In short
- A semantic content cluster organizes pages around a topic, with a parent page and child pages linked according to strict rules.
- On an e-commerce site, the category tree is already a cluster: parent category, subcategories, facets.
- Blog posts have their place in the cluster, as supporting pages that link back to category pages, not as the center of gravity.
- Every page in the cluster must target its own intent: otherwise the cluster creates cannibalization.
The semantic content cluster is one of the most popular notions in French-language SEO: 880 searches per month in France. Popularized by Laurent Bourrelly, it consists in organizing pages around a topic, with a parent page, child pages and internal links that follow the hierarchy. On content sites, it has proven itself. On e-commerce sites, it is often applied in the wrong place: the blog.
The site tree is already a cluster
Look at your category tree through the lens of a cluster:
| Cluster level | E-commerce page | Example | Query targeted |
|---|---|---|---|
| Parent page | Category | Fitted sheets | fitted sheet |
| Child pages | Subcategories and facets | 160x200 fitted sheets, linen, deep pocket | 160x200 fitted sheet, linen fitted sheet |
| Supporting pages | Guides and blog posts | How to choose the pocket depth | what pocket depth for a fitted sheet |
Child pages carry the transactional long tail, as we show in our analysis of the long tail. Supporting pages answer questions and link back to category pages. The parent page concentrates authority.
This cluster is the most profitable, because all of its main pages sell. A cluster built only from blog posts mostly attracts informational traffic, which converts little, for the reasons detailed in our analysis of e-commerce SEO traffic.
Cluster rules applied to a retail site
One intent per page
This is the rule that matters most. The "fitted sheets" category targets "fitted sheet", the "160x200" facet targets "160x200 fitted sheet", the blog post targets a question. If two pages in the cluster target the same query, the cluster creates cannibalization instead of preventing it.
Links that follow the hierarchy
- The parent page links to its children: a block of popular subcategories and facets, in-text links.
- Children link to their parent: breadcrumbs, in-text link.
- Siblings link to each other sparingly: only when it helps the buyer (fitted sheets to duvet covers).
- Supporting pages link to the category pages they relate to, in the text.
The details of the practices are in our guide to internal linking.
Consistent vocabulary
Each page uses the vocabulary of its own topic. The parent page covers the broad topic and its main variations; each child page goes deeper into its variation without redeveloping its parent's or its siblings' topic.
Building the cluster for your category pages
- List the queries for each major category, from Search Console and a keyword tool.
- Group them by intent: the generic query for the parent page, attribute-based groups for the children, questions for the supporting pages.
- Map each group to an existing page, or one to create, often a facet.
- Link the pages according to the rules above.
- Write tags and copy that respect each page's scope.
Where unblind comes in
Step 5 is the most voluminous and the hardest to keep: every page must cover its topic without encroaching on the others. unblind assigns a main query to each page through a targeting plan, detects pages competing for the same queries, and writes copy that names the parent category and the displayed subcategories once, where it helps the buyer choose, so they carry your internal links, without developing their topic in their place.
Frequently asked questions
What is a semantic content cluster?
It is a way of organizing a site's pages around a topic: a parent page covers the main topic, child pages cover its subtopics, and internal links follow the hierarchy (children link to the parent, siblings to each other, sparingly). The concept was popularized in France by Laurent Bourrelly.
Is a semantic content cluster useful for an e-commerce site?
Yes, provided you apply it first to the category and facet tree, which carries the buying queries. A cluster built only from blog posts mostly attracts informational traffic, which converts little.