Topic Clusters: The Pillar and Spoke Model
A topic cluster is one deep pillar page plus several narrower spoke posts that all link to it and to each other. Here is how to build one properly.
A topic cluster is one comprehensive pillar page on a broad topic, surrounded by several narrower spoke posts that each cover one sub-question in depth and link back up to the pillar. The pillar links to every spoke, and the spokes link to each other where relevant.
This is the structural half of an SEO content strategy: the plan decides what to write, and the cluster model decides how the pieces connect once they exist.
On this page
Why build a cluster instead of just publishing posts?
Because unlinked posts do not tell a search engine, or an AI assistant, that your site has depth on a subject. A single 800-word post about one narrow question reads as one opinion. Ten posts on related questions, all pointing back to one page that ties them together, reads as a site that has actually mapped a topic — which is closer to what topical authority means in practice: coverage a reader can trust because it is complete, not because any one page claims to be.
The linking also does real work for readers, not just for search engines. Someone who lands on a narrow spoke post because it answered their exact question is one click away from the pillar if they want the fuller picture, and one click away from a related spoke if their next question is adjacent. That path only exists if someone built it on purpose.
Depth also feeds the trust signal Google talks about directly. E-E-A-T — experience, expertise, authoritativeness and trust — is easier to demonstrate across a cluster than on a single page, because a reader (or a crawler) can follow the links and see that the coverage is not a one-off.
There is a practical ranking mechanic behind the model too, sometimes called link equity: a page that earns external backlinks passes some of that authority to whatever it links to internally. A pillar that attracts links from other sites — because it is the most complete resource on its topic — passes some of that strength down to every spoke it links to. A spoke published with no pillar, and no other page linking to it, has to earn every bit of its own authority from scratch.
How do you build the pillar first?
Start broad and comprehensive, then plan the spokes as the sections you deliberately left thin. A pillar that tries to be exhaustive on every sub-topic becomes unreadable; a pillar that is clearly a map — covering each area enough to be useful alone, while naming the deeper post for anyone who wants more — does both jobs.
Draft the pillar's outline first
List every sub-question a buyer or an AI assistant would fan out to on this topic. Each one is a candidate spoke.
Write the pillar section for each sub-question at medium depth
Enough to answer the question honestly, not enough to make the dedicated spoke redundant.
Publish the pillar, then write spokes against its outline
Each spoke expands one section the pillar covers briefly, and opens with a link back to it.
Come back and add spoke links into the pillar as each one ships
The pillar is a living document — its "what this covers" section grows as spokes publish.
How many spokes does a cluster actually need?
Enough that the internal linking has real material to work with — in practice, that is rarely fewer than four or five, and most effective clusters run eight to twenty spokes around one pillar. Below that floor, the pillar has too few links pointing at it from spokes, and each spoke has too few siblings to link sideways to, so the structure looks like a cluster in a diagram but behaves like a handful of disconnected posts in practice.
There is no ceiling that matters as much as the floor. A cluster can keep growing for years as new sub-questions appear, as long as each new spoke earns its place by answering something the pillar and the existing spokes do not already cover well.
Quality still outranks count. A cluster of six genuinely thorough, well-linked spokes beats a cluster of twenty thin ones every time a reader — or an AI assistant fanning out sub-questions — actually checks whether the answer is complete. If you are choosing between publishing a twelfth mediocre spoke and refreshing three existing ones that are starting to date, refresh the three; a content refresh cadence that keeps existing spokes current does more for the cluster's overall trust than a steady drip of new, thin ones.
| A real cluster | A folder of related posts |
|---|---|
| Every spoke links to the pillar with a descriptive anchor | Posts exist on similar topics with no deliberate links between them |
| The pillar names and links every spoke | The pillar, if one exists, was written after the fact as a summary |
| Spokes link sideways to siblings that genuinely relate | Any internal links are generic, added at the end as an afterthought |
| Each spoke answers one question the pillar left for it to cover | Several posts overlap on the same question with no clear primary |
How does this connect to what Google and AI assistants actually reward?
Indirectly, through depth and structure rather than through any special credit for the word "cluster" itself. Google's own guidance on site structure emphasizes clear organization and internal linking as a way to help both crawlers and readers understand how pages relate [1], and a well-linked cluster is simply a clean implementation of that advice at the scale of an entire topic rather than one page.
The same logic extends to AI-assisted content specifically. Clusters built by generating many thin, near-duplicate spokes with no real linking plan or editorial oversight are the pattern Google's guidance on AI-generated content treats as risk, not the cluster model itself — the model is fine; publishing at volume with no depth or human review under its name is what causes trouble.
AI assistants add a second reason the structure matters. When an assistant answers a broad question, it often reads several pages on a site to piece together a complete picture, the way a researcher would skim a table of contents before diving into one chapter. A pillar with a clear "what this covers" section functions exactly like that table of contents — it gives the assistant a map of what exists and where, which makes the whole cluster easier to cite from than the same content scattered with no page tying it together.
What actually belongs on the pillar versus a spoke?
The pillar earns the reader's trust that the topic is covered; the spoke earns their trust that one specific question got a real answer. Put definitions, an overview of the landscape, and a short answer to every sub-question on the pillar. Put the full method, the edge cases, and the worked example on the spoke. If a paragraph on the pillar is already doing what the spoke should do, the spoke will read as redundant and the pillar will run too long to finish.
A useful test: read the pillar's section for one sub-topic aloud. If it fully satisfies a reader who has no interest in going deeper, it is pillar-length and correct. If it leaves an obvious next question unanswered, that unanswered question is the spoke's actual title.
This split also solves a word-count problem that trips up a lot of first attempts at a pillar. Writers who try to make the pillar exhaustive end up with a six-thousand-word page that repeats, in less detail, everything the spokes were supposed to cover — and a reader who wanted the full pillar overview now has to scroll past paragraphs aimed at someone who wanted the deep version instead. Keeping the pillar at overview depth, with the deep material living only on spokes, keeps both documents doing the one job they are actually for.
A cluster is not a linking trick. It is what a site looks like when someone actually decided what they wanted to be known for and built the pages to prove it.
Questions people also ask
How is a topic cluster different from just writing a lot of blog posts?
The difference is the linking structure. A pile of posts on related topics that do not link to each other is still a pile of posts. A cluster is the same posts plus a deliberate link graph: every spoke links up to the pillar, and the pillar links to every spoke.
Does the pillar page have to be written first?
It should be. Spokes are written to support a specific argument the pillar makes, and a pillar retrofitted onto existing posts tends to read like a list of links rather than a guide with a point of view.
How many spokes does a cluster need before it is worth calling one?
There is no fixed number, but fewer than four or five spokes rarely shows a measurable lift, because the internal linking has too little to work with. Most effective clusters run eight to twenty spokes around one pillar.
Can two clusters share a spoke?
A spoke can link sideways into a second cluster's pillar as one of its outbound links without formally belonging to both. Making it a spoke of two pillars in frontmatter usually just means neither cluster fully claims it.
Sources
- SEO Starter Guide: The Basics — Google Search Central (checked 22 September 2026)
- Topic Clusters: The Next Evolution of SEO — HubSpot (checked 22 September 2026)
- Topic Clusters: What They Are & How to Create Them — Ahrefs (checked 22 September 2026)
Let us map your first cluster from a blank page
Book a free 30-minute consultation with Ibrahem and Jeremy. We look at your site while we talk, and you leave with a prioritised plan.
Who wrote this
Ibrahem HelalCo-founder
Leads strategy and delivery.
Works on: SEO, GEO
Reviewed before publishing by Jeremy Zava, Co-founder.
- SEO content strategy: the 2026 guideHow to plan, write, review, refresh and measure content that ranks in Google and gets cited by AI assistants, without tripping the spam policies.
- What is topical authority in SEO?Topical authority is how completely and consistently a site covers one subject. Here is how it is built, how long it takes, and what actually earns it.
- What is E-E-A-T in SEO?E-E-A-T is Google's shorthand for experience, expertise, authoritativeness and trust — a rating framework, not a ranking factor, and what it asks of a page.
- Google's AI Content Policy, ExplainedGoogle does not ban AI-written content. It bans content, AI or not, produced at scale mainly to manipulate rankings. Here is the actual line.