Cross-Cluster Fan-Out Analysis: A New Way to Plan Content for Brands With More Than One Story to Tell
Most SEO and GEO work happens one cluster at a time. You pick a seed topic, run a fan-out, and end up with a clean map of supporting pages. That works fine when a brand has one main thing to talk about.
It breaks down when a brand has more than one.
The moment you run a second fan-out for the same client, you start noticing things. Pages that should probably link to each other. Buyer journeys that overlap but were planned separately. Entity decisions made in cluster A that quietly conflict with cluster B. The fan-outs are not wrong on their own. They just don’t see each other.
Cross-Cluster Fan-Out Analysis, or CCFA, is the method I built to fix that.
What CCFA actually is
CCFA is a horizontal comparison of two or more fan-outs for the same brand. Instead of going deeper inside a single cluster, it looks across clusters and asks where they converge, where they diverge, and where the architecture has to make a real choice.
A single fan-out answers the question: what does the content around this seed look like?
CCFA answers a different one: what does our brand’s content architecture look like when we put the clusters next to each other?
That second question is the one most teams never ask. And it’s the one that determines whether your content compounds or competes with itself.
Why a single fan-out is not enough
A fan-out is a hypothesis. It tells you what a search engine or LLM expects to see around a topic, based on the signals available at the keyword surface.
But the keyword surface is missing things.
It does not show you the buyer’s actual context. It does not show you which entities are doing the heavy lifting in your category. It does not show you that two seed topics, which look unrelated in the SERP, share 60% of their supporting concepts under the hood.
If you build off a single fan-out, you get a plan that looks complete. Run a second fan-out for a related seed, and the gaps start to show. Run a third, and the architecture has already drifted.
CCFA catches the drift early. Before the briefs go out. Before the writers start. Before the internal linking gets locked in.
What the comparison surfaces
When you put two or more fan-outs side by side, four things tend to jump out.
Convergence. Topics that show up in more than one cluster. These are the places where a single hub page might serve multiple clusters, instead of writing three near-duplicate pages. Convergence points are also where you usually find your strongest internal linking opportunities.
Divergence. Topics that look related at the keyword level but are clearly serving different buyer journeys underneath. These are the places where you keep things separate, even when an editor would push to combine them.
Entity overlap. The named things, like products, methods, people, and categories, that show up across clusters. Entity overlap tells you where your knowledge graph is consistent and where it is contradicting itself.
Distribution gaps. Sources, formats, or platforms that one cluster relies on heavily and another ignores. If cluster A is built around third-party citations and cluster B is built around in-house content, that asymmetry has consequences when an AI tool decides who to quote.
None of these show up in a single fan-out. All of them matter.
Where this fits in the GEO conversation
The wider GEO conversation right now is mostly vertical. Most of the good work, like verifying entities, building citable blocks, and structuring claims for extraction, happens inside a single page or single cluster. That work is necessary, and I do it on every project.
CCFA sits one layer above it. It is horizontal. It looks at the relationships between clusters rather than inside them.
The reason that matters is simple: AI tools do not read your site one cluster at a time. They build a picture of your brand from everything they have seen. If your clusters were planned in isolation, that picture is going to be inconsistent, and inconsistency is one of the signals that pushes a brand out of the citation set.
Vertical work makes a page worth citing. Horizontal work makes a brand worth citing across queries.
When to run a CCFA
CCFA is not a default step. It is a specific tool for a specific situation. The signals that tell you it’s time:
- Your brand has two or more meaningful content pillars that have grown separately.
- You’re about to make architectural decisions, like site structure, internal linking, or entity schema, that will be hard to reverse.
- Your fan-outs are producing good individual plans but the team is debating which one is the “main” cluster.
- You’ve noticed pages from different clusters competing in the same SERP or showing up in the same AI Overview.
If any of those apply, the cost of running a CCFA is small compared to the cost of locking in the wrong architecture.
What it doesn’t do
A few things to be clear about, since this is the kind of method that can get oversold.
CCFA does not replace the per-cluster fan-out. It assumes you have done that work. It is a layer on top, not a substitute.
It does not produce a content plan by itself. It produces the architectural decisions that a content plan should be built around.
And it does not work without real data. The inputs are full fan-outs, with verified search intent, SERP analysis, and entity work already done. Skipping those upstream steps makes the comparison meaningless.
The short version
Single fan-outs answer questions inside a topic. CCFA answers questions between topics. For most brands, the second set of questions is where the long-term content architecture actually lives, and where the biggest mistakes get made when no one is looking at the whole picture.
That’s the method, and that’s why I built it.
If you’re working through a multi-cluster build and want to compare notes, reach out here.