
Key takeaways
- Ahrefs found branded web mentions correlated with AI mentions at 0.664 in ChatGPT, 0.709 in AI Mode and 0.656 in AI Overviews.
- YouTube mentions had the strongest correlation of any metric, at roughly 0.737 in ChatGPT, 0.740 in AI Mode and 0.712 in AI Overviews.
- Backlink counts correlated at only about 0.20 to 0.28, and Domain Rating at 0.266 to 0.326, so link metrics were weak predictors here.
- Ahrefs states that correlation is not causation and that improving these metrics will not automatically boost AI visibility.
- Every brand in the sample had a Domain Rating above 40, so the results say little about small or new sites.
Advice on getting a brand named by AI assistants usually rests on anecdote. Ahrefs published one of the larger attempts to measure it: a correlation study of about 75,000 brands, released on 12 December 2025 and written by Louise Linehan. This piece reports what that study did and found, and what it cannot show. It sits within AI search coverage.
What did Ahrefs measure?
Ahrefs wanted to know which metrics move together with brand mentions across three surfaces: ChatGPT, Google AI Mode and Google AI Overviews. It selected domains with a Domain Rating above 40, took each one’s highest-volume keyword with a monthly volume of at least 800, and used its Brand Radar tool to count mentions across millions of AI responses. Correlations are Spearman coefficients. Ahrefs says ChatGPT, Copilot, Gemini and Perplexity share one question pool, while AI Overviews and AI Mode share another.
Which metrics correlated most with AI mentions?
Off-site brand signals led. YouTube mentions were highest at roughly 0.74, branded web mentions followed at 0.656 to 0.709, and branded anchors at 0.511 to 0.628. Branded search volume ranged from 0.352 to 0.466.
Link and size metrics trailed. Domain Rating ran from 0.266 to 0.326 and backlink counts roughly 0.20 to 0.28. Ahrefs describes almost no relationship between content volume and AI visibility. Mention count edged out view-weighted impressions on YouTube, which Ahrefs reads as breadth mattering slightly more than reach.
Does this prove mentions cause AI visibility?
No, and Ahrefs says so: “correlation isn’t causation.” A well-known brand is likely to be discussed on the web, covered on YouTube and cited by AI at once, so one hidden factor could drive every column. The study also counts mentions of brands, not citations of pages, so it does not show that a given URL will be cited. For that question see this site’s write-up of the content length correlation.
What should a publisher take from it?
Treat it as a hypothesis generator. The consistent gap between mention-type signals and link metrics is worth testing on your own brand, for example by tracking whether third-party coverage precedes changes in AI mentions. Do not treat 0.7 as a recipe. The sample excludes low-authority domains, and Ahrefs’ own tool produced the data.
The evidence
- Sample
- About 75,000 brands; per-platform sample sizes not disclosed
Hypothesis: Ahrefs set out to identify which search and web metrics correlate most strongly with how often a brand is mentioned across ChatGPT, Google AI Mode and Google AI Overviews.
Method: Ahrefs filtered its web index for domains with a Domain Rating above 40, took each domain's highest-volume keyword with at least 800 monthly searches, and counted brand mentions in millions of AI responses using Brand Radar. It then computed Spearman correlations between mention counts and metrics including YouTube mentions, branded web mentions, branded anchors, branded search volume, backlinks, Domain Rating and site pages.
Findings
- Branded web mentions correlated with AI mentions at 0.664 in ChatGPT, 0.709 in AI Mode and 0.656 in AI Overviews; branded anchors reached 0.511, 0.628 and 0.527 on the same three surfaces.
- YouTube mentions were the strongest metric at roughly 0.737, 0.740 and 0.712, with YouTube mention impressions close behind at roughly 0.717, 0.724 and 0.712.
- Link metrics were weak: Domain Rating correlated at 0.266, 0.285 and 0.326, backlink counts at roughly 0.200, 0.250 and 0.280, and Ahrefs reports almost no relationship between content volume (site pages) and AI visibility.
- Branded search volume sat in the middle at 0.352 in ChatGPT, 0.466 in AI Mode and 0.392 in AI Overviews.
Limitations: Ahrefs states that correlation is not causation and that improving these metrics will not automatically boost AI visibility. The sample is limited to domains with Domain Rating above 40 and a keyword with at least 800 monthly searches, so smaller brands are not represented. Ahrefs did not disclose exact prompts or per-platform sample sizes, and the data comes from its own Brand Radar tool, so no independent party has reproduced it. Brands that are already well known plausibly attract mentions, YouTube coverage and links alike, which the design cannot separate.
Sources
- 1.Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied) - Ahrefs, December 12, 2025Primary
About the author

Founder
Founder, UpgradIQ, Inc.
Adam Hafez works on technical SEO and search measurement: how pages get crawled, indexed, ranked and now quoted by answer engines. He founded UpgradIQ, which reads Google Search Console and GA4 to tie ranking movement back to the changes that caused it. He publishes what the data supports and states the limits of it.
- Technical SEO
- Search Console and GA4 measurement
- Answer engine optimization
- Structured data
Related reading
Do AI assistants cite fresher content than Google? Ahrefs data
Ahrefs compared the age of 16.975 million URLs cited by AI assistants against organic Google results. Here is the method, the numbers and the limits.
A randomized field experiment on AI Overviews and organic clicks
A randomized Chrome-extension experiment hid AI Overviews for 1,065 searchers. This piece analyzes Agarwal and Sen's SSRN study of the CTR impact.
Who blocks AI crawlers: 60 robots.txt files, read in full
We read the robots.txt of 60 publishers and SEO vendors. News sites block AI crawlers almost universally. The SEO industry does not.
AI citation hallucination rates: what three real studies found
GPTZero, the Tow Center and a Cureus study each measured AI citation hallucination differently, and none of their numbers agree with each other.



