What is a topical map and how to build one from data

A topical map is the complete set of topics a site must cover to be the authority on its subject, organised as a tree. At the top are pillars, the three to six broad subjects the business is actually in. Under each pillar are clusters, the groups of related questions that a reader would expect a genuine authority to cover together. Under each cluster are the individual topics, each one a real query with a real search volume and a real intent, each one mapped to one page. Around the tree run the internal links: every article links up to its cluster and its pillar, sideways to its siblings, and out to the service page that the cluster exists to support.

That last part is what separates a topical map from a content calendar. A calendar is a list of things to publish. A map is an argument about what the site is for, and every page on it has a job: to answer one question, in one intent, in one cluster, and to send the reader one step closer to the service. About 30 Australians a month search for "topical map" and the term sits at difficulty 53, which tells you it is still a specialist idea. The businesses that use it are the ones whose sites read like a reference work rather than a scrapbook.

On 23 May 2023 Google published a Search Central post describing a system it calls topic authority, used to surface expert sources for queries in specialised areas. The post is framed around news, but the signals it lists are general: how notable a source is for a topic or location, how much original work the source produces on that topic and how often others cite it, and the source's history of quality on the subject. Google's Search Liaison later clarified the system had been in use for years rather than newly launched. Whichever way you read it, the message is that depth on a subject is measured and rewarded, and depth is exactly what a topical map plans.

The practical effect shows up in our own data. The physiotherapy guide we published in July 2026 reached an average position of 11 for "seo for physios australia" within weeks, and the strongest explanation we have is that it did not arrive alone: it landed in a cluster that already held dentists, lawyers, accountants, tradies, mortgage brokers and financial advisers, all cross-linked, all pointing at the same service. A single article on a site with no cluster around it would have had to earn that trust from nothing. The AI engines behave the same way, and their citations skew toward sources that cover a subject consistently rather than once; our guide to topical authority in content strategy covers that side of it.

Here is our map as it stood when this guide was last updated in July 2026, drawn from the taxonomy file that generates the hub pages. Three pillars, fourteen clusters, 103 articles. The bars show where the depth is and, more usefully, where it is not.

Two honest readings. The AI search cluster is the deepest on the site at sixteen articles, which is deliberate: it is the category we intend to own, and the depth is what earns the citations. The Local SEO cluster has two, and the Conversion cluster has four on a site that sells conversion optimisation, which is the map telling us what to build next and in what order. That is the point of drawing it. A map that only confirms what you have already written is a diagram; a map that embarrasses you into the next ten articles is a plan.

The expansion step is where the platform earns its place. One seed term returns its full cluster with volumes and intents in seconds; the GP seed in our journey mapping guide returned 25 variants in a single lookup. Done by hand from public tools it takes a week per pillar and is stale by the time it is finished. Our keyword research guide covers the lookups themselves and our search intent guide covers the label each node carries.

A topical map built with intents on every node is also, without any extra work, the plan for which pages are open and which sit behind a login. Every node whose intent is pre-purchase (a question, a comparison, a service, a place) is an open, indexable page in the tree, because it exists to win a reader who has not chosen you yet. Every node whose intent is post-purchase (how to use it, my account, the members' material, the deeper version of a guide) is a candidate for the gated layer, where it keeps the customer and builds first party data. The tree does not change; the login line runs through it.

This is why we build the map before the site rather than after. It decides the navigation, the URL structure, the breadcrumb trail, the service links, the content plan and the portal boundary in one document, and it does so from what customers search rather than from an org chart. Our client portal guide takes the map and turns it into the two-layer site it implies.

Built by hand from public tools, a three-pillar map is two to three weeks of strategist time and needs rebuilding every quarter. Our Snowball SEO platform holds twenty topical maps for this site alone and regenerates the expansions on demand, which is the heavy lifting other agencies bill by the hour, and it is why a map is included, not sold, at the start of every content strategy and keyword and topic research engagement. The budget the platform saves goes to the articles, the photography and the video that fill the tree.