Query Fan-Out: Why Ranking #1 Isn't Enough for AI Answers (and the Free Way to Fix It)

You finally hit #1 on Google. Champagne. Confetti. Then you ask ChatGPT the same question and it recommends... your competitor.

Rude. Also fixable.

The reason has a name: query fan-out. Kasra Dash broke it down on Oct 30, 2026, and once you see it, you can't unsee it.

Quick words (one line each):

• Query fan-out: when AI splits one question into several smaller searches and builds its answer from all of them.

• AI Mode / AI Overviews: Google's AI-written answers inside search.

• Cited: the AI links to or names your page as a source.

• Consensus: several independent sites saying the same true thing about you.

• Training data: the giant pile of web pages an AI learned from before it ever answers you.

What fan-out actually looks like

"Query fan-out" is Google's own name for it (Kasra Dash, Oct 30, 2026). AI Mode splits the question into subtopics and runs many searches at once.

Kasra's example: "best laptops for students under $1,000" turns into "best laptops," "best battery life," "student laptop reviews," and more.

You ranked for the big one. The AI built its answer from the small ones.

Why it matters (with numbers)

Kasra Dash showed two studies:

1. Surfer's look at AI Overviews: pages that ranked for the main search plus at least one fan-out made up about half of the cited pages. Pages that ranked only for the main search were about a fifth.

2. Profound ran 10,000 prompts and found most of ChatGPT's searches were different from one run to the next.

So you can't guess the fan-outs. You have to look.

Step 1: Capture the fan-outs (free, by hand)

1. Write down 3-5 real questions your buyers ask.

2. Type each one into ChatGPT, Perplexity and Google AI Mode.

3. Write down the searches and sources each tool shows you.

4. Repeat on a few different days. One run is not a ranking.

Kasra demos his own paid tool (RankOS) for this step. You don't need it. A spreadsheet and 20 minutes do the job.

Step 2: Sort them into 3 groups

Kasra's 3 buckets:

1. Win on your own site. You can answer it yourself.

2. Win on your site and other sites. You answer it, and others back you up.

3. Win only on other sites. "Is this company legit?" You can't answer that about yourself.

Step 3: Build what each group needs

Kasra's list of what to publish:

1. List articles for "best" searches.

2. Comparison pages for "versus" searches.

3. Real reviews and awards.

4. Case studies with the business name in them.

5. Articles about the founder.

6. Accreditations the company actually holds.

Group 1 becomes sections or pages on your site. Groups 2 and 3 need other sites.

Step 4: Get others to say it too

Kasra's example: "we've built over 1,000 websites" on your own site is just a statement. Other sites saying it is what makes it stick.

His 4-article rule: for every fan-out he wants to win, he publishes 4 articles on different sites, all saying the same true thing about the brand. He calls that consensus.

He points to Google's rater guidelines, which tell raters to be skeptical of what a site says about itself and to check independent sources.

Free version: pick 1-2 true facts (years in business, number of projects, location) and get them said the exact same way on your Google Business Profile, LinkedIn, one industry directory and one guest post or interview.

Step 5: Only publish what's true

Google's AI guide warns against fake mentions and mass-made content meant to game AI answers (as Kasra cites it). Real facts, real reviews, real case studies. That's the whole trick.

Step 6: Be in the training data and the live answers

Kasra notes Common Crawl (a free web archive running since 2007) made up 60% of GPT-3's filtered training data. Live answers come from each engine's own search. So be in both.

Before you follow any SEO tip (including this one)

Semrush (Oct 27, 2026) gave 3 questions to ask about any advice:

1. Is it based on fact or assumption?

2. Is their site similar to mine?

3. Which side of the argument is missing?

Their line: "The mistake is not learning from successful companies. The mistake is taking one example and deciding that it's truth." The Profound study above is the reason you test fan-outs on your own questions, more than once.

Do this today

1. Write down 3 real buyer questions.

2. Ask ChatGPT, Perplexity and Google AI Mode each one.

3. Write down the searches and sources they show.

4. Sort them into the 3 groups.

5. Pick 1 true fact and make it match on 4 sites.

Related reading

• Guest posts that tell AI who you are

• Make your business facts match everywhere AI checks

• Get your business on the best-of lists AI reads

Sources

• Kasra Dash - Ranking #1 on Google Isn't Enough Anymore (2026-10-30)

• Semrush - How to Make Sense of Conflicting SEO Advice (Without Losing Your Mind) (2026-10-27)

Don't worry if this seems like it's a pain in the ass for someone who doesn't build websites. That's because it is. The good news: just click the link below, pull the trigger, and we'll take care of the rest.