
What the Ahrefs study found about AI Overview citations
Only 38% of Google AI Overview citations now come from pages that rank in the top 10 for the same query, according to an Ahrefs analysis. That is down from 76% in an earlier Ahrefs study. In plain terms, a top ranking is now a weaker signal that Google’s AI answer will cite you. The link between rank and citation is loosening.
The study was large. Ahrefs looked at 863,000 keywords and about 4 million AI Overview URLs. The citations that did not come from the top 10 split almost evenly. About 31.2% came from pages ranking 11 to 100, and about 31.0% came from pages ranking beyond position 100. So roughly six in ten cited pages sat outside the first page of results.
Ahrefs offers two reasons. First, it points to query fan-out, where Google breaks one search into several smaller related searches. Pages that show up often across those sub-searches can get cited, even if they do not rank for the original term. Second, video is rising. Among cited pages that did not rank in the top 100 for the keyword, 18.2% were YouTube URLs. YouTube made up 5.6% of all AI Overview citations.
Who reported the AI Overview citation data, and one caveat
The data comes from Ahrefs and was reported by Search Engine Journal. Ahrefs published its own write-up as well. One honest caveat belongs here: Ahrefs says it improved how it detects citations since its earlier study, so the 76% and 38% figures are not a perfect like-for-like comparison. The direction of the change is clear, but the exact size of the drop should be read with that method change in mind.
How the ranking and AI citation gap fits B2B marketing in 2026
This finding sharpens a shift B2B teams already feel. AI search has passed SEO as the top content distribution channel for many B2B marketers, as covered in our report on AI search overtaking SEO. The old playbook was to win a top-10 rank and collect the click. The new question is whether the model cites you at all, from wherever your content sits. Control of that surface is contested too, seen in the UK ruling that lets publishers opt out of Google AI search. For B2B brands, the takeaway is that rank alone no longer protects visibility. Content has to be clear, specific, and useful across many related questions, not just one head term.
Talking Shift: why rank is no longer the moat for B2B brands
Start Some Shift reads this through the Binary Buyer Thesis: every B2B purchase passes through a Machine layer before it reaches a human, and that layer now decides what the buyer sees. For years the moat was a top ranking. Here is the line that should change your content plan: the model does not cite your rank, it cites your usefulness, and those two things have started to part ways. When Google fans one query into many, the brand that answers the whole cluster of related questions gets named, even from page two or deeper. The brands that win will treat content as a body of clear, specific answers a model can pull from anywhere, not as a single page fighting for one blue link. A number-one rank that no model quotes is a trophy the buyer never sees.
What B2B marketers should do about AI Overview citations
The data points to a clear content plan:
- Map the cluster of related questions around each buying topic, since Google fans one query into many.
- Write specific, self-contained answers a model can lift cleanly, not vague overview pages.
- Stop treating a top-10 rank as the finish line, and track AI citations as a separate goal.
- Build content in more than one format, including video, given YouTube’s growing share of citations.
- Structure pages with clear headings and direct answers so engines can parse them.
- Keep facts sourced and current, since credibility is what earns a citation.
- Watch your own appearances in AI answers, not just your ranking positions.
What to watch next for AI Overview citations
Watch whether Google leans further on query fan-out, which would keep pulling citations away from the top 10. Watch whether reliable tools for tracking AI citations mature, since brands need to see where they appear. And watch video keep rising as a cited source, which would push more B2B teams toward YouTube.