AEO Rankings
← Blog
Filed

The Page-One Myth: AI Overviews No Longer Repackage the Top 10

Ahrefs measured AI Overview citation overlap with the top 10 falling from 76% to 37% in eight months. BrightEdge puts it near 17%. Here is why both are right.

Bottom line

Ranking in Google's top 10 no longer predicts an AI Overview citation. Ahrefs measured overlap fall from 76% in July 2025 to 37% in March 2026. BrightEdge puts it near 17%. The methods differ, but the verdict does not: query fan-out now pulls citations from far beyond page one.

Last updated September 2026. The March 2026 Ahrefs update and the February 2026 BrightEdge report are the two most recent studies on this question.

For most of 2025, ranking in Google’s top 10 felt like the whole game. If you owned a page-one position, you assumed you owned the AI Overview slot above it too. That assumption is now measurably wrong, and two separate studies show exactly when it broke.

Most content plans still run on that old assumption. A team hits page one for its priority keyword, checks the box, and moves to the next term. The data below explains why that box no longer covers what teams think it covers.

Do AI Overviews only cite pages that rank in the top 10?

No. Ahrefs found that only 37.9% of AI Overview citations also rank in Google’s top 10, down from 76.1% eight months earlier. BrightEdge, using its own tracking, puts the overlap at roughly 17%, and says that figure has stayed flat for months. The exact number depends on who measured it, but both firms agree: most citations now come from pages outside page one.

This breaks the oldest assumption in AEO, that a strong SERP position is a reliable proxy for AI visibility. It was a fair bet in mid-2025. It is a bad bet now.

How much did the citation overlap actually fall?

Citation overlap fell by more than half in eight months, by Ahrefs’ own tracking. Its July 2025 study of 1.9 million citations found 76.1% pulling from the organic top 10. Its March 2026 update, built on 863,000 keyword SERPs and 4 million AI Overview URLs, found only 37.9%. BrightEdge’s separate tracking has held closer to 17% since early 2025, which makes it the steadier number if you want a stable baseline.

DateStudyTop 10 overlapSample
Jul 21, 2025Ahrefs, “76% of AI Overview Citations Pull From the Top 10”76.1%1.9 million citations
Feb 12, 2026BrightEdge, “AI Overviews at the One-Year Mark”~17% (11% to 24% by industry)Monthly tracking since Feb 2025
Mar 2, 2026Ahrefs, “38% of AI Overview Citations Pull From the Top 10”37.9%863,000 keyword SERPs; 4 million AI Overview URLs

Read the table left to right and the direction is clear, even with the methodology noise. Ahrefs’ own number cut itself in half in eight months. BrightEdge’s number barely moved, because it was already measuring something closer to what Ahrefs eventually found.

Where do the citations outside the top 10 come from?

They spread almost evenly across the rest of the web, not just positions 11 through 100. Ahrefs’ March 2026 update found 31.2% of citations ranking between positions 11 and 100, and another 31.0% ranking beyond position 100 or off Google’s index for that query entirely.

YouTube is one of the clearest winners of that shift. It accounted for 18.2% of the citations that ranked outside the top 100, and Ahrefs’ Brand Radar data shows YouTube citations up 34% over six months. If your competitor set includes video content, that share of voice now belongs to a channel your rank tracker never watches.

This is the part most keyword-first content plans miss. A page that never breaks the top 100 for your target term can still win the citation, as long as it answers one of the sub-queries Google generated better than anything else Google found. Format now competes with rank.

Why do Ahrefs and BrightEdge disagree by so much?

They are not measuring the same thing the same way. Ahrefs’ 2025 study only caught the citations its parser detected at the time; its 2026 update improved that detection, which alone explains part of the drop. BrightEdge runs its own Generative Parser Research on a separate dataset it has tracked monthly since February 2025. Different tools, different samples, different definitions of what counts as “the same query.” The two studies converge on direction, not on a shared number.

Both firms still land in the same place: most AI Overview citations no longer come from the traditional top 10. That is the finding that survives the methodology debate. For how a third-party study earns a place in our own scoring, see the methodology page.

What is query fan-out, and why is it replacing SERP-pull?

Query fan-out is Google’s practice of splitting one user question into several related sub-queries before it writes an AI Overview. Instead of pulling the top-ranked page for your exact search term, Google checks which pages appear most often across all the sub-queries it generated. A page that never ranks top 10 for your main keyword can still get cited if it ranks well for three of the five sub-queries Google fanned out to.

That is a structural break from the old model. Call the old model SERP-pull: it assumed the AI Overview simply borrowed from whatever already sat on page one. Fan-out retrieval assumes something different, that the best answer gets assembled from whichever pages cover the topic best across many angles, not just the one phrase you typed. Ahrefs points to Google’s January 2026 upgrade of AI Overviews to Gemini 3 as a plausible driver of the accelerated shift, though Google has not confirmed specific changes to fan-out behavior. The AEO glossary has more on how fan-out and related retrieval mechanics work.

Is this only a Google AI Overviews problem?

No, and that is the uncomfortable part. A separate Ahrefs study from August 2025 found that, on average, only about 12% of the pages cited by ChatGPT, Gemini, and Microsoft Copilot also ranked in Google’s top 10 for the same prompt. Perplexity was the outlier: roughly one in three of its citations pointed to a page already ranking in the top 10.

If you have been treating a Google top 10 ranking as a stand-in for AI visibility across every engine, the gap was already wider than the AI Overview numbers alone suggest. Each engine builds its own answer from its own retrieval process, and none of them read the SERP the way a rank tracker does.

Does ranking in the top 10 still help you get cited?

Yes, but it is no longer sufficient on its own. A top 10 ranking still raises your odds of a citation, and BrightEdge’s data shows the floor value of ranking varies by industry: overlap runs as high as 24% in Healthcare and as low as 11% in Finance. What changed is that ranking alone stopped being enough. You now need to win the sub-queries Google fans your topic out to, not just your primary keyword.

Three things follow from that shift:

  • Stop treating one keyword rank as your AI visibility score. It is one input among several sub-queries you are not tracking.
  • Cover the topic from multiple angles, not just the exact phrase you rank for, since fan-out rewards breadth.
  • Check your own overlap number instead of assuming the industry average applies to your pages.

How do you track your own citation overlap?

Start with a dedicated AI Overview tracker instead of guessing from anecdotes. SE Visible tracks daily AI Overview citations alongside your organic rankings, so you can calculate your own overlap number rather than borrow an industry average. Ahrefs’ Brand Radar, the tool behind the citation study above, layers citation tracking on top of its existing rank data. BrightEdge offers a similar view at enterprise scale, broken out by industry.

You can also build a rough version of this number by hand. Run your target queries in an incognito browser, record which URLs the AI Overview cites, and check how many of those URLs also sit in the organic top 10 for that same query. Do this across 20 to 30 queries and you have a manual approximation of the same measurement Ahrefs and BrightEdge run at scale, specific to your own topic and your own competitors.

If you want citation tracking bundled with the fixes that close the gap, content, FAQ schema, and monitoring across ChatGPT, Perplexity, Google AI Overviews, Gemini, and Microsoft Copilot, Temso covers that loop from $89/mo. It will not replace a dedicated rank tracker, but it shortens the distance between spotting a missing citation and doing something about it.

The page-one myth is dead. A top 10 ranking is still worth having, but it will not carry your AI Overview visibility alone anymore. Track your own citation overlap before you plan another content sprint around a single keyword. The full, ranked list of AEO tracking and execution tools is at /rankings/aeo-tools.

FAQ

Do AI Overviews only cite pages that rank in the top 10?

No. Ahrefs measured the share of AI Overview citations that also rank in Google's top 10 falling from 76.1% in July 2025 to 37.9% in March 2026, across a sample of 4 million AI Overview URLs. BrightEdge, using a different dataset, measures a flatter overlap of about 17%. Both studies agree that most citations now come from outside the traditional top 10.

What is query fan-out?

Query fan-out is Google's process of splitting a single search into several related sub-queries before it writes an AI Overview. A page can get cited because it ranks well across multiple sub-queries, even if it never ranks in the top 10 for your exact search term. Ahrefs points to fan-out as the main driver behind the drop in top 10 citation overlap.

Why do Ahrefs and BrightEdge report different citation overlap numbers?

They track different datasets with different tools and different definitions of a matching query. Ahrefs compares citation URLs to organic rankings for the exact same search term across a large keyword sample. BrightEdge runs its own Generative Parser Research and tracks the overlap monthly by industry. Ahrefs also improved its citation detection methodology between 2025 and 2026, which accounts for part of the drop in its own number.

Does ranking well on Google still help you get cited in AI Overviews?

Yes, but it is no longer enough by itself. A top 10 ranking still raises your odds of a citation, and the floor varies by industry: BrightEdge found about 24% overlap in Healthcare versus about 11% in Finance. What changed is that ranking for your primary keyword no longer guarantees you win the sub-queries Google fans your topic out to.

What tool should I use to track my own AI Overview citation overlap?

SE Visible and Ahrefs' Brand Radar both track daily AI Overview citations alongside your organic rankings, which lets you calculate your own overlap number instead of relying on an industry average. If you want citation tracking bundled with content and schema fixes in one subscription, Temso covers that loop starting at $89/mo.

Where do AI Overview citations come from if not the top 10?

Ahrefs' March 2026 study found the rest split almost evenly: 31.2% of citations rank between positions 11 and 100, and 31.0% rank beyond position 100 or off Google entirely. YouTube is a large beneficiary of that shift. It made up 18.2% of citations that ranked outside the top 100, and Ahrefs' Brand Radar shows YouTube citations up 34% over six months.