← Back to HomeBack to Blog List

Ahrefs Studied 75,000 Brands: Backlinks Barely Move AI Visibility—Mentions Win 3-to-1

📌 Key Takeaway:

Ahrefs studied 75K brands: mentions correlate 0.664 with AI visibility, backlinks just 0.218, YouTube mentions strongest at 0.737.

Ahrefs Studied 75,000 Brands: Backlinks Barely Move AI Visibility—Mentions Win 3-to-1

Ahrefs Studied 75,000 Brands: Backlinks Barely Move AI Visibility—Mentions Win 3-to-1

November 1, 2026 · GEO Industry Analysis

If you spent the last decade building backlinks for search traffic, the following data might sting.

Ahrefs ran a massive correlation analysis across 75,000 brands, and the conclusion fits in one sentence: in the AI search era, the backlinks you painstakingly built may be worth less than a casual mention of your brand.

0.737
YouTube mentions correlation with AI visibility (strongest signal)
0.664
Brand web mentions correlation
0.218
Backlinks correlation

1. The Data: SEO's Core Asset Is Depreciating

Quick primer on correlation coefficients: 1.0 means perfect correlation, 0 means none. Above 0.6 is strong; around 0.2 is "technically related, but barely."

Here's how Ahrefs' measured signals stack up against AI visibility:

SignalCorrelation with AI VisibilityVerdict
YouTube mentions~0.737Strongest single signal
YouTube mention impressions0.717Video reach matters too
Brand web mentions0.656–0.709Strong across engines
Branded anchors0.511–0.628Moderate
Backlinks0.18–0.30Weak, near-irrelevant
Number of site pages0.17–0.194"Almost no relationship"

That last row deserves attention: page count correlates at just 0.17 with AI visibility. Ahrefs' own words: "almost no relationship."

Publishing 500 more thin pages won't reliably increase your AI visibility. The content-flooding playbook is dead in AI search.

2. Why Mentions Beat Links

To understand this shift, you need to understand how differently AI and traditional search engines "see" the web.

Traditional search asks: "who links to you?" Links are votes. More backlinks = more authority. That's PageRank's core logic.

AI retrieval asks: "who's talking about you, and do those mentions reconcile across enough independent surfaces to trust?" AI doesn't count votes—it reads semantically. It processes how the entire web discusses you, then judges: Is this brand real? Widely recognized? Worth recommending?

This explains why backlinks fail:

  • Links can be bought; mentions are harder to fake: A backlink can be purchased. But genuine discussion across multiple independent sources is extremely difficult to manufacture
  • AI reads semantics, not code: An unlinked brand mention (e.g., "I use XX, it's pretty good" in a forum) does nothing for Google rankings—but AI understands it, and it feeds into your brand profile
  • Cross-validation: AI compares how different sources describe the same brand. Consistent, authentic mentions across platforms are the strongest trust signal

3. Why YouTube Became the Strongest Signal

The study's biggest surprise was YouTube. A 0.737 correlation—higher than any traditional signal. In Google AI Mode's B2B queries, YouTube appears in 38% of citations.

The reasons make sense:

  • YouTube content is AI's hardest source to fake: A genuine product review video includes visuals, audio, real people—AI inherently trusts multimodal sources more
  • Videos carry "experience evidence": Text can lie, but footage of a product actually working is much harder to fabricate
  • Incredibly accessible for smaller brands: GoGoChimp's case shows a low-authority domain earned 3,600 Bing Copilot citations in 90 days via YouTube—versus just 82 Google organic clicks in the same window. 44:1

For brands with limited budgets, YouTube may be the highest-ROI entry point into AI visibility.

4. The Other Sobering Number: 11%

The study also found: only 11% of domains are cited by both ChatGPT and Perplexity.

Each AI platform's "information diet" is remarkably different. Dominating ChatGPT doesn't mean Perplexity knows you exist. The only universal solution is entity coverage—consistent brand information across enough independent platforms:

  • Wikipedia/Wikidata (if eligible)
  • LinkedIn, Crunchbase, and other business directories
  • Trustpilot and review platforms
  • Industry association listings
  • YouTube channel
  • At least one vertical industry directory

The research suggests a minimum of 8 independent surfaces for AI to form a stable brand entity.

5. What to Do Now

  1. Reallocate your budget: If you're still buying backlinks, stop. Redirect that budget toward generating genuine brand discussion—user reviews, community engagement, media contributions
  2. Start a YouTube channel: It doesn't need to be polished—authentic product demos and tutorials work. To AI, one genuine demo video beats ten advertorials
  3. Run a "mention audit": Ask major AI platforms "What do you think of [your brand]?" and examine what sources they cite. If your mentions cluster on 2-3 platforms, your entity coverage is too thin
  4. Unify brand information everywhere: AI cross-validates. Inconsistent company descriptions across your website, directories, and social platforms directly lower your trust score

In the SEO era, links were currency. In the AI era, mentions are currency, and authentic third-party discussion is gold. The monetary system changed—has your portfolio?

Sources: Ahrefs "Across 75,000 Brands" research series (December 2025 data re-analysis), GoGoChimp "Entity SEO and Brand Mentions: Why AI Cites Brands 2026" (2026-10-26). All correlation data cited from the above public studies.

Want Better SEO Results?

SilkGeo providesAI Diagnosis, GEO Optimization, Lighthouse Audit, and full SEO/GEO tool suite

Use SilkGeo for free