Topical authority clusters

Audience: Merchant·Plan: pro·Read: 8 min·Updated: 2026-08-03

In 30 seconds

• A cluster is a programme, not a document. Every page in the plan becomes a trackable row with a status you move through.
• Gaps are grounded — diffed against your real catalogue, your existing content and your actual rankings, not generated from the model's assumptions.
• The authority score is explainable: coverage, ranking, depth and interlinking, each shown separately.
• Execute from the cluster: generate real drafts into the content library with maker-checker review, add pages to the content calendar, and apply hub-and-spoke internal links.
• Measure per cluster: clicks, revenue and average position, with a next-best-action.

How to use it

1. Generate a cluster plan for a topic, choosing a region and a size.
2. Review the grounded gaps — the pages you do not have yet.
3. Build drafts for the gaps worth writing, or schedule them on the calendar.
4. Apply hub-and-spoke links so the cluster is connected, not just written.
5. Come back to the measurement panel to see whether the topic is gaining ground.

⏱ ~20 min per cluster · 💳 Pro+ · 🎯 Own a topic, not a keyword

What a cluster is

A topical cluster is a hub page plus the spokes that support it. The plan enumerates both, and each page in it is a row you can track:

Page state	Meaning
Planned	In the plan, nothing written
Drafted	A draft exists in the content library
Scheduled	On the content calendar
Published	Live
Existing	You already had a page covering this

Sizes range from a focused five-page cluster to a twenty-page programme, and a cluster is scoped to a region so that a US and an India plan for the same topic do not collide.

Grounded gaps

The difference between this and a generated content outline is grounding.

A page is only a gap if you genuinely do not cover it. Pages you already have — whether as products, collections, blog posts or ranked URLs — are marked as existing and linked, rather than proposed again as new work. That is the difference between a plan you can execute and a list you have to manually de-duplicate.

The authority score

The score is decomposed rather than delivered as a single opaque number:

Component	What it measures
Coverage	How much of the planned cluster actually exists
Ranking	How well the cluster's pages rank for their terms
Depth	Content depth across the cluster
Interlink	How well the cluster's pages link to each other

Each component is shown separately, so "we have written everything and still score badly" resolves into a specific answer — usually interlinking, which is the component most often skipped.

Executing from the cluster

The point of the cluster is that it does things:

Action	What happens
Build drafts	Generates real drafts into the content library, entering the normal maker-checker review flow
Add to calendar	Schedules a planned page on the content calendar
Apply internal links	Applies hub-and-spoke links between the cluster's pages using the internal link platform, with its policy guardrails
Enrich	Re-runs grounding against your current catalogue and rankings
Update page status	Moves a row through the states by hand

Draft generation goes into the content library, which means it inherits approval rules, brand voice and versioning — nothing is published to your storefront straight from a cluster plan.

Hub-and-spoke linking runs through the internal link platform, so the same guardrails apply: root-relative hrefs, inbound and outbound caps, anchor-repetition limits.

Measuring a cluster

Per cluster, over the recent Search Console window:

• Total clicks
• Attributed revenue
• Average position
• A next-best-action, derived from which authority component is weakest

Lifecycle

Clusters can be archived, restored and deleted. Generating a plan for a topic you already have a cluster for de-duplicates rather than creating a second one, and re-enrichment updates an existing plan against your current data instead of replacing it.

Plan tiers

Capability	Free	Starter	Pro	Agency	Enterprise
Cluster plan generation	—	—	✓	✓	✓
Grounded gap diff	—	—	✓	✓	✓
Explainable authority score	—	—	✓	✓	✓
Draft generation into the content library	—	—	✓	✓	✓
Calendar scheduling	—	—	✓	✓	✓
Hub-and-spoke internal linking	—	—	✓	✓	✓
Per-cluster clicks, revenue, position	—	—	✓	✓	✓

Cluster generation and enrichment make model calls, so the whole surface sits at Pro with the other AI-backed SEO tools.

Best practices

• One cluster per commercial topic, not per keyword. If two clusters share most of their spokes, they are one cluster.
• Enrich before you plan the quarter. Grounding against a stale catalogue proposes work you have already done.
• Do not skip the linking step. The interlink component is usually what separates a cluster that ranks from a folder of posts.
• Publish the hub first. Spokes linking to a hub that does not exist yet is the most common way a cluster stalls.
• Measure at 8–12 weeks. Topical authority is the slowest-moving thing in this suite; weekly readings are noise.

See also

• Content gap — single-page gaps rather than whole topics
• Internal links — the linking engine clusters use
• Blog engine — where cluster drafts are written
• Content calendar — scheduling the programme
• Keyword portfolio — the terms a cluster is built from