Content strategy for B2B: Building topic clusters and a hub-and-spoke structure effectively

More blog posts do not lead to better rankings. This reality costs B2B marketing five-figure budgets every year, because the logic is so tempting: one more article, one more keyword, one more editorial slot. What you end up with is an archive of 80 or 120 individual posts that exist side by side thematically and steal visibility from one another. Google sees no authority. Google sees a hodgepodge.

The solution isn’t to publish more frequently, but to adopt a different architecture. A B2B content strategy that will still be effective in 2026 organises content into topic clusters according to the hub-and-spoke principle: so that search engines recognise that this domain has in-depth expertise on a particular topic. It is precisely this signal, topical authority, that determines whether you rank. And, more recently, it also determines whether an AI-generated response will quote you at all. This guide shows exactly how the structure works, where it often falls short in practice, and how to turn an existing blog archive into a cluster without having to start from scratch.

Why individual blog posts no longer rank

For a long time, SEO was a one-to-one game: one keyword, one optimised page, one ranking position. Anyone wanting to rank for ‘Shopware migration costs’ would build a page optimised specifically for that phrase. That no longer works, and the reason is not a trend but a fundamental overhaul of the search algorithm.

Google now evaluates content semantically. The search engine understands that “move online shop”, ‘switch from SW5 to SW6’ and ‘migrate shop system’ all refer to the same concept, and it no longer evaluates individual pages in isolation, but rather how comprehensively a domain covers a subject area. This is called topical authority: a signal at domain level, not at page level. A single brilliant page on a topic, surrounded by silence, appears to this system as a fluke. Twenty related pages that examine a topic from every relevant angle and are logically interconnected convey an impression of expertise.

HubSpot popularised this model in 2017 with the Topic Cluster approach, in direct response to this very shift: away from ranking individual pages for individual keywords, towards assessing authority across entire topics. Since then, the approach has gone from being a marketing insider’s tip to the industry standard. AI-powered search is currently giving it a second, more significant meaning, which we’ll return to at the end of this post.

In short: search engines no longer rank texts; they rank subject matter expertise. And expertise cannot be packed into a single article.

Hub-and-spoke in a nutshell

Imagine a wheel. In the centre is the hub, with spokes radiating outwards. Applied to content: the hub is a comprehensive page on a major topic, the pillar page. The spokes are individual articles, each of which covers a specific aspect of this topic in depth. Each spoke links back to the pillar page, and the pillar page links out to each spoke. This reciprocal linking isn’t just an afterthought; it’s the core mechanism.

What this image deliberately does not show is the connection between the spokes across the wheel. This is intentional. The linking takes place primarily via the hub, not criss-crossing; otherwise, rather than a clear signal of authority, a tangled mess would result. The structure is hierarchical, and it is precisely this clarity that a search engine can read.

Pillar, Cluster, Spoke: the three building blocks

Three terms that are constantly getting mixed up. And as long as they are, you’re building the wrong thing. So, let’s make a clear distinction.

The Pillar Page is the central page for the main topic. It covers the entire topic broadly, without going into every sub-aspect down to the last detail; that’s what the Spokes are for. In practice, a Pillar Page is long: A typical length is 2,500 to 4,000 words (depending on the topic and competition), because it needs to cover enough ground to serve as a comprehensive go-to resource. The article you’re reading right now is itself designed as a Pillar on the topic of content strategy.

The cluster is the combination of the pillar and all associated spokes – in other words, the whole wheel. When someone talks about ‘building a cluster’, they mean this complete unit.

The spoke (also known as a cluster article or cluster content) is a single post on exactly one sub-topic. It goes into depth where the pillar only touches on the subject, and is typically between 800 and 1,500 words. A good spoke provides an exhaustive answer to a specific search query and refers back to the pillar for the broader context.

It may sound like nit-picking, but it isn’t: anyone who confuses the Pillar and the Spoke will either end up with a Pillar that doesn’t rank for anything, or a pile of Spokes without an overarching structure. The connecting element between all three is internal linking. It is not a decorative element that you ‘just throw in’ at the end, but the wiring without which the cluster does not even exist as a cluster for a search engine. A ‘spoke’ without a link to the ‘pillar’ is technically an orphan: a lone fighter.

Structuring topic clusters correctly: from keyword to topic

The most common mistake when clustering happens right at the start, when asking what a cluster actually is. If you think in terms of keywords, you’re on the wrong track. “Shopware costs”, “Shopware prices” and “Shopware licence” aren’t three different topics, but the same search intent expressed in three different ways. Turning these into three ‘spokes’ leads to keyword cannibalisation: your own pages compete for the same ranking position, and Google doesn’t know which one to choose.

A topic should be segmented based on search intent, not on search terms. The key question is: What distinct questions does someone exploring this overarching topic have – questions that each deserve their own complete answer? For ‘Shopware migration’, these might include: What are the technical differences between SW5 and SW6? How do I migrate my product data? What happens to my plugins? How do I safeguard my SEO ranking during the migration? How much does the whole thing cost? Each of these questions is a separate ‘spoke’, because it has its own self-contained answer, and because people actually Google them individually in this way.

In practical terms, this means: before a single line is written, there’s a topic map. The main topic goes in the centre, surrounded by the subtopics, and for each subtopic, the actual search query it addresses. This map immediately reveals two things: gaps (sub-topics you haven’t covered yet) and duplicates (two planned ‘spokes’ that actually mean the same thing). You want to know both of these before you start writing, not afterwards in the performance report. A poor structure costs you twice over: you produce content that pushes each other out of the top 10, and only realise this after months in which two of your own pages have been competing for the same position, instead of holding one position together.

The specific structure in six steps

Enough theory. Here’s how to build a cluster from scratch.

1. Choose a core topic that suits your business. It’s not the topic with the highest search volume that wins, but the one where high relevance to your offering coincides with genuine search demand. For a Shopware service provider, ‘Shopware migration’ is a strong core topic, whereas ‘E-commerce trends 2026’ is merely a high-volume decoy with no business relevance. Rule of thumb: if a perfectly ranked cluster on this topic wouldn’t bring you any qualified enquiries, it’s the wrong topic.

2. Map out sub-topics and their search intentions. This is where the topic map from the previous section comes into play. Eight to fifteen sub-topics is a healthy range for a medium-sized cluster. Fewer than five rarely provide enough depth to signal authority; more than twenty is often a sign that two clusters are combined into one and would be better separated.

3. Plan the pillar pages, not just the topic. The pillar page is the most demanding page in the cluster because it must comprehensively outline the main topic whilst also linking to every spoke. Plan its structure so that each main section corresponds to a sub-topic: the section introduces it, provides the core answer and links to the relevant spoke for further detail. The Pillar is thus both the content and the table of contents of the cluster.

4. Write the spokes, each one complete in its own right. A spoke must be able to answer its question on its own, even if someone lands on it directly from a Google search and never sees the pillar. ‘Read the pillar first’ is not an acceptable answer. Each spoke includes a link back to the pillar at a natural point in the text, and ideally one or two links to thematically related spokes, where the content justifies it.

5. Link internally – in a structured rather than decorative way. Now it’s time to interlink. The pillar links to every spoke, and every spoke links back to the pillar. The anchor texts are descriptive and varied, rather than saying ‘click here’ twenty times or using exactly the same keyword twenty times. This step is most often botched, which is why it gets its own section right away.

6. Measure and expand. A cluster is never ‘finished’. After eight to twelve weeks, Google Search Console shows which spokes are performing well and where the gaps lie: search queries for which you’re collecting impressions but don’t have a page that actually ranks for them. Each of these gaps is a spoke you’re still missing. This is how the cluster grows based on real demand rather than gut feeling.

Internal linking: the real secret

If a cluster isn’t working, it’s almost never down to the text. It’s down to the linking. This is precisely where clean architecture sets itself apart from a pretty-looking mess.

Three rules are key. Firstly: no orphans. Every spoke must have at least one incoming internal link from the pillar and one outgoing link back to it. A post that nothing links to and that links to nothing is invisible to the cluster logic, no matter how good it is. Secondly: anchor text should describe the destination, not the click. “How to transfer your product data during migration” tells Google and readers what to expect; “find out more” tells them nothing. Thirdly: link contextually, not as a list of links at the bottom of the page. A link belongs in the body text at the point where the linked topic is mentioned. A tacked-on block titled ‘You might also be interested in’ with eight links is precisely the kind of pattern that readers and search engines recognise as filler, and which fails to establish the semantic context that matters.

The effect of this linking structure is twofold. For search engines, it makes the thematic relationship explicit and directs link equity specifically to the pillar article, which thereby builds authority for the highly competitive main keyword. For people, it creates a path: anyone landing on a ‘spoke’ can find the whole topic via the ‘pillar’ and the logically next question via the neighbouring ‘spokes’. Internal linking is therefore not an SEO trick. It is the wiring that turns individual parts into a system.

An example cluster, from start to finish

To ensure this doesn’t remain abstract, here is a worked-out example. Take a B2B service provider that implements Shopware projects and wants to rank for the topic of migration.

The pillar page is titled, for example, “Shopware migration: the complete guide from SW5 to SW6”. It outlines the entire migration process (reasons, phases, risks, effort involved, outcome) and links to the relevant spoke in each section.

The spokes cover the individual search intentions:

Eight spokes, one pillar, fully interconnected. What matters in this example is not the number, but the cohesion: Every serious query that a prospective migration client Googles has its own in-depth answer within the cluster, and all answers lead back to the Pillar. Covering the subject in this way signals the same thing to a search engine month after month: this domain understands Shopware migration. That is the definition of topical authority, and it is not the result of a stroke of genius, but of consistent comprehensiveness. Anyone who cannot or does not wish to manage this in-house can bring in the architecture and implementation via a data-driven SEO and content strategy: the effort lies not in the individual text, but in the well-thought-out structure of the cluster.

Topic clusters and AI visibility (GEO)

This is the part that will transform the topic in 2026 from a purely SEO-related discipline into a matter of survival. Since the roll-out of AI answers in Google Search across the DACH region, AI Overviews since 26 March 2025, and the conversational AI Mode since 7 October 2025, a machine is increasingly intercepting the answer before a click even takes place. The question is no longer simply ‘am I ranking?’, but ‘will I be cited as a source by the AI answer?’. This targeted focus on AI citability is known as Generative Engine Optimisation (GEO).

And this is precisely where the cluster architecture pays off a second time. AI systems do not primarily evaluate individual pages, but rather how consistently a domain demonstrates expertise on a topic across many pages. This is known as entity authority. A domain that covers a topic within a cohesive cluster provides precisely the signal these systems are looking for: coherent, widely substantiated subject expertise rather than an isolated hit. Topical authority, which has been an SEO bonus for years, thus becomes a prerequisite for appearing in an AI response at all. How to technically make your own domain citable for AI crawlers is a topic for another chapter. We’ve covered this in detail in the article Does AI even find your website?.

The uncomfortable consequence: anyone who continues to rely on individual posts loses out twice over in AI search. They don’t rank, and they aren’t cited. Clusters are the answer to both problems at once.

The most common mistakes

A few pitfalls crop up so regularly that they deserve their own warning.

Orphan content. Spokes that have been written but never linked to. This almost always happens when linking is left as the final step to be done ‘later’ and then gets overlooked. Treat linking as part of the publishing process, not as an optional afterthought.

Keyword cannibalisation. Two or three spokes that cater to the same search intent because they were structured around keywords rather than intent. They compete internally and weaken one another. If in doubt, merge them rather than split them.

The thin pillar. A pillar page that, whilst having a perfect heading structure, contains only two sentences under each heading. This is a skeleton without substance, and both readers and search engines recognise it immediately. A pillar that lives up to its claim as a central hub needs real depth in every section.

No refresh. A cluster that is built once and then forgotten becomes outdated. Particularly in technical B2B topics, versions, prices and best practices change. A good cluster is checked for up-to-date information at least once a year, and the pillar is linked to every new spoke.

The ‘boil the ocean’ cluster. Attempting to build five clusters in parallel usually ends up with five half-finished ones. A complete, excellently interconnected cluster is miles ahead of three incomplete ones.

How to get started

The good news for anyone with a full blog archive: you’re not starting from scratch. The first step is to take stock. Export your existing posts and sort them by topic. In most B2B blogs, there are already three or four half-formed clusters scattered about – enough material for a pillar and a handful of spokes, which were simply never conceived or linked as a unified whole.

So the rule is: one cluster, not ten. Choose the core topic with the highest business relevance, develop it fully, link it up properly and measure the results over a quarter. A single, cohesive cluster that ranks for its main topic and is cited in AI responses is worth more than an archive of a hundred individual posts that nobody can find. Only once the first cluster is bearing fruit should you move on to the second.

By 2026, content strategy for B2B will no longer be a question of volume, but of architecture. Those who think in terms of topics rather than texts and take the interlinking seriously will build something that grows stronger over time, rather than gathering dust in the archive. Anyone can write more posts. Few can truly own a topic.