A formal representation of the concepts, categories, and relationships within a specific domain — defining how things are classified and how they relate to each other. In information architecture, an ontology provides the conceptual foundation for a taxonomy, describing not just what content categories exist but why, what they include and exclude, and how they are semantically connected.
Common contexts
- Defining whether 'article', 'guide', and 'tutorial' are three content types or three labels for the same type before building a CMS
- Establishing what 'product' means across teams — physical SKU, digital license, and subscription bundle — before redesigning navigation
- Resolving whether 'client' and 'customer' refer to the same entity in a B2B platform's data model and interface vocabulary
Use when
Invest in an ontology when multiple teams or systems use the same words to mean different things — particularly before designing a large navigation system, a search experience, or a shared component library where naming inconsistencies will compound.
Avoid when
Skip formal ontology work for small, contained products with a single content type — the overhead of documenting conceptual relationships is wasted when a flat list of pages describes the site in full.
Ontology disagreements almost always surface late as 'content strategy' arguments — but they're really unresolved business definition conflicts that no amount of IA work can fix until someone decides what the domain terms actually mean.
Real-world examples
- Wikipedia's category system is a user-facing ontology: articles belong to multiple overlapping categories (Biology > Mammals > Primates > Humans), mirroring how knowledge actually organises rather than a rigid hierarchy.
- Airbnb's property type ontology (Entire home, Private room, Shared room, Hotel room) was redesigned in 2022 into 56 named categories (Treehouses, Boats, Castles), shifting from accommodation type to experience type.
- Google's Knowledge Graph is a massive ontology linking 500 billion entities; its structure directly determines which information appears in featured snippets and knowledge panels for any given search query.