the score. The count of sites linking to you is closer to the raw material, though it is still the vendor's count, from the vendor's index.
5. Ignore fixed tier charts, and do not buy links. The vendors' own guidance is to compare your score with similar sites; Semrush labels its scores from low authority to industry leader without attaching ranges. Bought links that pass ranking credit break Google's policies.

What this does not establish

• Three searches are three examples, and we chose two of them knowing who ranked first. They show the highest score does not always win; they cannot show how often it does. The vendors say their scores correlate with rankings; Ahrefs, for one, says so of Domain Rating.
• The score is Serpstat's. Moz's Domain Authority or Ahrefs's Domain Rating would give different numbers for the same sites, and we did not check whether they would order them differently.
• Positions are Serpstat's record of Google's results, and Serpstat does not say when it last checked Google. The scores for the first two searches were pulled up to a week after their positions. One snapshot each, no repeats.
• Google ranks pages and these scores rate whole domains. A strong domain with a weak page and a weak domain with a strong page are both in this data.
• We track these searches because they touch work we sell: website audits are one of our services, and we publish an SEO ROI calculator. The sites in them include businesses we compete with.
• We cannot see Google's own site quality judgement, so we cannot say how close any vendor's score comes to it. Nobody outside Google can.

Want help growing your online presence?

This is the kind of check we run for small businesses that want more of the right people to find them, with your own searches and your own competitors in it. Our site and page optimization work starts by finding exactly what is wrong with your website, ranked by what it costs you, before anybody spends money changing it. Search and content care then tracks your positions against named competitors month by month, checks term by term whether an AI answer sits above you, and reviews the sites that link to you.

We do not promise rankings, because nobody honestly can. If you want to know which of your searches are winnable on the page and which need links first, tell us what you sell and who you want to reach.

Questions people ask

What is a good domain authority score?

One that is close to or above the sites already ranking for the searches you want. The scores are relative, so no fixed number is good everywhere. A quick check is the lowest score on page one: in three searches we measured in October 2026, those floors were 23, 16 and 13.

Does Google use domain authority?

Not Moz's score or any other vendor's; Google has said so repeatedly. Google does keep its own judgement of quality, described in a 2025 court exhibit as largely about the site rather than the query. Vendor scores are best read as rough outside proxies for that, built mostly from links.

What is the difference between domain authority and page authority?

Domain authority rates a whole website. Page authority, Moz's companion score, rates a single page with the same kind of link data. Google ranks pages, so a strong page on a modest site can outrank a weak page on a famous one.

Why did my domain authority drop?

Usually for one of three reasons: you lost links, other sites gained links and pushed you down a relative scale, or the vendor updated its model or index. Only the first says anything about your site. Check your count of linking sites, itself the vendor's count, before worrying about the score.

How do I increase my domain authority?

Earn links from sites related to what you do that are themselves well linked: suppliers, customers, trade associations and local press. The score follows the links. Buying links that pass ranking credit breaks Google's spam policies, and Google says sites that break them may rank lower or not appear at all.

How do I get my website on the first page of Google?

Nobody can promise it, and anyone who does is guessing. Pick searches where page one already includes sites scoring near yours, make the page say exactly what the searcher asked for in their own words, and earn links from trusted sites in your trade. Then check your position monthly in Google Search Console, which is free, and adjust.

Sources

1. Moz. Moz to Launch Updated Domain Authority Score Model, press release, February 2019, as reprinted by MarTech Series. Moz's own site refuses our research crawler. https://martechseries.com/sales-marketing/programmatic-buying/moz-launch-updated-domain-authority-score-model/
2. Search Engine Journal. Domain Authority: Is It A Google Ranking Factor? (Googlers' statements, 2015 to 2022). https://www.searchenginejournal.com/ranking-factors/domain-authority/
3. Wikipedia. 2024 Google Search documentation leak, including Google's statement to The Verge of 29 May 2024. https://en.wikipedia.org/wiki/2024_Google_Search_documentation_leak
4. U.S. Department of Justice. United States v. Google LLC, 1:20-cv-03010-APM, trial exhibit PXR0356: notes of a February 18, 2025 call with a Google engineer, redacted public version. https://www.justice.gov/atr/media/1398871/dl
5. Ahrefs Help Center. What is Domain Rating (DR)?. https://help.ahrefs.com/en/articles/1409408-what-is-domain-rating-dr
6. Semrush Knowledge Base. Authority Score. https://www.semrush.com/kb/747-authority-score
7. Google Search Central. Spam policies for Google web search: link spam. https://developers.google.com/search/docs/essentials/spam-policies
8. Highland Willow. Google US page-one results for three searches as recorded by Serpstat, pulled 24 and 26 October 2026, and Serpstat Domain Rank for each ranking site, pulled 26 and 31 October 2026 through Serpstat's API.