Does Domain Authority Still Matter for GEO?

Domain authority has been the shorthand for “is this site worth a link from” for over a decade. So when a client asks whether a placement is worth pursuing for GEO, the DA score is usually the first thing they check. It’s a reasonable habit. It’s also the wrong question for a growing share of what actually drives AI citations.

Here’s what domain authority actually tells you, what it misses for GEO specifically, and what to check instead.

What domain authority actually measures

Domain authority (DA from Moz, or similar scores like Ahrefs’ Domain Rating) is a third-party metric built to estimate how well a site is likely to rank in Google search results. It’s calculated from the quantity and quality of backlinks pointing at a domain, modeled against how Google’s own ranking algorithm is believed to behave.

It was never a Google metric. It’s an outside approximation of one, built specifically to predict search rankings. That’s an important distinction, because it means DA was designed to model one system, not the systems doing the citing in GEO.

Why GEO doesn’t run on the same signal

LLMs don’t have a link graph in the way a search engine does. When a model generates an answer, it’s drawing on patterns learned from training data and, for tools with live retrieval, on what a search or fetch call turns up at query time. Neither of those processes runs a backlink-weighted authority score the way Google’s ranking system does.

What tends to matter instead is whether a source contains a clear, specific, extractable claim, and whether that claim shows up consistently across the sources a model has access to. A niche industry forum with a DA of 25 can carry more weight for a specific citation than a DA-80 general news site, if the forum thread is exactly the kind of firsthand, detailed content a model pulls from when someone asks a narrow question.

This is the same shift covered in our guide to link building for SEO vs GEO: the target moves from authority and anchor text to being present, accurately, in the sources a model actually reads.

Where DA still matters indirectly

None of this means DA is irrelevant. It correlates with a few things that do matter for GEO, just indirectly.

What DA correlates with	Why it still matters for GEO
Editorial standards and fact-checking	Models tend to weight sources with a track record of accuracy more heavily, and high-DA sites often have stronger editorial processes
Volume of existing coverage	A well-established, high-DA site is more likely to already be part of a model’s training data
Site stability and crawlability	High-DA sites are usually well-maintained and easy for crawlers to access, which retrieval-based tools depend on
General brand credibility	A placement on a recognizable, high-DA outlet still signals legitimacy to a human reader, even if it isn’t the deciding factor for a model

The correlation is real, but it’s a byproduct of what tends to come with high DA, not something a model checks directly. Treating DA as a proxy is fine. Treating it as the goal itself is where teams waste budget.

What to prioritize instead of chasing DA

A few checks matter more than the DA number when you’re evaluating whether a placement is worth pursuing for GEO:

1. Does the piece contain a specific, quotable claim? “Company X cut onboarding time in half” survives summarization. “Company X is a trusted leader” gets dropped.
2. Is this the kind of source a model in your category actually retrieves from? Review platforms, forums, and documentation sites often outperform general media for narrow, high-intent questions, as covered in our piece on how G2 reviews influence AI software recommendations.
3. Does the surrounding context support the claim? A single positive line surrounded by skepticism or caveats is weaker than the same line in a piece that consistently backs it up.
4. Is the topic narrow enough to be the clear answer to a specific prompt? Broad, generic coverage is easy to get and easy for a model to ignore. Narrow, specific coverage of a real use case is harder to get and far more likely to get cited.

A practical way to evaluate a placement without DA

Before chasing or turning down an opportunity based on its authority score, ask a simpler question: if a model were summarizing this page to answer a specific question a buyer might ask, would this page’s claim about your brand survive the summary? If the answer is yes, the placement is probably worth pursuing regardless of what the DA number says. If the answer is no, a higher DA score won’t fix it.

This doesn’t mean ignoring authority entirely when you’re planning outreach. It means using it as one input among several, rather than the filter that decides everything else. A PR and link building program built for GEO should be evaluating placements on citation potential first, with DA as a secondary signal, not the other way around.

The brands treating every high-DA placement as automatically valuable, and every low-DA one as automatically skippable, are optimizing for a metric that predicts a different system than the one deciding whether AI tools mention them at all.