Covering a subject properly rather than repeatedly
A topic cluster is one broad page introducing a subject, a set of focused pages covering its parts, and links tying them together so the whole reads as depth rather than repetition.
Learn
Short answer
A topic cluster is a group of related pages organized around one central page. The central page introduces the subject broadly and links to focused pages covering each part, which link back. It demonstrates depth on a subject and prevents pages competing with each other.
The shape
One page covers the subject broadly and links out to everything beneath it. Each focused page covers one part in depth and links back to the center.
Focused pages also link sideways to their closest siblings where a reader would naturally want both.
That is the whole structure. It is not complicated, and almost no small site does it deliberately.
It is not complicated and almost no small site does it deliberately, which is precisely why it is worth doing. Most sites have the pages already and have never connected them.
Why it works
It tells a search engine you cover a subject rather than mentioning it. Depth across linked pages is a far stronger signal than one long article.
It prevents cannibalization by construction, because each page has a defined part of the subject and the center does not compete with its own children.
And it works for readers, who arrive on one page from search and can see immediately that the rest exists.
Depth across linked pages is a far stronger signal than one long article, and it prevents cannibalization by construction.
It also makes the site easier to grow. A new page has an obvious home, an obvious set of links, and an obvious reason to exist, rather than being added to a pile.
Choosing the center
The center should be the subject as somebody would name it, broad enough to introduce everything beneath and specific enough to be one subject.
It is not a list of links. It should be readable on its own, giving somebody the overall picture and pointing them at the detail.
If the center is thin, the cluster reads as a directory rather than as coverage, and the center itself will not rank.
It is not a list of links. If the center is thin, the cluster reads as a directory and the center itself will not rank.
Choosing the parts
Each focused page should answer a question somebody actually asks, distinct enough that it would not be covered adequately inside the center.
Three or four focused pages is a cluster. One focused page is a section of the center that got promoted for no reason.
Resist splitting for the sake of page count. Two pages that overlap are worse than one page that covers both.
Three or four focused pages is a cluster. One is a section of the center that got promoted for no reason.
The linking
Centre links to every part, with wording describing what each one covers. Every part links back to the center, usually near the top.
Parts link to each other where a reader would want both, not exhaustively. Six links between six pages is a web nobody follows.
Use descriptive wording throughout, because those links are also how an engine understands what each page covers.
- Center links to every part, with wording describing each
- Every part links back to the center, near the top
- Parts link sideways where a reader would want both
- Not exhaustively — six links between six pages is a web nobody follows
- Descriptive wording throughout
- One primary home per page for the breadcrumb
How it differs from a blog
A blog is chronological and unstructured. Articles accumulate, overlap, and are found by date rather than by subject.
A cluster is organized by subject and stays organized as it grows, because every new page has a defined place.
You can run a blog inside this structure. The difference is whether an article joins a cluster or lands in a pile.
A blog accumulates chronologically; a cluster stays organized as it grows because every new page has a defined place — see site structure.
How ours are built
Sites here group pages into hubs, where the hub introduces the subject and lists its pages, and every page carries a breadcrumb and related links back into the group.
The structure comes from the plan rather than from tagging articles afterwards, which is what stops a cluster becoming a category with a label.
Every page passes a check that runs before anything publishes, validated against Google Search Essentials and modern web standards, and a page with no place in the structure does not publish.
The structure comes from the plan rather than from tagging articles afterwards, which is what stops a cluster becoming a category with a label on it.
Questions people ask
How many pages make a cluster?
A center and three or four focused pages is a working cluster. Fewer and the center should simply cover the subject itself.
Should the center page be long?
Long enough to introduce the subject properly and give somebody the overall picture. It should be readable alone, not just a list of links.
Can a page belong to two clusters?
It can be linked from both, with one primary home for its breadcrumb. Two parents in the trail confuses both readers and engines.
Is this different from site structure?
It is site structure applied to a subject. Your sections are the clusters, which is why deciding sections properly does most of this work for you.
Keep reading
See your website built from a conversation
15-day free trial. Card required. Cancel before day 15 and you pay nothing.