Earn a spot within the high 10, shut the tab glad, and head out to blissful hour.
That’s the way it was once. When you appeared excessive on the search outcomes web page, you felt assured about your web page’s efficiency.
Up to now, pages in Google’s high 10 supplied most citations in AI overviews. Nonetheless, this share has modified considerably in lower than a 12 months attributable to a brand new problem: the question fan-out. Whereas it’s essential to acknowledge that AI Overviews can typically embrace errors, specializing in being cited quite than simply ranked will assist your technique thrive by means of these modifications.
Key takeaways
- Rating and quotation have break up aside. The share of AI Overview citations that additionally rank in Google’s high 10 fell from about 76% in July 2025 to roughly 38% by March 2026.
- Question fan-out is the trigger. Google’s AI breaks one query into many sub-queries and builds its reply from the pages that floor most persistently throughout all of them—not simply the web page that ranks for the typed question.
- High-10 rating nonetheless issues. A robust natural place stays essentially the most dependable feeder into AI Overviews and the clearest authority sign Google has—it will get you into the candidate pool.
- Quotation takes depth and credibility. Reply engine optimization (AEO) rewards self-contained sections, topic-level protection, and E-E-A-T alerts that permit a mannequin elevate a clear, quotable declare.
- Monitor citations, not simply rankings. Place monitoring alone now misses many of the image of the place your site visitors comes from.
What’s a question fan-out?
Question fan-out is how an AI search system breaks one consumer question into a number of sub-queries. It collects data for every, then combines the outcomes into one response. LLMs depend on it to provide richer solutions.
While you ask a query in Google’s AI experiences, the system expands it earlier than it solutions. Behind the scenes, an AI mannequin breaks your query right into a set of associated sub-queries—equal phrasings, follow-ups, broader framings, narrower specs—and runs them unexpectedly.
The AI Overview is then constructed from the pages that floor most reliably throughout that entire set. A web page can rank first for its headline question, but it surely won’t present up within the fan-out. This occurs as a result of the mannequin appears to be like at a number of associated searches the place different pages present extra detailed data.
Here’s what that appears like in apply:
“How do I measure the ROI of our B2B content material advertising and marketing program to show its worth to executives?”
As a substitute of operating that one question and stopping, the LLM pulls the query aside into shorter searches drawn from inside it:
- measure content material advertising and marketing ROI
- B2B content material advertising and marketing metrics
- content material advertising and marketing worth
- show content material ROI to executives
- content material program efficiency
That shift—discovering solutions primarily based on essentially the most constant pages, not simply the typed query—is what separates rating from quotation.
Why rating nonetheless issues
Roughly half of Google searches already floor an AI abstract, and McKinsey initiatives that determine will cross 75% by 2028. In a McKinsey survey of 1,927 US customers, half now actively hunt down AI-powered search, and it has turn into the main digital supply they use for purchasing choices. With most searches headed towards an AI reply, the pages that get cited resolve most of your site visitors.
In July 2025, about 76% of pages cited in Google’s AI Overviews additionally ranked within the high 10 for a similar question. Ahrefs’ March 2026 examine checked out 863,000 key phrases and round 4 million AI Overview URLs. It discovered that determine had dropped to about 38%.
The remainder of the citations moved elsewhere on the net. Ahrefs discovered them break up virtually evenly: roughly 31% from pages rating 11 to 100, and one other 31% from pages rating previous 100 or not rating for the question in any respect. Rating and getting cited now not go collectively.
Don’t throw out your plans to rank simply but. A 38% overlap continues to be a big minority, and top-10 pages stay essentially the most dependable feeder into AI Overviews. A robust natural place continues to be the clearest authority sign Google has. Rating properly will get you thought-about. Getting cited takes extra.
Consider it as two gates. Conventional web optimization will get you into the candidate pool, and fan-out decides which candidates get quoted. A web page that ranks and covers its subject with actual depth clears each. A web page that ranks for one key phrase and stops there clears the primary and stalls on the second.
What AEO really asks of your content material
That is the place reply engine optimization enters. Construction helps: clear headings, self-contained sections, schema, and a direct reply close to the highest all make content material simpler for a mannequin to parse and extract.
What you shouldn’t ignore is protection and credibility. If the AI samples sub-queries, your content material has to reply the encircling questions along with the principle question. Which means depth over key phrase breadth: one useful resource that resolves the true query and its pure follow-ups, written with sufficient specificity {that a} mannequin can elevate a clear, citable declare from it. The identical E-E-A-T (expertise, experience, authoritativeness, and trustworthiness) alerts Google has all the time rewarded are what make a passage price quoting.
AEO is identical demand for good content material with the stakes raised, the place each part now has to face by itself.
The place to spend your effort now
Fan-out rewards content material that anticipates the questions a reader really has, and anticipating these questions is editorial judgment.
Figuring out which sub-questions matter, which framings are trustworthy, the place to be particular and the place to be temporary, what declare is price stating cleanly sufficient to be quoted—that’s the work of an skilled editor or subject-matter professional.
The manufacturers that get cited persistently share a trait: their content material carries a transparent viewpoint and the depth to assist it throughout a subject. Quantity of output has little to do with it.
Methods embrace:
- Construct round matters, not single key phrases. Map the cluster of questions a reader asks earlier than and after the headline one, and canopy them in depth inside a web page or a good set of linked pages.
- Make each part independently citable, since a mannequin will elevate passages out of context.
- Keep the course on incomes the rating, as a result of top-10 placement continues to be your surest path into the candidate pool.
- Watch the place you might be cited, not solely the place you rank. Place monitoring alone now misses many of the image.
- Spend money on editorial depth and experience. The factor that makes content material price citing holds regular: depth, judgment, and a viewpoint a reader, or a mannequin, can belief. That’s editorial work, and it’s precisely what Contently’s community of professional editors and subject-matter writers is constructed to do.
Regularly requested questions
What’s a question fan-out in AI search?
Question fan-out is the approach an AI search system makes use of to interrupt a single consumer question into a number of associated sub-queries—equal phrasings, follow-ups, broader framings, and narrower specs. It runs them unexpectedly, then builds its reply from the pages that floor most persistently throughout the entire set quite than from the one web page that ranks for the typed query.
What’s the distinction between web optimization and AEO?
web optimization (SEO) works to earn a excessive rating on the outcomes web page, which will get your web page into the pool of candidates an AI can draw from. AEO (reply engine optimization) works to get your content material quoted within the AI reply itself, which will depend on self-contained sections, topic-level depth, and E-E-A-T alerts a mannequin can elevate a clear declare from. web optimization will get you thought-about; AEO will get you cited.
Does rating in Google’s high 10 nonetheless matter for AI search?
Sure. Despite the fact that the overlap between top-10 rankings and AI Overview citations fell to about 38% by March 2026, top-10 pages stay the one most dependable feeder into AI Overviews and a robust natural place continues to be the clearest authority sign Google has. Rating will get you into the candidate pool; quotation takes extra depth and credibility.
How do I get my content material cited in Google’s AI Overviews?
Cowl a subject—not a single key phrase—deeply sufficient to reply the encircling sub-questions a reader and the fan-out will ask. Construction every part to face by itself with clear headings, schema, and a direct reply close to the highest, and write with sufficient specificity and demonstrated experience (E-E-A-T) {that a} mannequin can extract a clear, quotable declare.
What’s E-E-A-T and why does it matter for AEO?
E-E-A-T stands for expertise, experience, authoritativeness, and trustworthiness—the alerts Google has lengthy rewarded in rankings. For AEO they matter as a result of the identical qualities that make a passage credible to Google are what make it price quoting to an AI mannequin: particular, well-sourced, professional content material is the type a mannequin is most keen to quote.
