Backlinks vs AI Citations: Which One Actually Predicts Traffic

“More backlinks equals higher rankings equals more traffic” was close enough to true for two decades that it became close to unquestioned SEO orthodoxy. What’s happening now with AI-generated answer surfaces is a genuinely different question: what predicts being cited by an AI system isn’t necessarily the same thing as what predicts ranking well in the classic algorithm, and treating them as interchangeable is starting to cost people traffic they thought their link profile had already secured.

What Ahrefs Actually Found

Ahrefs published a study in August 2025 analyzing 75,000 brands to measure which factors correlate with visibility inside Google’s AI Overviews. The ranked correlation results were stark, and off-site brand signals dominated the top of the list rather than classic link metrics:

• Branded web mentions: 0.664
• Branded anchor text: 0.527
• Branded search volume: 0.392
• Domain Rating: 0.326
• Number of referring domains: 0.295
• Branded traffic: 0.274
• Raw backlink count: 0.218
• Ad traffic: 0.216
• Ad spend: 0.215
• URL Rating: 0.18
• Number of indexed site pages: 0.17

Brand mentions correlated roughly three times more strongly than raw backlink count. Brands in the top mention-volume quartile earned up to 10x more AI Overview mentions than the quartile directly below them, and 26% of the brands studied had zero mentions inside AI Overviews at all.

Why an Unlinked Mention Can Outperform a Backlink Here

Classic ranking algorithms are built substantially around the link graph — who links to whom, with what anchor text, from what authority. The systems generating AI Overview and AI Mode answers appear to weight something closer to a reputation and frequency signal: how often and how consistently a brand or entity gets talked about across the open web, including plain-text mentions with no hyperlink attached at all. That’s a plausible mechanical explanation for the gap in Ahrefs’ numbers, but it’s worth being precise that it’s an inference from the correlation data rather than a confirmed detail of how Google’s systems actually work — Google hasn’t published the underlying mechanics publicly.

The Caveat Ahrefs Itself Flagged

Ahrefs’ own researchers were explicit that correlation isn’t causation here. Brands with strong cross-platform presence — an active YouTube channel, regular press coverage, a healthy footprint on review sites — tend to have both high mention volume and high AI visibility at the same time. That doesn’t prove that starting a YouTube channel causes AI citations; it shows the two travel together, likely because both stem from the same underlying brand strength rather than one directly producing the other.

Separately, reported organic click-through rate averaged 2.07% for brands appearing inside an AI Overview, versus 0.94% for brands that didn’t. Worth treating as directional rather than a precise universal figure — CTR studies like this vary considerably by query type and industry, and this specific comparison came through the same body of research rather than a controlled experiment.

What This Actually Changes About the Work

Backlinks haven’t become worthless. Domain Rating and referring domain count still show a real, if smaller, correlation — 0.326 and 0.295 respectively — which suggests link building still functions as something closer to a threshold signal rather than a wasted effort. What’s shifting is where the next dollar or hour of effort produces more return: PR coverage, product reviews, YouTube presence, podcast mentions, and other off-site brand signals that never carried a hyperlink are now competing directly with link acquisition for budget that used to go there by default.

In practice, that means treating digital PR, review generation, and being mentioned by name on forums, comparison sites, and third-party YouTube content as measurable channels in their own right — not just a byproduct of a good product — and tracking unlinked brand mentions the same deliberately as a link-building campaign would track new referring domains. Most rank-tracking tools weren’t built for that; brand-monitoring tools built for reputation management are closer to the right instrument, even though they weren’t originally designed with SEO in mind.

A backlink was always a proxy for “other sites think this is worth pointing to” — an unlinked brand mention is turning out to be a more direct version of the same signal, and AI systems appear to be reading it that way even when there’s no link attached at all.

The practical shift isn’t abandoning link building — the data doesn’t support that — it’s recognizing that a link is no longer the only unit of earned attention worth tracking, and for AI visibility specifically, it may not even be the most predictive one.