Topic clusters are groups of linked pages on one subject. A pillar page covers the broad topic, and each cluster page answers one narrower question inside it. Every cluster page links back to the pillar, and the pillar links out to each of them. To plan one, you choose a pillar topic your buyers search for and split it into subtopics with one search intent each. Then you draw the links before anyone starts writing.

Plan a topic cluster on paper before anyone writes a page, because fixing a bad structure after publication means redirects and rewrites. The order below works for a new site and for a blog with years of loose posts. Most of the effort goes into steps 2 and 3, since that's where you decide how many pages the cluster needs.
The tool you plan with decides how much of this work you do by hand. Plan prices below are monthly list prices in 2026.
| Option | Price (2026) | What it covers | Who it suits |
|---|---|---|---|
| Spreadsheet with Google Search Console | Free | Your own query data. Grouping and link maps are manual | One site, one or two clusters, a writer who knows the subject |
| Semrush | Pro at $139.95 a month | Keyword research, rank tracking, site audit | Teams that need competitor keyword data at volume |
| Ahrefs | Lite plan at $129 a month | Keyword and backlink research, site audit, credit limits on Lite | SEOs who research competitors' link profiles as well as keywords |
| Snoika | Not stated in public sources found | Site audits, brand visibility tracking in Google and AI assistants, researched content maps, publishing in 135+ languages, with human review from the same team | Teams that want the plan, the writing and AI visibility tracking in one place |
Spreadsheet with Google Search Console
Semrush
Ahrefs
Snoika
If you only need keyword data and you'll write in-house, Semrush or Ahrefs will cover the research. A spreadsheet is enough for a first cluster on a small site. A platform like Snoika makes more sense once you're running several clusters across languages and need to know whether AI assistants cite them.
A topic cluster is a site structure where one page owns a broad subject and several pages each own a narrow part of it, held together by internal links. People also call it the hub and spoke model: the pillar is the hub, and the cluster pages are the spokes. The idea came out of HubSpot. Anum Hussain and Cambria Davies, then at HubSpot, tested it on the company's own content in their 2015 Topics Over Keywords research. They found that more internal links between related pages pushed those pages higher in search results and earned them more impressions.
The model replaced the habit of writing one post per keyword with no plan for how the posts relate. In a cluster, each page has one job and a known place in the structure. Search engines read the links as a sign that the pages belong together, and readers can move from the overview to the detail they came for.
A good pillar topic is broad enough to hold a full cluster yet narrow enough that one page can give a complete overview of it. Leslie Ye, who rebuilt a large share of HubSpot's content into clusters, wrote in HubSpot's topic cluster report that a pillar page should answer every question the reader who searched X keyword had and be broad enough to be an umbrella for 20-30 posts. If you're trying to rank the page for a long-tail phrase, she added, it isn't a pillar.
There is no fixed number of cluster pages a pillar needs, and the published ranges differ. Ye's test points to 20 to 30. The agency Clariant Creative asks whether a topic can support eight to 22 niche posts. I'd count the distinct intents you found in step 3 of the plan and stop there. Padding a cluster with near-duplicate pages to hit a target makes the cluster weaker, because those pages compete with each other.
Before you commit, check that your pillar topic matches something you sell or support. A cluster on a subject your buyers never search for can bring traffic and still bring no customers.
Every cluster page targets one search intent, and the search results tell you what that intent is. Type the query into Google and look at what ranks. Guides and product pages point to different reader needs; comparison tables point to a third. Two queries that return mostly the same URLs belong on one page. Two queries with different result sets need two pages, however close the wording looks.
The pillar usually takes the broad informational query, such as "email marketing". Cluster pages take the narrower ones, such as "email subject line length". Commercial queries like "best email marketing software" deserve their own cluster page, because that reader is comparing options and wants a verdict.
Write the intent into the spreadsheet column next to each URL. When two rows end up with the same query and the same intent, merge them before writing starts. It costs far less to merge a spreadsheet row than a published page.
Internal links are what turn a set of pages into a cluster. John Mueller, Search Advocate at Google, said in March 2022 that "internal linking is super critical for SEO" and called it "one of the biggest things" a site can do to guide Google and visitors to the pages it considers important.
Follow these rules for every cluster:
Links in the body text tell the reader why the next page matters, which a footer or sidebar list can't do. Once the cluster is live, crawl the site and confirm that no cluster page is an orphan and the pillar sits within a few clicks of the homepage.
Measure a cluster as one unit, because the pages succeed together. In Google Search Console, filter the Performance report by a URL pattern that matches the cluster, such as a shared folder, and track clicks and impressions for the whole group. Then check how many distinct queries the cluster ranks for. That number grows as the pages gain authority on the topic.
Look at the pillar on its own as well. If cluster pages rank but the pillar doesn't, check its links first, then whether it really answers the broad query. Tie the cluster to business results in your analytics tool by counting sign-ups or leads that start on cluster URLs.
Expect timing to vary with the domain's existing authority and the competition on the topic. HubSpot's research paper reports that the agency Campaign Creators tripled its organic traffic within 6 months of moving to clusters. Your own baseline is the only fair comparison, so record it before you publish.
Topic clusters help content get cited by AI because AI search breaks one question into many smaller searches, and a cluster has a page for each one. Elizabeth Reid, Head of Google Search, described the technique at Google I/O 2025. AI Mode, she said, uses Gemini "to break the question into different subtopics, and it issues a multitude of queries simultaneously on your behalf." Google's documentation says AI Overviews use the same query fan-out technique and that there are no extra technical requirements beyond being indexed and eligible for a snippet.
That changes which pages get picked. An Ahrefs study of 863,000 keyword results found that only 38% of cited pages also ranked in the top 10 for the original query. Ahrefs attributes part of the gap to fan-out queries. So a page can be cited because it answers a subquery well, even if it doesn't rank for the main query.
Write cluster pages so a model can lift the answer. Each section should open with the direct answer, then give the evidence. Ask ChatGPT and Gemini the questions your cluster covers, and note which sources they cite. Repeat the check with Perplexity.
The common topic cluster mistakes come from skipping planning steps, and every one of them is cheaper to prevent than to fix later.
When two cluster pages answer the same question, Google has to choose between them, and neither ranks well. This keyword cannibalization starts when old blog posts sit next to new cluster pages on the same query. Merge the weaker page into the stronger one and add a 301 redirect.
A pillar built on a long-tail phrase can't support a cluster. A pillar on a whole industry can't give a complete overview in one page. Apply Leslie Ye's test from the pillar section before you write a word.
Writers publish a cluster page and forget the link to the pillar, or the pillar never gets updated with the new page. The structure exists in the spreadsheet but not on the site. Add a link check to your publishing checklist and run a crawl every month.
Volume alone is a poor reason to build a cluster. If the topic doesn't connect to what you sell, the traffic won't convert, and the cluster pulls writing time away from topics that would.
The clusters that work share a tight scope and pages that link to each other on purpose.
HubSpot rebuilt its own blog around pillars after the 2015 research. Its marketing site still groups posts under pillar pages, each of which links to dozens of narrower articles. The same research paper documents Campaign Creators, whose organic traffic rose by over 500% within a year of adopting clusters.
The agency Noble Studios built pillar pages for a Fortune 500 insurance client around product lines such as commercial auto insurance. Between July 2017 and July 2024, the client's nonbrand traffic grew 520%, according to the agency's Semrush data.
Smaller sites see the effect too. Search Engine Land cites a Minuttia case where a single cluster ranked for more than 1,100 keywords. One cluster can cover that many queries because each page targets a different part of the topic and passes authority to the others.
Choose your first pillar topic this week. Run it through the planning order before you commission any writing. Start with the topic closest to what you sell, because that cluster pays back fastest. Before you plan pages, ask the main AI assistants the questions your topic clusters will cover. You'll see which competitors they cite now, and that shows you the gaps your cluster pages need to fill. Snoika tracks brand visibility in Google and AI assistants and builds researched content maps for clusters like these. Start with a free AI visibility report.
Topic clusters can answer the smaller questions that AI search generates from a broader query. A focused page may be cited for one of those questions even when it doesn't rank for the original query. Open each section with a direct answer, followed by evidence.
The number depends on how many distinct reader needs the topic covers. Give each different intent its own page, and combine queries that need the same answer. Published ranges vary, so don't add near-duplicate pages just to meet a target.
Common mistakes include publishing pages that answer the same question and leaving out links between the pillar and cluster pages. Compare planned pages with existing URLs before writing, and check links in both directions when publishing.
Yes, a spreadsheet and Google Search Console can support a first cluster on a small site. You'll need to group queries and map links manually. Paid research tools become useful when you need competitor keyword data at volume.
There's no fixed timeline because results depend on your domain's existing authority and competition for the topic. Record clicks and impressions before publication, then compare the cluster's performance against that baseline rather than treating another site's results as a deadline.
Keep reading
All articles


See how AI talks about your brand