Hub-and-Spoke: The Site Structure AI Can Actually Understand

We recently helped an IT consulting firm move to a completely new website. They brought eight years of consistent publishing with them: an impressive archive of blog posts and resources, and no organizing structure whatsoever. Every post was a leaf with no branch.

Rather than migrate the pile as-is, we used the move to give the archive a skeleton. We codified eight years of content into eight or nine categories, built themes around each one, and arranged the site so each category had a central page with its posts gathered around it. Then we went one step further: within each theme, we organized content by the problems clients face and the capabilities that solve them, so the structure told a story, not just a taxonomy. Without ever using the term, we had built a textbook hub-and-spoke architecture.

What hub-and-spoke means

The model is simple enough to sketch on a napkin.

• One hub per money topic. A hub (often called a pillar page) is the definitive page for a subject you want to own: broad, well-organized, linking out to everything you have on that subject.
• Spokes answer single questions. Each spoke is a focused post that answers one question inside the hub's territory. Not three questions. One.
• Every spoke links to its hub. The link says, in machine-readable terms, "this specific answer belongs to that body of expertise."

Keep the plumbing boring and consistent. URLs should reflect the hierarchy (/it-security/ as the hub, /it-security/zero-trust-basics/ as a spoke). Headings should do the same: the page's question in the H1, the sub-answers under descriptive H2s. Nothing clever, everything predictable.

Why this is the structure machines can read

Hub-and-spoke is the buildable version of the topical authority we keep telling you machines reward. Authority is breadth, depth, and connectedness; a hub gives the breadth, spokes give the depth, and the links give the connectedness. When an AI system evaluates whether your site knows a subject, a hub-and-spoke cluster hands it the answer pre-organized: here is the subject, here are the twelve questions inside it we answer, here is how they relate.

A flat archive forces the machine to reconstruct those relationships itself, and machines mostly do not bother. Eight years of expertise, published flat, reads like eight years of miscellany.

What changed for the IT firm

Three things, in order of how quickly they showed up.

First, content decisions became easy. The structure exposed its own gaps: a category with a strong hub and two thin spokes is an assignment sheet, not a mystery. The perpetual "what should we write next" meeting shrank to minutes.

Second, sales conversations improved. Prospects arriving through one spoke found their way to the hub and read laterally. The site started doing the qualifying that the first sales call used to do.

Third, search performance compounded. Individual posts that had ranked modestly for years began performing better once they were connected to a hub, because each spoke now inherited context from the cluster around it.

Where the gaps hide

Here is the catch with hub-and-spoke: it is easy to draw and hard to audit. Once you have nine hubs and a hundred spokes, nobody can hold the whole structure in their head. Which hubs are missing spokes? Which spokes attached themselves to the wrong hub? Which subject has no hub at all?

That audit is exactly what a topic map is for. Topic Modeler reads your site and renders the actual structure, hubs, spokes, strays, and holes, so you can see where the architecture is solid and where it only exists in the original plan. Build the wheel, then look at it the way the machines will.