Topic Clusters: How to Build and Audit Them
Topic clusters organize related content around a broad pillar page, with focused pages for specific subtopics and internal links connecting them, as described in HubSpot’s topic cluster model. To build or audit a cluster, map the questions it covers, assign each page a clear purpose, and check for overlapping answers and missing links.
What a topic cluster contains
The model is also called a content cluster or hub-and-spoke model in HubSpot’s explanation. Its parts have different roles:
- Pillar page: the broad overview that introduces the subject and points to deeper coverage.
- Cluster pages: focused articles that answer specific questions within that subject.
- Internal links: connections from the pillar to its supporting pages, back to the pillar, and between relevant subtopics.
A pillar page is one part of the cluster; the whole cluster includes its supporting pages and their connections. Use the overview for orientation and the supporting pages for detailed answers.
Google’s SEO Starter Guide says logical site organization can help people and search engines understand how pages relate. For cluster planning, apply that principle by making each page’s scope and connections explicit. The same guide explains that SEO practices offer no automatic first-place ranking; evaluate a cluster by its coverage and navigation as well as its search performance.
How to plan a topic cluster
1. Define the subject and its boundary
Write a short scope statement identifying the subject, the questions included, and the adjacent subjects excluded. Choose a topic for which useful, well-supported coverage can be produced and maintained. Google’s people-first content guidance asks whether a site has a clear focus and whether its content gives people enough information to accomplish their goal.
Use those questions to set the boundary. Record which parts need a definition, a procedure, a comparison, or a troubleshooting answer. Keep adjacent subjects separate when their main questions fall outside the chosen scope, while allowing relevant links between them.
2. Group search terms by the answer they require
Collect relevant queries, then group them by meaning and intent. Google’s guidance on search terms explains that its language matching systems can relate a page to many queries without every exact wording appearing in the text. Apply that guidance by grouping synonymous queries before assigning URLs.
For each group, record the central question and the answer format it needs. Distinguish an explanation from instructions, a comparison, or a diagnosis. Treat keyword variations as candidates for the same page until the underlying answer justifies a separate one.
3. Decide what deserves a separate page
Assign one primary question to each proposed supporting page. Write down the answer it must provide, what belongs in the overview, and what belongs elsewhere. Keep enough context for the page to make sense independently, but place detailed background on the most appropriate page and link to it.
Use Google’s content-quality questions to check whether the proposed answer will be complete, accurately titled, and valuable beyond a restatement of other material. A separate URL should have enough substance to fulfill its own promise.
Set the cluster’s size through these page decisions. Add a page for a distinct question that needs depth; expand an existing section for a question that fits comfortably within its current scope. Make the page list follow the coverage plan rather than a predetermined publishing quota.
4. Map existing content before commissioning more
Create a working inventory with one row per existing or proposed URL. Include the following fields:
| Field | What to record |
|---|---|
| Primary question | The main question the page must answer |
| Page role | Overview, explanation, procedure, comparison, or troubleshooting |
| Scope boundary | What the page covers and hands off elsewhere |
| Existing coverage | Complete, incomplete, outdated, or duplicated |
| Internal routes | Pages that link in and useful destinations to link out to |
| Proposed action | Keep, revise, combine, relink, or create |
Read the existing pages before marking a gap. Record the actual answer, including useful sections hidden under an inaccurate title. Reuse or improve that coverage before adding another URL for the same question.
How to audit overlapping pages
Start with the content inventory, then add search evidence. The Search Console query and page dimensions allow performance data to be grouped and filtered by search term and URL. Use them to find pages that appear for similar searches and queries associated with unexpected pages.
Use a repeatable review sequence:
- Select a query and inspect its page rows within a stated date range.
- Open the URLs associated with that query and compare their titles, introductions, and main answers.
- Record each page’s distinct contribution and the sections they repeat.
- Decide whether the pages need clearer boundaries, stronger answers, or consolidation.
- Check the links into the pages before choosing the repair.
Treat shared queries as a reason to investigate. The editorial test is whether the pages provide different, useful answers. Different titles alone are insufficient grounds for keeping separate pages, and shared vocabulary alone is insufficient grounds for merging them.
Use the following decision guide after reading the content:
| Finding | Recommended repair |
|---|---|
| Distinct questions with complete answers | Keep both pages and clarify their relationship |
| Distinct questions with repeated background | Sharpen each scope and link to shared context |
| The same question with substantially repeated answers | Combine useful material into one complete answer |
| A useful answer with outdated information | Refresh the existing page and its supporting evidence |
| Existing coverage with no relevant incoming route | Add contextual internal links |
| A distinct question with no adequate answer | Expand a suitable section or create a focused page |
Before combining articles, select the destination URL and list the material worth preserving. Rewrite the resulting article as a coherent answer, then update its title, description, and incoming links. Retain specific instructions, sources, and distinctions that remain useful; remove repetition that adds no substance.
Review the whole answer before retiring a page. Low search visibility should prompt a content and access check, not an automatic removal decision.
How to fix missing internal links
Google’s internal linking guidance recommends at least one link from another page on the site to every important page. Check incoming links to each cluster URL to find pages with no internal route.
From the pillar, link to supporting coverage where the relevant subtopic is introduced. From supporting pages, link to the overview when broader context is useful. Add links between supporting pages where one answer directly leads to another question. Check the destination against the sentence containing the link.
Follow Google’s anchor text recommendations: use concise wording that describes the destination and fits the surrounding text. Google specifies no ideal number of links per page.
Inspect rendered links as well as visible navigation. According to Google’s crawlable-link guidance, links generally need an HTML <a> element with an href; script-only interactions cannot be reliably extracted. Open each destination and verify that it contains the promised answer.
When to use redirects and canonical tags
Resolve the content decision before changing URL signals. Google’s canonical documentation describes canonicalization methods for duplicate or very similar pages. Membership in the same topic cluster does not establish that pages are duplicates.
For a duplicate page being discontinued, Google’s method comparison recommends a permanent redirect. Use it after selecting and completing the replacement answer. For duplicate versions that remain accessible, a rel="canonical" annotation can indicate the preferred URL. Google describes these methods as signals expressing a preference.
Keep internal links and canonical declarations consistent with the preferred URL, following Google’s canonical best practices. Keep distinct answers independently available; use scope changes and relevant links to resolve their editorial overlap.
For consolidation, Google’s URL-change guidance permits redirects from old pages to the combined answer and warns against irrelevant destinations. The same guidance calls for HTTP 404 or 410 responses for removed content that is not being retained. Choose the URL treatment according to the replacement available, and remove outdated internal references.
How to measure and maintain a topic cluster
Save the URL inventory, query groups, date range, and planned edits before making changes. Record the completion date for content revisions, new links, merges, and URL changes. Review the same page set afterward, including the destination pages for any consolidated content.
Search Console’s Performance report overview identifies four available metrics: clicks, impressions, click-through rate, and average position. It also explains how to change dates and grouping dimensions. Compare search visibility and traffic at both the page and query level, then read the results alongside the content changes.
Allow for the report’s limits. Google’s data-grouping documentation says that some queries are anonymized, not all rows are shown, most performance data is assigned to canonical URLs, and recent periods can contain preliminary data. A missing row does not establish that a subject has no demand.
Use comparable, complete periods and note other changes affecting the site. Google’s SEO Starter Guide says changes can take hours to months to be reflected and recommends allowing weeks before assessing their effects. Interpret later movement as evidence to review, without attributing every change to the cluster edits.
Repeat the coverage and link checks when adding or substantially revising a page. Confirm its primary question, identify the existing answer closest to it, update the relevant connections, and recheck overlap. Keep the inventory current so the next publishing decision starts with the coverage already available.