topic-cluster-design
This skill should be used when the user asks to "build topic clusters", "design topic clusters for SEO", "create a pillar and cluster strategy", "topic cluster strategy", "organize content into clusters", "content clustering for SEO", "hub and spoke content", "topical authority strategy", or any variation of designing, building, or organizing topic clusters for SEO and topical authority.
Topic Cluster Design
A topic cluster is a group of related pages organized around a central hub page. The hub page covers a topic comprehensively. Spoke pages cover specific subtopics in depth. Every page in the cluster links to every other relevant page. This structure signals to Google that your site has topical authority — deep expertise on a subject — which improves rankings for the entire cluster.
Topic clusters replaced the old approach of targeting individual keywords with individual pages. Instead of 20 disconnected blog posts about "lead scoring," you build one hub and 10 connected spokes that collectively dominate the topic.
Topic Cluster Architecture
Hub Page: "The Complete Guide to Lead Scoring"
(comprehensive, 2,000-4,000 words, targets head term)
│
├── Spoke: "What is Lead Scoring?" (definition, targets informational)
├── Spoke: "Lead Scoring Best Practices" (how-to, targets informational)
├── Spoke: "Lead Scoring Models: Demographic vs Behavioral" (comparison)
├── Spoke: "How to Set Up Lead Scoring in HubSpot" (tutorial)
├── Spoke: "How to Set Up Lead Scoring in Salesforce" (tutorial)
├── Spoke: "Lead Scoring vs Lead Grading" (comparison)
├── Spoke: "MQL Scoring Criteria" (deep-dive)
├── Spoke: "Lead Scoring Software Comparison" (listicle, MOFU)
├── Spoke: "Lead Scoring Mistakes" (common errors)
└── Spoke: "AI Lead Scoring" (emerging topic)
How linking works within a cluster
| From | To | Link type |
|---|---|---|
| Hub → Every spoke | All spokes | Contextual links in the hub's body content |
| Every spoke → Hub | Hub page | Contextual link back to the hub (usually in intro or relevant section) |
| Spoke → Adjacent spoke | 1-2 related spokes | Contextual link where relevant |
| External pages → Hub | Hub page | Backlinks should target the hub (concentrates authority) |
Designing a Topic Cluster
Step 1: Choose the topic
A good cluster topic meets all of these criteria:
| Criterion | Check |
|---|---|
| Your company has deep expertise | Can you write 10+ pages on this topic with unique insight? |
| Buyers search for this topic | Does keyword research show demand across multiple related queries? |
| It's core to your product | Is this topic directly related to what your product does? |
| You can differentiate | Can you add proprietary data, unique frameworks, or original POV? |
| It supports business goals | Will ranking for this topic drive pipeline? |
Step 2: Map the hub and spokes
Process:
- Identify the head term (the hub's primary keyword). Example: "lead scoring"
- Use keyword research to find all related subtopics (the spokes)
- Check search intent for each subtopic — this determines the spoke's page type
- Ensure no two spokes target the same keyword cluster (prevents cannibalization)
- Map each spoke to a keyword cluster with primary and secondary keywords
Cluster mapping template:
| Page | Role | Primary keyword | Volume | Intent | Page type |
|---|---|---|---|---|---|
| The Complete Guide to Lead Scoring | Hub | lead scoring | 5,400 | Mixed | Comprehensive guide |
| What is Lead Scoring? | Spoke | what is lead scoring | 2,900 | Informational | Definition page |
| Lead Scoring Best Practices | Spoke | lead scoring best practices | 720 | Informational | How-to guide |
| Lead Scoring in HubSpot | Spoke | lead scoring hubspot | 590 | Informational | Tutorial |
| Best Lead Scoring Software | Spoke | lead scoring software | 880 | Commercial | Listicle |
| Lead Scoring vs Lead Grading | Spoke | lead scoring vs lead grading | 320 | Informational | Comparison |
Step 3: Define the hub page
The hub page is the cluster's centerpiece. It must be the most comprehensive page on the topic on your entire site.
Hub page requirements:
| Requirement | Detail |
|---|---|
| Word count | 2,000-4,000 words |
| Coverage | Touches every subtopic covered by spoke pages (at overview level) |
| Linking | Links to every spoke page with contextual anchor text |
| Keyword | Targets the head term for the topic |
| Structure | H2s for each subtopic, with links to corresponding spoke pages for deeper reads |
| Schema | Article + FAQPage (if FAQ section included) |
| Update frequency | Quarterly — hub freshness signals apply to the whole cluster |
Step 4: Build and publish
| Phase | Action | Timeline |
|---|---|---|
| 1 | Publish the hub page | Week 1 |
| 2 | Publish 3-5 highest-priority spoke pages | Weeks 2-3 |
| 3 | Link spokes to hub and hub to spokes | On each publish |
| 4 | Publish remaining spoke pages | Weeks 4-8 |
| 5 | Update hub with links to all new spokes | After each batch |
| 6 | Build backlinks to the hub page | Ongoing |
Publish the hub first. Then add spokes. Each spoke strengthens the hub's authority. Each hub link strengthens the spoke's discoverability.
Cluster Sizing
| Topic breadth | Cluster size | Hub word count | Example |
|---|---|---|---|
| Narrow (1 concept) | 3-5 spokes | 1,500-2,000 | "Email deliverability" |
| Standard (methodology or category) | 6-12 spokes | 2,000-3,000 | "Lead scoring" |
| Broad (entire category) | 12-20+ spokes | 3,000-4,000 | "CRM for SaaS" |
Don't force it. If a topic only has 3 natural subtopics, build a 3-spoke cluster. Don't create thin spoke pages to inflate the cluster size.
Measuring Cluster Performance
| Metric | How to measure | Target |
|---|---|---|
| Hub ranking | GSC position for head term | Top 5 within 6 months |
| Spoke rankings | GSC positions for spoke keywords | Top 10 for 80%+ of spokes |
| Cluster traffic | Sum of organic traffic across all cluster pages | Growing 15%+ MoM |
| Internal link health | Screaming Frog → all cluster pages linked correctly | 100% of links in place |
| AI citations | Test hub's head term in AI engines | Cited in 2+ engines |
| Topical authority signals | When you publish a new spoke, does it rank faster than average? | New spokes ranking top 20 within 4 weeks |
Pre-Build Checklist
- [ ] Topic selected (meets all 5 criteria: expertise, demand, product-core, differentiation, business value)
- [ ] Hub keyword identified with volume and difficulty data
- [ ] All spoke keywords mapped (no two spokes target the same cluster)
- [ ] Search intent classified for each spoke
- [ ] Page types assigned per spoke based on intent
- [ ] Hub page structure outlined (H2s covering each subtopic with links)
- [ ] Cluster map documented (table with all pages, keywords, intent, page types)
- [ ] Publishing order defined (hub first, then priority spokes)
- [ ] Internal linking plan defined (hub↔spoke links, spoke↔spoke links)
- [ ] Backlink strategy defined for the hub page
- [ ] Measurement plan set (rankings, traffic, citations tracked per page)
Anti-Pattern Check
- Building spokes without a hub → Spokes without a hub are just disconnected pages. The hub is what ties the cluster together and concentrates authority. Always build the hub first
- Hub page is thin (500 words) → A hub that doesn't cover the topic comprehensively fails its purpose. Build a 2,000-4,000 word hub that touches every subtopic with overview-level coverage
- Two spokes targeting the same keyword → This creates cannibalization within the cluster. Each spoke must target a unique keyword cluster. If two subtopics overlap, merge them into one spoke
- No links between pages → A cluster without internal links is just a collection of pages. Link every spoke to the hub. Link the hub to every spoke. Link related spokes to each other
- Building a 20-spoke cluster for a narrow topic → If the topic only supports 5 subtopics, build 5 spokes. Don't create thin pages to fill a quota. Quality per page matters more than page count
- Never updating the hub → The hub should be refreshed quarterly. As you publish new spokes, update the hub to link to them. Hub freshness signals apply to the entire cluster
Want agents that use skill files like this?
We customize skill files for your brand voice and methodology, then run content agents against them.
Book a call