Originally published: April 18, 2022 · Last updated: August 16, 2026
Information architecture is the structure behind a website: how content is grouped, labeled, connected and exposed through navigation. When it works, visitors can predict where information lives. When it fails, even strong content becomes difficult to find.
The CMS should implement that structure, not invent it by accident.
Start with user tasks, not menu labels
Before drawing a sitemap, list what visitors are trying to accomplish. They may want to compare services, find an artist, read a guide, check a price, contact someone or use a tool.
Those tasks help reveal which sections deserve prominence and which content belongs together. Internal organizational charts are rarely a good substitute for user-facing structure.
Inventory the content you already have
For an existing site, collect the current pages, posts, custom post types, categories and important assets. Group them by purpose and topic.
This exposes three common problems:
- several pages doing the same job;
- important subjects with no clear home;
- content types that were built as one-off pages but should share a repeatable structure.
A redesign becomes much easier when you know what content actually exists before creating a new menu.
Choose categories people can predict
Good labels use language the audience understands. Avoid internal department names, clever terminology and categories so broad that almost anything could fit inside them.
If users repeatedly hesitate between two categories, the structure may be reflecting your organization rather than their mental model.
Keep the hierarchy as shallow as the content allows
Not every page needs to be one click from the homepage, but deep hierarchies create maintenance and navigation problems when each level exists only to hold another level.
Use hierarchy when it communicates a real relationship. A section for WordPress can contain maintenance, security and performance resources because those topics belong together. Creating several empty intermediate layers adds little value.
Navigation and information architecture are related but not identical
The information architecture describes the content relationships. Navigation is one interface for moving through them.
A site can expose the same architecture through the main menu, contextual internal links, topic hubs, breadcrumbs, filters and search. The primary menu does not need to contain every page.
Use headings and regions to help orientation
Structure exists inside pages too. Clear headings let readers scan sections and help assistive technologies navigate content. W3C guidance also recommends meaningful page regions and descriptive labels when several regions serve similar functions.
Good information architecture therefore supports accessibility as well as visual navigation.
Design taxonomies deliberately
WordPress makes categories and tags easy to create, which is exactly why they can become messy. A taxonomy should help users browse a meaningful collection, not simply record every word associated with a post.
Before adding a new category or tag, ask whether it will contain enough useful material to justify an archive and whether visitors will understand the distinction from existing terms.
Use internal links to connect related intent
Hierarchy alone cannot represent every useful relationship. Contextual internal links allow one article to point to a supporting guide, comparison or tool without forcing both into the same menu branch.
Link when the destination answers a likely next question. This creates a network of useful pathways instead of a rigid tree.
Card sorting can expose labeling problems
For larger sites, card sorting is a simple research method: give representative users a set of content topics and ask them to group or label them. The goal is not to let a handful of participants design the site, but to discover where your proposed categories conflict with how people naturally think about the material.
Prototype the structure before designing pages
A simple sitemap, spreadsheet or low-fidelity wireframe is enough to test whether the main content has a logical place. You do not need finished visual design to discover that a section is overloaded or that two navigation labels mean almost the same thing.
Plan migrations around the architecture
During a redesign, map every important old URL to its future state: retained, rewritten, merged, redirected or removed. The new information architecture should guide those decisions.
Do not force every legacy page into the new structure merely to avoid making editorial choices. A redesign is an opportunity to reduce duplication and clarify purpose.
Test the paths that matter
After building the structure, test realistic journeys rather than asking whether the menu looks neat. Can a new visitor find pricing? Can someone move from a broad guide to a specific tutorial? Can a mobile user locate the same destination without guessing?
Search logs, analytics, usability sessions and support questions can reveal where the architecture is failing.
A practical IA workflow
- List the site’s primary audiences and tasks.
- Inventory existing content.
- Group content by purpose and relationship.
- Define a small set of clear top-level sections.
- Choose labels using audience language.
- Plan categories, tags and custom content types.
- Prototype navigation and key user paths.
- Add contextual internal links between related resources.
- Test the structure on mobile and with real users where possible.
- Review the architecture as the content library grows.
The practical rule
Information architecture is not about producing the perfect sitemap. It is about making the content model predictable enough that users, editors and search systems can understand how the site fits together. Start from user tasks, keep labels clear and let the structure evolve deliberately instead of accumulating by accident.
Official reference: W3C WAI page structure guidance.