Topic cluster

What is a topic cluster?

A topic cluster is a set of pages about one subject, connected by links. One broad pillar page covers the whole topic, and shorter pages go deep on each subtopic. Every subtopic page links back to the pillar, and the pillar links out to each of them.

Let's say you sell revenue software for sales teams. Your pillar is a guide to sales forecasting. Around it, you publish pages on forecasting methods, how to forecast in a spreadsheet, how to forecast with a weighted pipeline, how to run a forecast call and a comparison of forecasting tools. Each page answers one question well, and together they cover the topic from basics to buying.

HubSpot made the idea popular. Its marketing team rebuilt the HubSpot blog around pillars and clusters, an approach its team credits to Matthew Barby, who led acquisition there at the time.

Why topic clusters matter for B2B SaaS

Google and AI tools judge whether your site knows a subject by how fully it covers it. One great article on sales forecasting is a start. A set of ten connected pages that answers every question a sales leader has sends a much stronger signal, and that's how you build topical authority.

AI search makes this even more true. Google's AI features run related searches across subtopics before they answer, a process called query fan-out. In a Surfer study of 10,000 keywords, pages that ranked for the main search and its fan-out searches were 161% more likely to be cited in AI Overviews. It shows correlation only, but the direction is clear.

Here's why I plan almost every content program in clusters:

• Each page has one job: When you map every subtopic first, you stop writing three articles that chase the same search. That keeps your pages from competing with each other.
• New pages rank faster: A new page that joins a strong cluster gets links and context from the pages already there. It doesn't start from zero.
• Readers keep going: A sales leader who lands on "how to run a forecast call" can click through to the methods guide and then to your tool comparison. That's a path toward a demo.
• Gaps become obvious: A cluster map shows you which questions you haven't answered yet, so your editorial calendar almost writes itself.

How to build a topic cluster

Most of the work happens before anyone writes a word. Here's my process:

• Pick a topic close to what you sell: It should be broad enough for at least eight subtopic pages and close enough that readers could become buyers. "Sales forecasting" works for revenue software, while "sales" is far too wide.
• List the subtopics from real searches: I research them with SEMRush, Google and Reddit. Google's People Also Ask box, related searches and Reddit threads show the questions buyers ask in their own words.
• Give each search intent one page: Group keywords that want the same answer onto one page. "Forecasting methods" and "types of sales forecasts" usually belong together, while "best forecasting software" needs its own page.
• Include a buying page: Every cluster should hold at least one page for buyers who are ready, like a comparison or alternatives page. Otherwise the cluster teaches people and sends them nowhere.
• Map the links before you publish: Write down which page links to which, with the anchor text. Then follow the plan with real internal links in the body text, where readers will actually click them.

A topic cluster example

Here's how the hypothetical revenue software company could map its sales forecasting cluster:

The pillar links to all five pages. Each one links back to the pillar, and the spreadsheet guide links to the software comparison, because someone tired of spreadsheets is the most likely buyer.

For real pillars, I like Gong's and Cloudflare's. We walk through those and ten more in 12 pillar page examples.

How to measure a topic cluster

Measure the cluster as one unit, since its pages lift each other. Here's what I track:

• Rankings across the cluster: In Google Search Console, filter by the cluster's URLs and count how many queries they rank for in the top ten. A healthy cluster ranks for more searches every month.
• Clicks and impressions for the group: Compare total clicks for all cluster pages quarter over quarter. One page dropping matters less when the group keeps climbing.
• Internal clicks: In GA4, check how many readers move from one cluster page to another. Low numbers usually mean the links are buried or the next step isn't clear.
• AI citations: Run prompts about the topic in ChatGPT, Perplexity and Google AI Mode each month. Count how often your cluster pages show up as sources.
• Visits to the buying page: Track how many sessions reach the comparison or alternatives page from inside the cluster, and how many of those book a demo.

Common topic cluster mistakes

• Choosing a topic that's too broad: A cluster on "marketing" never ends and never wins. Narrow it until a small team could cover it well in a few months.
• Splitting one intent across pages: Separate pages for "forecasting methods" and "forecasting techniques" end up competing. Merge them and redirect the weaker URL.
• Leaving orphan pages: A subtopic page with no links pointing to it is hard for Google and readers to find. Check new pages for links in and out before you call them done.
• Building only for learners: A cluster of guides with no comparison, alternatives or pricing page brings readers who leave without a next step.
• Publishing and forgetting: Clusters go stale as the topic changes. Update the pillar each time you add a page, and review the whole cluster at least once a year.

Here's a good first exercise. Pick the one topic your best customers asked about before they bought, and list every question they had. How many of those questions does your site answer today?