Thematic clusters

A topic cluster (English: Topic Cluster) is a method of organising content in which a subject area is covered not by individual, isolated pages, but by a coherent group of linked content. At the centre is a comprehensive pillar page on the main topic, around which several specialised cluster articles (spokes) are arranged, each dealing with a specific aspect in depth. The pillar and spokes are interconnected via internal links. This architecture signals to search engines that a domain has comprehensive expertise in a topic – the foundation for topical authority.

The topic cluster model is the response to a fundamental shift in how search engines operate: away from ranking individual pages for individual keywords, towards assessing a domain’s expertise across entire topics. The approach was popularised in 2017 by HubSpot, which described the concept in its report ‘Topic Clusters: The Next Evolution of SEO’. Since then, it has evolved from the specialist knowledge of individual SEO teams to become the standard for modern content strategies – particularly in the B2B sector, where complex topics raise many interrelated questions.

Why topic clusters work

Search engines such as Google now evaluate content semantically. They understand that different phrases such as ‘moving an online shop’, ‘migrating a shop system’ or ‘switching to a new platform’ all refer to the same issue, and they measure how comprehensively a domain covers a subject area. A single, isolated page on a topic appears to this system as a fluke. A group of related, logically linked pages that examine a topic from every relevant angle, on the other hand, conveys proven expertise.

The second mechanism at work is internal linking. When every ‘spoke’ links to the ‘pillar’ page and the ‘pillar’ page links to every ‘spoke’, a clear hierarchical signal is created: the ‘pillar’ is the central authority page, and the ‘spokes’ support it. At the same time, link equity (PageRank) is distributed strategically within the cluster, which helps the pillar page to rank for the highly competitive head keyword. For human readers, this simultaneously creates a clear path through the topic.

Topic clusters and AI search

With the emergence of AI-powered search results – Google AI Overviews have been rolled out across the DACH region since 26 March 2025 – the cluster model takes on a second significance. AI systems do not primarily evaluate individual pages, but rather how consistently a domain demonstrates expertise on a topic across many pages (entity authority). A cohesive cluster architecture provides precisely this signal and thus becomes a prerequisite for being cited as a source in AI responses at all. Topic clusters are therefore no longer just an SEO bonus, but the foundation for AI visibility.

The three building blocks of a topic cluster

A complete topic cluster consists of three elements, which must be clearly distinguished:

• Pillar Page (Hub): The central page for the main topic. It covers the entire topic in depth (typically 2,500–4,000 words) and links to each spoke. It serves as both the content and the table of contents for the cluster.
• Spokes (cluster articles): Individual articles, each covering a sub-topic (typically 800–1,500 words). Each spoke comprehensively addresses a specific search intent and links back to the pillar.
• Internal linking: The connections between the pillar and the spokes. Without them, the cluster does not exist as a cluster for a search engine. A spoke without a link to the pillar is an ‘orphan’ fighting on its own.

Structuring topic clusters correctly

The most common mistake when clustering occurs at the outset: structuring clusters based on keywords rather than search intent. ‘Shopware costs’, ‘Shopware prices’ and ‘Shopware licence’ are not three separate topics, but the same search intent expressed in three different ways. If these are turned into three spokes, keyword cannibalisation occurs: your own pages compete for the same position, and the search engine doesn’t know which one to choose.

The correct guiding question is instead: What distinct questions does someone have when dealing with the main topic – questions that each deserve their own complete answer? For a medium-sized cluster, eight to fifteen such subtopics is a healthy range. Fewer than five rarely provide enough depth to signal authority; significantly more than twenty often suggests that two clusters are combined into one and would be better separated.

Practical example: a topic cluster on ‘Shopware migration’

A concrete example from the e-commerce sector illustrates what a cluster looks like. A service provider wishing to rank for the topic of Shopware migration creates a pillar page entitled ‘Shopware migration: the complete guide from SW5 to SW6’, which outlines the entire migration process. Spokes for the individual search intentions are grouped around this:

Each of these questions is given its own in-depth ‘spoke’, and all spokes lead back to the ‘pillar’. By covering topics in this way, you signal the same message to search engines over a period of months: this domain understands Shopware migration. This is topical authority in action.

Advantages of topic clusters

Compared to a collection of isolated individual articles, the cluster model offers several concrete advantages that grow over time:

• Stronger rankings: The bundled topic coverage and internal linking particularly help the pillar page to rank for highly competitive top-tier keywords that a single page would struggle to achieve.
• Protection against cannibalisation: By aligning with search intent rather than keyword variations, you avoid your own pages competing against one another.
• Better user guidance: Readers can explore the entire topic via the pillar page and find the relevant detailed answer via the spoke pages – a clear path rather than scattered fragments.
• Plannable expansion: The topic map highlights which content is still missing, enabling the editorial team to grow in line with genuine demand rather than gut instinct.
• Lasting impact: A cohesive cluster grows stronger with every added ‘spoke’ and every refresh, rather than gathering dust in the archive like a single article that’s only relevant for a day.
• Foundation for AI visibility: Coherent topic coverage provides the signal that AI systems use to cite a source.

Common SEO tools help with implementation: keyword tools for researching sub-topics and their search volumes, Google Search Console for measuring gaps and success, and tools for analysing internal linking that uncover orphaned pages and untapped linking opportunities. However, none of these replaces the work on the content at the cluster level – the real effort lies not in the individual text, but in the well-thought-out structuring of the topics.

Common mistakes with topic clusters

• Orphan content: Spokes that have been written but never linked to – invisible to the cluster logic.
• Keyword cannibalisation: Multiple spokes targeting the same search intent, which weaken one another.
• The thin pillar: A pillar with a perfect heading structure but too little substance per section – a skeleton without content.
• No refresh: A cluster that is never updated and becomes outdated, particularly for technical B2B topics with frequent version changes.
• Too many clusters at once: Five clusters started in parallel often end up as five half-finished ones. One complete cluster is better than three incomplete ones.

Building a topic cluster: the steps

Building a topic cluster follows a clear sequence. If you stick to it, you’ll avoid most typical mistakes from the outset:

1. Choose a core topic that fits your business. It is not the topic with the highest search volume that wins, but the one with the greatest overlap between genuine search demand and relevance to your own offering. A perfectly ranked cluster with no business relevance generates traffic, but no qualified enquiries.
2. Map subtopics and search intentions. Create a topic map: place the main topic in the centre, surrounded by the subtopics along with the actual search queries they address. The map reveals gaps and duplicates before a single line is written.
3. Plan the pillar. Structure the pillar so that each main section corresponds to a subtopic and links to the associated spoke.
4. Write the spokes. Write each spoke so that it answers its question on its own – even if a reader lands on it directly from a search and never sees the pillar.
5. Create internal links. Link from the Pillar to each Spoke, and from each Spoke back to the Pillar, using descriptive, varied anchor text.
6. Measure and expand. After eight to twelve weeks, the data will show which spokes are performing well and where there are gaps. These gaps form the roadmap for the next phase of expansion.

Topic clusters in B2B

In B2B marketing, topic clusters demonstrate their strength particularly clearly. B2B topics are often complex, require explanation and raise many interrelated questions – ideal conditions for a cluster. A decision-maker facing a platform migration, a software implementation or a marketing investment typically goes through several phases of information-seeking, from initial orientation to the specific selection of a provider. A well-structured cluster accompanies this entire journey: The pillar provides the overview, whilst the spokes answer the specific, detailed questions that arise throughout the decision-making process.

Furthermore, B2B search queries often have a lower volume but a higher purchase intent. This is precisely where comprehensive coverage pays off: By addressing even the niche, specific queries with dedicated spokes, you reach precisely those qualified prospects who are on the verge of making a decision – whilst simultaneously building the subject authority that makes your content visible across a broad range of queries.

Frequently asked questions about topic clusters

How many articles does a topic cluster need?
A medium-sized cluster typically comprises one pillar page and eight to fifteen spokes. It is not the number that matters, but the comprehensiveness: every distinct search intent relating to the main topic should have its own spoke.

What is the difference between a topic cluster and a hub-and-spoke model?
Both terms refer to the same architecture. ‘Topic cluster’ emphasises the grouping of content, whilst ‘hub-and-spoke’ emphasises the structural form (hub and spokes). The hub is the pillar page; the spokes are the cluster articles.

Do I have to start from scratch for a topic cluster?
No. Most B2B blogs already have several half-formed clusters scattered about. Taking stock of existing posts by topic often reveals enough material for a pillar and several spokes, which simply were never conceived of or linked as a single unit.

How do I measure the success of a topic cluster?
Via Google Search Console: after eight to twelve weeks, the data will show which ‘spokes’ are ranking and for which search queries impressions are being generated without a matching page existing. These gaps form the roadmap for further expansion.

Further background on the concept is described in the original HubSpot report on topic clusters.