Website Information Architecture

Information architecture and its representations

In this guide, information architecture means the organization and labeling of information and the relationships through which people encounter it. A conceptual structure can be represented in a sitemap, category index, navigation menu, or collection of related links. These representations overlap, but each reveals only selected relationships. A menu is therefore evidence of an architecture, not necessarily a complete inventory of it.

There is also structure inside individual pages. Headings identify sections, while regions distinguish navigation, main content, and other page areas. W3C explains that meaningful page structure supports orientation and navigation, including navigation through assistive technologies. Page-level structure and site-level organization operate at different scales: an article can have coherent headings while its relationship to other articles remains unspecified. W3C WAI — Page Structure Tutorial

Hierarchy and parent–child relationships

A hierarchy arranges items into broader and narrower levels. In a strict tree, each non-root item has one immediate parent. Website classification need not remain a strict tree: one subject can relate to more than one broader subject. SKOS explicitly permits concepts to have multiple broader concepts, so its semantic relationships can express structures that are more complex than a single folder tree. SKOS Reference

In a hypothetical reference library, “Materials” could contain “Packaging,” which contains “Paper.” This is a conceptual example, not a report about an existing website. A URL such as /resources/paper does not, by itself, express all those relationships. Conversely, nested URL segments do not establish a formal taxonomy. Location, classification, and the routes shown in an interface are separate descriptions of the same content.

Navigation as an interface to content

Navigation provides routes between destinations or within a document. A global menu can expose broad areas, a local menu can expose neighboring pages, and contextual links can connect related explanations. An in-page contents list links to section identifiers rather than separate pages. These are different navigation mechanisms, even where they use the same visual treatment or repeat the same labels.

W3C’s explanation of WCAG 2.2 Success Criterion 2.4.5 describes multiple ways of locating pages within a set, with an exception for pages that are a step in or result of a process. Its examples include search, sitemaps, and tables of contents. This is a Level AA accessibility criterion, not a rule that every menu must display every page. Multiple Ways

Taxonomy, concepts, and labels

Taxonomy concerns classification: which concepts exist, how they are named, and how they relate. Terminology varies between organizations, so “taxonomy” does not always imply the same implementation. SKOS is a W3C Recommendation dated 18 August 2009 for representing knowledge organization systems, including thesauri and classification schemes. It supplies a precise vocabulary without requiring websites to publish their classifications in that format. SKOS Reference

SKOS distinguishes a concept from its labels. A concept can have a preferred label, alternative labels, and hidden labels; language tags identify the language of lexical labels. Its broader, narrower, and related properties distinguish hierarchical from associative relationships. Thus “related to” does not necessarily mean “contained within.” This distinction separates subject association from parentage and permits different labels to refer to the same underlying concept. SKOS vocabulary

A further SKOS distinction concerns immediate and transitive relationships. By SKOS convention, the broader property records a direct hierarchical assertion, while broaderTransitive supports reasoning across a chain of broader relationships. An interface that shows only the immediate parent therefore presents different information from an expanded ancestry view. This distinction describes the classification model; it does not require a website to expose every inferred ancestor in its navigation. SKOS Reference

Categories, tags, and facets

A category is a grouping used to classify content. A category landing page is a presentation of that grouping, not the grouping itself. If an article is assigned to two categories, the assignments describe two memberships; they do not necessarily create two separate article records. The storage arrangement is a CMS question, while the meaning of membership is a classification question.

Terms such as “category” and “tag” are implementation-dependent rather than universal technical opposites. In a hypothetical catalog, subject categories could coexist with tags for material and format. A faceted interface could expose those dimensions independently, allowing “paper” and “poster” to describe different attributes of the same item. SKOS collections group concepts, whereas broader and narrower properties express semantic relationships; grouping alone does not establish hierarchy. SKOS Reference

Breadcrumbs and contextual paths

W3C’s ARIA Authoring Practices describes a breadcrumb trail as a list of links to parent pages in hierarchical order. It gives a navigation landmark with a distinguishing label as the pattern’s semantic container. When the current page is represented by a link, aria-current="page" identifies it; when the current page is not a link, that attribute is optional. Breadcrumb Pattern

A breadcrumb therefore communicates a location or contextual ancestry rather than necessarily reproducing the visitor’s click history. In the hypothetical library, “Resources → Materials → Paper” could describe a hierarchy even if the visitor arrived directly from another website. A browser’s Back command answers a different question: which entry preceded the current one in that browsing context. The two mechanisms can lead to different destinations.

Page structure and limits of representation

Heading levels describe relationships between sections of a document. W3C identifies ranks from h1 to h6 and explains that lower-ranked headings introduce subsections. A heading’s semantic rank is not the same as its font size. Likewise, a visual cluster of links does not explain its purpose to every user unless its structure and labeling convey that relationship. W3C WAI — Headings

No single representation resolves every organizational question. A heading outline describes one page; a taxonomy describes concepts; breadcrumbs expose one contextual path; and navigation offers selected routes. The Crawling, Rendering, and Indexing reference separately explains how search systems retrieve and process pages. The related website service page describes a commercial offering, not the definitions or accessibility requirements presented here.

Frequently asked questions