Taxonomy for content designers: what to know before tagging a single page

Taxonomy on Australian government websites usually gets handed to information architects or records managers. That is a category error.

Taxonomy is a language decision: what to call a thing so users can find it. Language decisions belong to content designers.

Taxonomy decisions are linguistic. Records teams had built tags around how the agency organised its work. Users searched in their own words. The two need a content designer to bring them together and write the term list.

This is the part of the website nobody puts on a content designer's brief. It is also one of the most consequential things on the page.

What taxonomy actually is, and what it is not

Three terms get used interchangeably. They should not be.

Taxonomy is the controlled vocabulary your site uses for search, filtering, and browsing. It is the list of approved category names: "Eligibility," "Apply," "Service finder."

Ontology is the relationship between those concepts. It tells the system that "Apply" connects to "Eligibility," and that "Concession card" is a type of "Identity document."

Content model is the structure of each content type. The fields a service page must contain, the rules for what is required, how content blocks relate.

The three are connected. They are not the same. Taxonomy is the layer closest to the user, and the layer most often built without a content designer in the room.

Why content designers should own it

Taxonomies fail in government when they use policy language instead of user language. "Community Services Access Program" rather than "support payments." "Subordinate legislation" rather than "regulations." "Notifiable instruments" rather than "rules that apply to you."

That is not a technical error. It is a content design error. It is the same error that produces unreadable web pages, and it gets fixed with the same instinct: use the words your reader uses.

The Australian Government Style Manual is direct on this point. Use the words your audience uses, not the words your organisation uses. The same rule applies to category names as to body copy. You would not let a policy officer write your H1. Do not let them write your taxonomy.

Australia has a reference. Most teams have not heard of it.

The Australian Governments' Interactive Functions Thesaurus (AGIFT), maintained by the National Archives of Australia, is the closest thing the APS has to a national taxonomy. It maps the functions of government (health, education, defence, social services) into a controlled vocabulary used for records management and discoverability across agencies.

AGIFT is a starting point. It is not a content design taxonomy. AGIFT is built around what government does. A content design taxonomy is built around what users come to do. Those are not the same thing, and confusing them is the most common reason a government taxonomy reads like an org chart.

Use AGIFT where your records and reporting obligations require it. Build a separate, user-facing taxonomy for the website. They are different jobs.

Before-and-after: a worked example

Think about the language you’ve seen on pages before. It probably looked a bit like this, written by the program team:

  • Threshold requirements and exclusion provisions

  • Assessment pathway and administrative arrangements

  • Mandatory attachments, declarations and certifications

Nothing on that list would match a search term anyone outside the directorate or department would type. After a content design pass, the same content sits under:

  • Who can apply

  • How to apply

  • What you need to provide

Same content. Different doors. The second set is what taxonomy is for.

Where to start

If you are a content designer who has never been asked to think about taxonomy, three moves to make this week:

  1. Find out who currently owns the controlled vocabulary on your website. It is usually somebody in records, information architecture, or CMS administration. Get a copy of the term list.

  2. Compare it against the top 50 search terms in your site analytics. Note every gap between what users type and what the system tags.

  3. Run a five-minute test with three people outside your team. Read them the top tags and ask what they think each one means. The gaps will not be subtle.

That is enough to make the case that taxonomy is a content design problem, not a records one.

The question to ask

Who in your agency decides what users are allowed to search by? If the answer is not a content designer, the words on your category pages are not under content design control. That is the gap to close.

Photo by Clay Banks on Unsplash.

Previous
Previous

AI-Era Content Strategy: How search and AI consume government information

Next
Next

Responsible AI use in government content isn't a policy question. It's a workflow question.