Definition
A content cluster is a group of pages that covers one topic together: a pillar page that gives a broad overview and several cluster pages that each go deep on one subtopic, all connected by internal links. The model is also called hub and spoke or pillar-cluster. How the groups are found and assigned is covered under content clustering.
HubSpot popularized the structure in 2017 after testing it on its own blogs. A pillar on a topic such as Instagram marketing linked out to posts on profiles, captions and Stories, and each post linked back to the pillar. The aim was a cleaner architecture in which similar posts stopped competing for the same keywords and search engines could see how pages related.
A cluster has three working parts. The pillar page answers the broad question and links to every subtopic. Cluster pages each answer one narrower question well enough to stand alone. The links between them use descriptive anchor text, so a crawler reading any page can find the rest of the topic. A cluster on cloud migration might have a pillar plus pages on cost estimation, a security checklist, a provider comparison and a rollback plan.
In AI search, clusters matter mostly because of query fan-out. ChatGPT, Perplexity and Google's AI Overviews break one question into several searches and cite pages that rank for those sub-searches. Ahrefs found in March 2026 that only 38% of URLs cited in AI Overviews ranked in the top 10 for the original query, down from 76% in July 2025, and suggested that more citations now come from fan-out results (Ahrefs). A cluster whose pages cover the likely sub-questions has more chances to be retrieved than a single article. It also supports topical authority, since a site that covers a subject thoroughly is easier to recognize as a source on it.
Clusters are not a page-count exercise. Google's May 2026 guidance for its AI features says its systems can find the relevant passage on a page that covers several topics, and that making pages for every query variation can violate its scaled content abuse policy. A subtopic deserves its own page when it has its own intent, enough substance to answer properly, and a reason for a reader to land on it directly. Otherwise it belongs as a section of the pillar.
Clusters also need upkeep, because pages decay at different speeds. In Promptwatch's ChatGPT content half-life study, published in October 2026, ChatGPT still cited 72% of comparison pages 30 days after first citing them, while half of news pages had dropped out after 36 days. Track rankings and citations per cluster rather than per URL, check which page gets cited for each prompt, and refresh the pages that slip first with content freshness in mind.
Examples of Content Clusters
- A cybersecurity consultancy builds a small business security cluster: a pillar plus pages on phishing training, backup policies, cyber insurance requirements and incident response. Within a quarter, Perplexity cites different pages of the cluster for different sub-questions instead of citing one competitor guide for all of them.
- A payroll software company finds its 6,000-word payroll guide is retrieved but rarely cited for pricing questions. It moves the pricing material into a dedicated page linked from the pillar, and that page starts appearing in ChatGPT answers about payroll costs.
- A nutrition clinic's family meal planning cluster has dietitian-reviewed pages by age group. The pillar links to each page and each page links to two siblings, so crawlers reach every page within two clicks of the pillar.
- A running shoe retailer audits its cluster and merges four overlapping posts about shoes for flat feet into one page, ending the internal competition for the same prompts.
