Ahrefs study tracks rising citation share in Google AI Overviews

Citation share is the KPI hiding in plain sight

Ahrefs’ June 2026 study makes the shift obvious: in Google AI Overviews, the question is no longer just who ranks first, but who gets cited. Ahrefs defines that as mention share, the percentage of total citations captured by a domain among the top sources, and that is the kind of number growth-stage agencies can actually build around.

The scale matters. Ahrefs says the June 2026 dataset covers more than 3 million U.S. queries across all topics, and Google says AI Overviews now reach more than 2.5 billion monthly active users. That puts citation share in a place every agency should care about, because the answer layer is no longer a niche experiment. It is a massive visibility surface in Google Search, available in more than 200 countries and territories and more than 40 languages.

The top of the chart tells the story fast. YouTube leads with 20.9% mention share, Reddit follows at 19.6%, and Facebook sits at 11.6%. Google is next at 6.0%, then Instagram at 5.2%, Wikipedia at 4.8%, with Amazon and Quora tied at 4.0%. The top three sources alone account for more than half of all citations among the top domains, which should immediately change how you think about “winning” in AI search.

Why the benchmark matters more than a rank report

This is not a simple organic rankings story. Ahrefs’ March 2026 update found that only 37.9% of URLs cited in AI Overviews also appeared within the first 10 result blocks for the same query. Another 31.2% came from positions 11 through 100, and 31.0% were beyond the top 100. That is the biggest clue in the whole study: AI Overviews are not just repackaging the obvious page-one results.

Ahrefs says improved parsing and Google query fan-out may help explain the gap from its earlier 76% top-10 overlap study, and that lines up with what anyone watching Search has felt for months. AI systems are pulling from a wider pool than classic blue-link thinking assumes. If you are only measuring whether a client owns the first organic slot, you are missing a large slice of the citations that shape what users actually see.

Google’s own product direction reinforces that point. The company says AI Overviews are meant to help people ask new kinds of questions, and it is now testing a Search Console control that lets site owners decide whether their site can appear in and help ground generative AI Search features. That is a strong signal that citation presence is becoming measurable, contestable, and strategically important.

How to turn the monthly list into an agency workflow

For agencies, the Ahrefs list is more than a curiosity. It is a monthly operating tool.

Use it for prospecting:

• Find categories where a client has decent organic visibility but almost no citation share.
• Spot brands or publishers that dominate AI Overviews without owning the old-school rankings.
• Build outreach lists around domains that are already trusted by the answer layer.

Use it for competitive benchmarking:

• Track citation share month over month, not just keyword positions.
• Compare a client’s cited presence against the domains that keep surfacing in the same topic cluster.
• Separate Google-style authority from answer-layer authority, because the two do not always line up.

Use it for reporting:

• Put mention share in the dashboard next to organic traffic, branded search, and conversions.
• Show which content types are earning citations even when they are not rank number one.
• Tie content, PR, and authority investments to a metric clients can understand without having to decode rank volatility.

Ahrefs’ own setup is part of the lesson. The company says the list updates automatically every month using Agent A, its SEO and marketing agent. That makes the study useful on two levels: it tracks AI citation behavior, and it shows a workflow agencies can imitate if they want reporting that keeps up with the pace of AI Search.

What the citation mix says about content strategy

The domain mix is the loudest strategic signal in the whole dataset. YouTube, Reddit, Facebook, Instagram, Wikipedia, Amazon, and Quora are not the usual lineup of traditional publishers. That tells you AI Overviews reward more than editorial depth alone. They also reward the formats people use to explain, demonstrate, compare, and validate things in the open.

That matters for content planning. The domains leading citation share suggest that some of the most citeable assets are not polished landing pages at all, but video explainers, community discussions, reference-style pages, product listings, and Q&A threads. In practice, that means a client can earn AI citations even when its main page is not the organic leader, as long as the underlying content is the clearest source for the specific question being asked.

For growth teams, the test is simple: do your assets answer the query in a form AI systems can lift confidently? A short product explainer, a tightly structured comparison, a how-to video, or a well-supported FAQ can be more useful to the answer layer than a broader page that ranks well but never gets quoted. The new game is not just “own the keyword.” It is “be the source Google wants to cite.”

The real takeaway for agencies

Google says AI Overviews are already at internet-scale, with more than 2.5 billion monthly active users and availability across more than 200 countries and territories and more than 40 languages. Ahrefs’ June 2026 study shows the citation winners are often platforms built for explanation, discussion, and proof, not just classic publishing. Put those two facts together and the agency KPI becomes obvious.

If you are still treating AI Overview visibility as a side effect of organic rankings, you are already behind. Citation share is becoming its own benchmark, and the agencies that learn to report it cleanly will have a much sharper story to tell about visibility, authority, and growth.