With the new Google update (June 2026), every search you rank for now drags ten invisible ones behind it, known as a query fan-out. Here’s what that actually means and how to fix your SEO plan to reflect this.
Google isn’t just weighing your page against everyone else’s page for that phrase.
Behind the scenes, it’s already run the equivalent of ten other searches, such as “average agency cost,” “should I use a small or global agency,” “best agency in Worcester”, and it’s building its answer from whatever content covers that whole cluster best.
This is AI Overviews, and it has quietly rewritten what “ranking” even means.
We’ve spent the past few weeks pulling apart how this actually behaves in practice: not the theory, the real patterns showing up in the data.
Here’s what we found, and why most of it is good news if you’re willing to adapt.
Fan Queries Instead of Individual Searches
For every search someone types, the system generates a cluster of related questions and pulls in content that satisfies the cluster as a whole and not just the literal words typed in. Google calls it “query fan-out.”
Ahrefs (2026) found that a single search generated an average of 10.7 related fan-out queries.
When you look these fan-out queries up in a keyword tool, 95% of them show zero search volume. That’s because almost nobody types them- they’re synthetic, generated on the fly, and they change every time you run the same search twice. They’re also personalised, so two people searching the identical phrase can trigger different fan-outs entirely.
You’re no longer optimising for a keyword. You’re optimising for a cluster of questions a real person would never think to ask.
Measuring Your AI Presence? Ranking is out. Visibility is in.
Because citation is filtered for randomness by design, being ranked #1 no longer guarantees you’ll be the one AI cites, BUT ranking lower doesn’t rule you out either. The more useful question has shifted from “where do we rank” to “how likely are we to show up at all?”
A few things move that needle:
- Consensus. Multiple independent sources saying the same thing, in more places, builds the trust AI is looking for.
- Freshness. 75% of AI-cited content had been refreshed within the previous 30 days. It’s key to make sure your content isn’t outdated and replaceable.
- Existing authority. 76% of AI Overview citations were already ranking on page one, so this isn’t a replacement for good SEO; it’s an amplifier of it.
- New territory. YouTube and Reddit are increasingly pulled into AI answers alongside traditional web pages.
Metrics worth tracking now
Rank tracking alone doesn’t capture what’s happening anymore. Alongside it, we’re now tracking three additional signals for every client:
- Mentions: How often a brand is referenced inside an AI-generated answer.
- Citations: How often that mention comes with an actual link back to the site.
- Impressions: How often the content appears in AI results at all, cited or not.
How to Pick the Right Keyword In 2026
Long-tail, exact-match phrases are largely a dead end here. What works better is a short seed keyword layered with distinct modifiers:
Seed keyword (1–2 words): “coffee” “sandwiches
Add an amplifier (5+): → best · how to · near me · vs. · for beginners
Before committing to any of them, we run three checks:
- Is there real buying intent behind it?
- Is the current page-one format even one we can compete with?
- Would an AI answer alone fully satisfy the search, leaving no reason left to click through?
You need to bear in mind that even if you write the best content in the world, if the AI can directly pull this into a snippet, people have no reason to click on your site. Below’s what AI can’t replace.
The content AI genuinely can’t take from you
Anything interactive is naturally resistant to being summarised away, because AI can describe it, but it can’t do it for the user in the answer box itself. Think calculators, checkers, generators, templates, finders, planners, makers.
If you can build a tool which people need to use instead of reading via an AI snippet, you’re guaranteeing that AI can’t steal your traffic.
Think interactive quizzes, a pricing calculator, or a template on creating content.
Writing so AI can Actually Use Your Content
The single biggest structural shift in the Google update: write assuming any one paragraph might get lifted with zero surrounding context. That means leading with the answer first (BLUF: Bottom Line Up Front), keeping every section genuinely self-contained, and labelling your work clearly so credit sticks where it’s due. Think BUZZGEN’S Google Update Guide, and not just a guide.
For example, if you read this paragraph as standalone content, you can see in the first sentence that this content covers structures needed when writing for the Google update in 2026.
It also means depth beats breadth. Going deep on one topic at a time, building real density before moving on, consistently outperforms thin coverage spread across many.
The technical check almost everyone skips
None of the above matters if your robots.txt file is quietly blocking the crawlers doing this work. It’s a two-minute check with an outsized cost if it’s wrong. If AI can’t access your content, it can’t cite it, no matter how good it is.
Get the full guide on the Google Update for June 2026. Everything above, plus the complete keyword framework, content structure playbook, and a ready-to-run 7-step action plan.
Leave a Reply