Key takeaways
- Depth compounds; breadth scatters. Cover one topic completely before starting another.
- The pillar targets the broad term; cluster pages target the specific sub-questions.
- Internal links must run both ways — pillar down to cluster, cluster up to pillar.
- A cluster only pays off if it connects to a page that actually generates enquiries.
The term "topic cluster" gets used loosely enough that it has stopped meaning much. Stripped of jargon, it describes a simple observation: search engines rank sites that cover a subject thoroughly more easily than sites that touch many subjects lightly.
Why depth beats breadth
A site that covers one subject completely — the main question, the sub-questions, the comparisons, the edge cases — accumulates credibility on that subject. New pages within it then start ranking faster, because the domain has already demonstrated it knows the area.
A site with thirty posts across ten unrelated subjects has demonstrated nothing in particular about any of them. This is why small sites almost always do better narrowing their focus than widening it.
The structure
- 01
The pillar page
A comprehensive page covering the whole topic at a useful level and targeting the broad head term. It links out to every cluster page.
- 02
The cluster pages
One page per meaningful sub-question, each targeting a specific long-tail search with its own intent. Each links back up to the pillar.
- 03
The commercial connection
Both pillar and cluster link into the service or product page that converts, so the authority they earn flows somewhere useful.
A concrete example
Take a dental clinic that wants to grow its implant work. The topic is dental implants; the commercial page is the implant treatment page.
| Page | Targets | Role |
|---|---|---|
| Dental implants: a complete guide | "dental implants", "dental implant treatment" | Pillar. Covers the whole topic and links to every cluster page. |
| Dental implant cost — what affects the price | "dental implant cost", "implant price" | Cluster. High commercial intent, usually the highest-volume sub-question. |
| Implant vs bridge: how to choose | "implant vs bridge" | Cluster. Comparison intent, captures people still deciding. |
| How long do dental implants last? | "how long do implants last" | Cluster. Reassurance intent, common pre-decision question. |
| Are dental implants painful? | "is implant surgery painful" | Cluster. Anxiety-driven, and heavily searched. |
| Implant recovery and aftercare | "implant recovery time", "aftercare" | Cluster. Post-decision, builds trust and topical completeness. |
| Who is a candidate for dental implants? | "am I suitable for implants" | Cluster. Qualifying content that also filters unsuitable enquiries. |
Seven connected pages on one subject. Each ranks for its own search, each reinforces the others, and all of them route to the page that produces appointments.
The internal linking rules
This is where most implementations fail. Publishing seven related articles is not a cluster if nothing connects them.
- Pillar links down to every cluster page, with descriptive anchor text
- Every cluster page links up to the pillar — this is the link most often forgotten
- Cluster pages link to each other where genuinely relevant, not indiscriminately
- Both link into the commercial page the cluster supports
- Anchor text describes the destination. "Dental implant cost" tells Google something; "read more" does not
- Do not link every page to every other page. Indiscriminate linking dilutes the signal it is meant to send
Choosing the right topic
A cluster is a substantial investment — six to ten pages properly written. Choosing the wrong topic wastes a quarter.
- 1Does it map to something you sell? Traffic on a topic you cannot monetise is a hobby
- 2Are there genuinely six or more distinct sub-questions? If you can only think of two, it is a page, not a cluster
- 3Can you cover it credibly? Clusters need real expertise. Thin coverage of a deep topic is worse than not attempting it
- 4Is the competition beatable? Check the live results for the pillar term — if the first page is national health portals, target a narrower pillar
- 5Would you finish it? A half-built cluster performs worse than a single good page
Common mistakes
| Mistake | Why it fails |
|---|---|
| Publishing the cluster pages but not the pillar | Nothing targets the broad term, and the cluster has no hub to concentrate authority. |
| A pillar that is only an index of links | It has no content of its own to rank with. A pillar must be a genuinely useful page. |
| Cluster pages that overlap | Two pages targeting the same sub-question compete with each other and destabilise both. |
| No link from cluster back to pillar | Authority flows down but never concentrates. The pillar stays weak. |
| Starting a second cluster before finishing the first | Neither reaches the depth needed to establish authority. |
| No connection to a commercial page | The cluster earns traffic that never reaches anything that produces enquiries. |
How long a cluster takes to work
Individual cluster pages targeting specific long-tail questions often start ranking within four to eight weeks, because the competition for them is thin. The pillar takes considerably longer — three to six months at minimum — because it targets the broad, contested term and depends on the cluster underneath it maturing first.
That sequencing is useful: the cluster pages produce early signals that tell you whether the topic was the right choice, well before the pillar has had a chance to move.
More on how this fits into a wider plan: content SEO.