One Cluster, One Page, Decided Before You Write

One Cluster, One Page, Decided Before You Write

Treats each keyword cluster as a hypothesis about a single page: page type before title, existing URL before new one.

scalenut seo keyword-research clustering site-architecture planning

What it is for

Keyword clustering produces neat groups, and the neatness hides the real question: is this one page, and what kind of page. This checks each cluster against what actually ranks, fixes the page type before the title, and sends matching intents to existing URLs. Use it in the gap between the keyword report and the content calendar.

The skill

A keyword cluster is a hypothesis about one page. Test it before it becomes
a row in the publishing calendar.

For each cluster:
1. Read the keywords aloud. If you would answer two of them differently, for
   two different people, it is two clusters. Split it.
2. Search the main keyword yourself and look at what actually ranks. If the
   results are product or category pages and your plan says blog post, the
   cluster is fine and the page type is wrong.
3. Decide the page type BEFORE the title: guide, comparison, product,
   category, calculator, or short answer page. Most cluster plans fail here
   and never recover.
4. Pick the single keyword that IS the page. The rest are sections, not
   phrases to be crammed into the introduction.

Then build one row per cluster:
  cluster | page type | new or existing URL | intent in one sentence | owner

Rules:
- Two clusters never get the same URL. If they want the same URL, they were
  one cluster and the grouping was too tight.
- An existing URL beats a new one whenever the intent matches. New URLs are
  a cost you pay forever, not an achievement.
- Do not size a cluster by total volume. One keyword with real buying or
  fixing intent beats forty phrasings of the same idle question.
- A cluster with no commercial or navigational purpose can still be worth
  writing, but name the page it feeds. If it feeds nothing, drop it.
- Anything left over after mapping is not a backlog. It is noise the tool
  produced, and keeping it makes the plan look bigger than it is.

Failure this prevents: forty articles mapped from forty clusters, where half
are the same page written twice and a third should have been one paragraph
added to a page that already exists.

How to install it

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from

Skills app Original