Topic clusters instead of single pieces: how to occupy a field

The usual approach is a topic list: forty ideas worked through in whatever order they occur to you. The result is forty pages, each standing alone — and several of which compete with one another because they phrase the same question differently.

A cluster reverses that: instead of breadth, one field gets occupied in depth.

How a cluster is built

Defining the scope

The hardest part, and the one where most clusters fail. Three questions decide it:

1. Is the field narrow enough? "Marketing" is not a cluster but an industry. "Email deliverability" is a cluster. Rule of thumb: if you cannot name six to twelve concrete questions off the top of your head, it is too broad or too narrow.
2. Do you have something of your own for every piece? A cluster where three of ten pieces only summarise what exists elsewhere is a cluster with three weak points — and they drag the whole thing down.
3. Does it fit what you sell? A cluster occupies a field for years. If the field is only indirectly connected to your offering, the visibility brings little.

The order: from the inside out

The obvious route — overview first, then the parts — is the worse one. This sequence works better:

Step four is nearly always skipped, and it is the one that turns a collection into a cluster.

A cluster, scoped

Using "email deliverability" as an example — a field narrow enough and with enough concrete questions:

• Overview: why emails do not arrive and what helps
• What SPF, DKIM and DMARC actually do
• How to tighten DMARC step by step without losing mail
• How to tell your own domain is being abused as a sender
• Why a grown recipient list lowers the delivery rate
• What to watch when changing sending service
• How to warm up a new sending domain
• Which phrases actually trigger spam filters — and which do not

Seven individual pieces, each with a clearly bounded question, no overlap. Together with the overview that makes a field a small company can genuinely occupy — because the concrete questions are rarely answered fully anywhere.

How to tell it is holding

Three signs, in this order:

After two to three months: the first individual pieces appear in Search Console with queries you did not plan for. That is the first signal the field is being recognised.

After four to eight months: the overview page gains search visitors it did not have before. That is the actual proof the cluster is working — the overview benefits from the individual pieces.

After eight to twelve months: new pieces in the same field rank faster than earlier ones did. That is the point at which the effort starts paying back.

Help me scope a topic cluster.

Starting point:
- What we offer: [offering]
- Audience: [as narrow as possible]
- Field I want to occupy: [topic]
- What we know from our own work: [bullet points, figures, cases,
  recurring customer questions]
- How many pieces we manage per quarter: [number]

Tasks:
1. Judge whether the field suits a cluster: too broad, too
   narrow, or right. Justify it.
2. For every planned piece, phrase the one question it answers –
   in exactly one sentence. Then name every overlap between two
   questions and propose which pieces belong together.
3. Name the 2 to 3 individual pieces we should start with – the
   most concrete ones with the least competition.
4. Draft the outline of the overview page, with one section per
   individual piece.
5. For each planned piece, tell me whether our own substance is
   sufficient – and if not, which detail I am missing.

Do not invent search volumes or difficulty scores.

In closing

A cluster is a decision about a year, not about a piece. In return you get something single pieces cannot deliver: content that supports itself instead of getting in its own way, and a field where new material takes hold faster.

The three things most often missing: a sharp scope with no overlaps, starting at the concrete questions rather than the overview, and revising the first pieces once the rest stands.

Common questions