Topic Clusters: Cover Every Question Buyers Ask AI

If ScoutRival’s Site Audit flagged “Topic fan-out coverage,” it means buyers are asking AI a whole set of sub-questions about your service — and your site only answers one or two of them. This guide shows you how to map the full question set and build a topic cluster that covers it, on every major platform.

What “query fan-out” is, and why it changes the game

When you ask Google AI Mode or Gemini a question, it rarely runs just your query. It decomposes your prompt into many related sub-queries — typically 8 to 12 — runs them in parallel across the index, and synthesises one answer from the strongest passages across multiple pages. Google calls this query fan-out. ChatGPT and Perplexity do a similar multi-step retrieval.

Here’s why that matters for you. Someone asking “should I get laser hair removal?” triggers hidden sub-queries like how much does laser hair removal cost, is laser hair removal safe, how many sessions does it take, laser hair removal near me, laser vs waxing, and laser hair removal reviews. If your site answers only “what is laser hair removal,” you’re competing for one of those twelve sub-queries. If you have a focused, genuinely useful page for each of them, you can be pulled into the answer from several angles at once.

This is a STRONG lever. It reflects how these systems actually retrieve, and 2026 citation data backs it up: the majority of AI citations for commercial queries are “best/top” lineups and comparison-style pages, and buyers ask the sub-questions individually before they buy. A cluster of focused pages simply gives the fan-out more surfaces to land on.

The sub-question set (memorise this)

For almost any local service, the same seven-part question set fans out. Map your coverage against it:

1. Cost — “how much does [service] cost in [city]?”
2. Vs / alternatives — “[service] vs [alternative]” (e.g. laser vs waxing, repair vs replace).
3. How long / process — “how long does [service] take?” / “what’s the process?”
4. Is it safe / worth it — “is [service] safe?” / “does [service] work?”
5. Near me / [city] — “[service] near me” / “[service] in [city].”
6. Reviews / results — “[service] reviews” / “before and after.”
7. How to choose — “how to choose a [service provider] in [city].”

Most service businesses have a single page for #1–#5 mashed together and nothing for #2, #6 or #7. Those gaps are exactly where competitors (or aggregator sites) win the answer.

Why a cluster beats one fat page

It’s tempting to answer everything on one long service page. Two reasons that loses:

1. Focus wins the match. A page whose title is the sub-question (“How much does a roof replacement cost in Austin?”) matches that sub-query far more cleanly than a general “Roofing Services” page that mentions cost in paragraph seven.
2. Coverage compounds. Seven focused pages can each earn long-tail impressions and each get cited independently. One giant page can only be one result.

The right structure is a hub and spokes: a main service page (the hub) surrounded by focused sub-pages (the spokes), all interlinked. This is a genuine topic cluster — and it’s the same structure this very blog uses.

How to build your topic cluster

1. Pick the primary service and list its sub-questions

Choose one core service. Write out the seven sub-questions above, filled in with your service and city. Add any extras your buyers actually ask (from chat logs, Google Business Profile Q&A, and “People Also Ask”).

2. Audit your current coverage

For each sub-question, ask: do I have a page whose main job is answering it? Most businesses find they cover two or three and miss four or five. ScoutRival scores this directly — covered ÷ expected — and lists the exact gaps, so you don’t have to guess.

3. Publish one focused page per gap

For each missing sub-question, create one focused page whose title is the sub-question. Not a section on an existing page — its own page. Each should be answer-first (the answer up top), use real facts, and be well-structured with lists and tables.

4. Interlink the cluster

Every spoke links back to the hub (the main service page) and to its siblings (the other sub-pages). The hub links out to all spokes. This tells engines the pages belong together and helps crawlers discover them. Add the new pages to your nav or a “Popular questions” section so they’re reachable.

5. Let BlogCraft draft the gap pages — then review them

This is the one-click part of the loop. ScoutRival’s BlogCraft can draft a page seeded with the sub-question title — a “how much does [service] cost in [city]?” price guide, a “vs” comparison, a “how to choose” buyer’s guide. It gives you a real first draft in minutes.

The non-negotiable step: you review it for local accuracy before publishing. BlogCraft doesn’t know your real local prices, your specific service area, or the honest pros and cons — you do. Never auto-publish a draft. Fill in the true local details, correct anything generic, and then hit publish. The draft saves you the blank-page problem; your review keeps it truthful.

Platform-by-platform: building the cluster

The concept is identical everywhere — new focused pages, interlinked. Here’s the mechanics:

WordPress (Yoast or Rank Math)

1. Pages → Add New for each sub-question; title it as the question (“How Much Does a Roof Replacement Cost in Austin?”).
2. Write answer-first, add a pricing table or FAQ block where relevant.
3. Interlink: in the editor, select text and link to the hub and sibling pages.
4. Add the pages to a menu (Appearance → Menus) or a “Popular questions” block so they’re crawlable.
5. Yoast/Rank Math will let you set each page’s SEO title and meta — keep the sub-question in the title.

Wix

1. Add a new page per sub-question from the Pages panel; name it as the question.
2. Build the content, then add text links between hub and spokes.
3. Add the pages to the site menu and Publish.

Squarespace

1. Add a new page for each gap and title it as the sub-question.
2. Link between pages using inline text links; add them to navigation.
3. Save/publish.

Webflow

1. Create a static page (or a CMS Collection item, if you’re templating many) per sub-question.
2. Interlink hub and spokes, add them to nav, and Publish to your live domain.

Shopify

1. Use Online Store → Pages (or blog posts) for each sub-question; title them as questions.
2. Link between them, add to a menu, and save.

Any other site

Create one real, crawlable URL per sub-question, interlink the hub and spokes, and make sure each page is in your sitemap and nav.

How to confirm it’s fixed

1. Re-run your ScoutRival Site Audit — the “Topic fan-out coverage” score should climb as gaps fill, and covered ÷ expected rises.
2. Check that each new sub-page is a real, indexed URL (search site:yourdomain.com for it, or check GSC/Bing coverage).
3. Over a few weeks, watch for long-tail impressions on the new pages in Google Search Console and Bing Webmaster Tools — a sign they’re matching sub-queries.
4. Ask an AI engine a couple of the sub-questions plus your city, and see whether you start surfacing.

Common mistakes to avoid

• Cramming everything onto one page. A single “Services” page can’t win the cost query and the vs query and the how-to-choose query. Split them into focused pages.
• Spinning near-duplicate pages. Don’t clone one page and swap the city or a few words. Thin, templated doorway pages get discounted (and risk a spam classification). Each page must genuinely answer its own sub-question.
• Publishing BlogCraft drafts unreviewed. A draft that guesses your local prices or service area is a liability. Always review and correct for local accuracy before publishing — that owner review is a required step, never optional.
• Orphan pages. A great sub-page nobody links to is hard to discover. Interlink every spoke to the hub and siblings, and add it to nav.
• Ignoring the comparison and price-guide gaps. “Vs” and “cost” pages are among the most-cited formats for commercial queries — they’re usually the highest-value gaps to fill first.
• Covering questions no one asks. Base your set on real buyer questions (chat logs, GBP Q&A, People Also Ask), not invented ones.

The bottom line

AI answers are assembled from many pages, because AI Mode fans one prompt out into a dozen sub-questions. Map the full set — cost, vs, how-long, is-it-safe, near-me, reviews, how-to-choose — build one focused page for each gap, and interlink them into a hub-and-spoke cluster. It’s the difference between competing for one sub-query and showing up across the whole answer.

Want your coverage scored automatically? Run a free Site Audit with ScoutRival. It measures covered-versus-expected for your services, shows the exact gaps, and BlogCraft drafts each missing page for you to review and localise before you publish.