What is domain authority, and who produces it?
Domain Authority is a score, usually expressed from one to a hundred, produced by a search-marketing software company as a prediction of how likely a domain is to rank well. Competing vendors sell similar scores under their own names, each computed by its own model from its own crawl of the web.
The important structural fact is in that description. These are third-party estimates, built by companies that do not have access to Google's ranking systems, inferring a general propensity to rank from signals they can observe from outside — principally the pattern of links pointing at a domain.
That is a legitimate thing to build, and the vendors are generally candid that it is what they have built. A prediction of ranking likelihood is genuinely useful for comparing link-building targets at scale, which is the job it was designed for.
The distortion happens in transmission. A predictive score, produced by a vendor, about a domain's general tendency, becomes in ordinary conversation a property of the outlet — its authority, a number that placement can transfer to a buyer. Each of those steps loses something, and the last one gives away the whole thing: nothing about a publication's score is conferred on a company written about by it.
Does Google use it?
Google publishes a guide to the ranking systems it uses in Search. It names seventeen: BERT, crisis information systems, deduplication systems, the exact match domain system, freshness systems, link analysis systems and PageRank, local news systems, MUM, neural matching, original content systems, removal-based demotion systems, the passage ranking system, RankBrain, reliable information systems, the reviews system, the site diversity system, and spam detection systems. It also lists four retired systems, including Panda and Penguin.
Domain authority is not among them. Neither is any equivalent site-wide authority score, under that or any other name. The guide contains no reference to such a metric at all.
That absence is the primary evidence, and it is worth preferring to the alternative. Google representatives have said in public remarks, widely quoted in the search trade press, that Google does not use Domain Authority and that it is a metric from an SEO company. Those remarks are consistent with the documentation, but they reach a reader through second-hand reporting, and a published Google document is the better thing to rely on.
What follows is narrow and sufficient. A buyer told that a placement carries weight because the outlet has a high domain authority is being told something about a vendor's model, not about a system Google has described.
Presence in Google's published ranking systems guide
Source: Google Search Central, guide to Google Search ranking systems.
Bars encode whether the guide names the item, not its importance. Google does not publish the relative weight of any system.
But doesn't Google judge sites as a whole?
Yes, and this is where an honest account has to slow down, because the simple version — Google has no site-level view — is not what the documentation says.
Google's guide states: “Our ranking systems are designed to work on the page level, using a variety of signals and systems to understand how to rank individual pages. Site-wide signals and classifiers are also used and contribute to our understanding of pages.” Site-level information plainly exists and plainly matters.
The guide then constrains what that means: “Having some good site-wide signals does not mean that all content from a site will always rank highly, just as having some poor site-wide signals does not mean all the content from a site will rank poorly.”
So the accurate position has three parts. Google uses site-wide signals; Google publishes no single number summarising them; and those signals do not determine the fate of individual pages. The vendor score compresses all of that into one figure, which is why the score can be directionally informative and still misdescribe the mechanism. A placement buyer reasoning from one number is reasoning about a simplification of a system whose actual shape its operator has described in different terms entirely.
Why does the metric persist in media buying?
Because it solves a real problem badly, and nothing better is publicly available. Anyone choosing between outlets needs some way to rank them, and a single comparable number across every domain is enormously convenient. Editorial judgement does not scale into a price list; a score does.
It also serves a commercial function. A placement priced against a high score looks justified, and the score is verifiable in the sense that a buyer can look it up and see the same figure. Verifiable and meaningful are different properties, and the first is doing most of the work in that transaction.
The clearest evidence that these scores are models rather than measurements is that they disagree. Different vendors publish materially different numbers for the same domain, and both numbers are correct outputs of their respective models. If either were reading a value out of Google, they would converge.
None of which makes the scores worthless. As a rough comparator between link targets — the use they were designed for — they carry information. The failure is of transfer: using a prediction about one domain's ranking tendency as a proxy for what an article on that domain will do for somebody else.
What should guide outlet selection instead?
Questions with checkable answers. Does the outlet's audience contain the people you are trying to reach? That is answerable from the publication's own readership and subject matter, and it is the question a score cannot address at all, because a general-purpose ranking prediction knows nothing about fit.
Is the coverage the kind that survives scrutiny elsewhere? A placement's usefulness is decided by different rules in different places, and none of those rules mention any vendor score. Wikipedia's guideline asks whether the source is independent in authorship and content. Immigration adjudicators ask whether material is about the person and appeared in professional or major media. Search engines discount links that were paid for regardless of where they sit.
Will the coverage remain findable and citable later? Durability is a real property of an outlet and is entirely absent from an authority score.
And what is the label? Whether a piece is presented as editorial or as sponsored changes how every downstream system treats it, and that is a fact about the specific article rather than about the domain it appears on.
What this means for anyone buying placement
Treat a domain authority figure as what it is: a third-party vendor's prediction about a domain's general ranking tendency, useful for coarse comparison and not a property that transfers to you.
Be direct with any seller who prices against it. Ask what the number measures, who calculates it, and where Google says it is used. A seller who cannot answer the third question is pricing against a proxy, and the honest version of that conversation is available: this outlet tends to rank well, which is a reason to think a page there may be findable, and is not a ranking benefit conferred on your company.
Watch particularly for the combination of a high score and a paid or sponsored label. Where a placement is disclosed as paid, its links are supposed to be marked so that they do not pass ranking credit, which means the domain's ranking strength is precisely the thing the arrangement is designed not to transfer. A high score attached to such a placement is decorative.
Media Morpheus sells placement and publishes its prices. What a placement buys is an article on a specific publication, with a specific audience and a specific label, which stays findable for as long as that publisher keeps it up. It does not buy a share of that domain's authority, because there is no such quantity to buy.
Methodology and limitations
This report relies on Google's own published guide to its Search ranking systems, read both for the systems it names and for the site-level language it uses. Four limits apply. First, the absence of a domain authority metric from Google's documentation is evidence about what Google has published, not proof about the internals of a system Google does not fully describe; Google states plainly that site-wide signals exist. Second, public remarks by Google representatives on this subject are referred to here as reported speech, because they reached us through search trade publications rather than as primary documents, and they are not relied on for any finding. Third, no vendor's scoring methodology is examined or evaluated here, and nothing in this report asserts that any such score is inaccurate at the job it was built for. Fourth, no correlation data between vendor scores and ranking outcomes is cited, because published studies on that question are produced largely by the vendors whose metrics are under examination.
Conclusion
Domain authority is a useful tool wearing a misleading name. It is a vendor's prediction of how a domain tends to rank, built from outside Google by companies with no access to the systems they are modelling, and it is sold onward as though it were a property of the publication and a benefit conferred on whoever appears there. Google's published guide names seventeen ranking systems and describes site-wide signals in terms that explicitly refuse the idea of a single summary score. The number can still be worth glancing at when comparing targets. It should not be the reason a placement is priced, and it answers none of the questions — audience fit, labelling, independence, durability — that determine whether coverage is worth having.
Frequently Asked Questions
Does Google use Domain Authority as a ranking factor?
Google's published guide to its ranking systems names seventeen systems and contains no reference to domain authority or any site-wide authority score. Domain Authority is a third-party vendor metric.
What is Domain Authority, then?
A proprietary score produced by an SEO software company predicting how likely a domain is to rank well, computed from that vendor's own crawl and model.
Does Google look at sites as a whole?
Yes. Google states that “site-wide signals and classifiers are also used and contribute to our understanding of pages”, while adding that good site-wide signals do not mean all content from a site will rank highly.
Why do different tools give different scores for the same site?
Because each is the output of a different proprietary model rather than a reading of a value inside Google. If they were measuring the same underlying quantity they would converge.
Should domain authority guide which outlets I pitch?
Only coarsely. It says nothing about audience fit, how the piece will be labelled, whether the coverage counts as independent elsewhere, or how long it stays findable.
Does a placement transfer a site's authority to my company?
No. The score is a prediction about that domain's own ranking tendency, and paid placements are separately expected to mark their links so they do not pass ranking credit at all.
Sources
- official Google Search Central. A guide to Google Search ranking systems, including the list of current and retired systems and the description of page-level and site-wide signals.
- trade press Search trade publications reporting public remarks by Google representatives on Domain Authority. Cited as reported speech only; no finding rests on it.