Ontology

An ontology is a formal specification of the concepts in a subject area and the relationships between them: what types of things exist, what properties they can have and how they may connect. It is the blueprint; a knowledge graph is that blueprint filled with real data.

How it works

An ontology defines classes (Person, Organisation, Place), properties (birthDate, foundingDate), relationships (born_in, employed_by) and rules about them, such as which class a relationship can point to, or that every Course is a kind of CreativeWork. As MLforSEO’s knowledge graph guide puts it, the ontology says a Person can have a birthDate (a Date), be born_in a Place and be employed_by an Organisation; the graph instantiates it with Steve Jobs, born 1955 in San Francisco, employed by Apple.

On the web, formal ontologies are written in languages such as RDF Schema and OWL, while schema.org provides a widely shared vocabulary of types and properties that structured data uses. Ontologies are what make semantic consistency and automated reasoning possible: if a system knows the rules, it can infer facts that were never stated.

How it’s used in SEO and marketing

• Designing your site’s entity model before writing schema markup: which entities you have, which properties matter and how they link.
• Planning taxonomies, categories and internal linking rules that mirror real relationships.
• Structuring semantic keyword research with the entity-attribute-variable model.
• Defining your brand’s “ontological core”: the essential concepts and definitions that should never change across channels.

Example

A training provider might define: a Course has a provider (Organisation), an instructor (Person), a level and a price, and teaches one or more topics. Each Person worksFor the Organisation and has sameAs links to external profiles. Writing this down first means every course page, instructor bio and schema block uses the same types and relationships, which is the consistency AI systems reward when they reconcile your entities.

Limitations and things to watch

• It is easy to over-engineer; model only what you will maintain and use.
• Your ontology is not Google’s. Map your types to schema.org and to public identifiers where you can.
• Schema.org is deliberately flexible and does not enforce your rules, so validation tools will not catch logical inconsistencies.
• Ontologies need governance as products, people and offers change.

Learn it on MLforSEO

• What Is a Knowledge Graph? How to Use It for Smarter Keyword Strategy
• How Entity Blueprints Turn Brand Knowledge into Search Signals
• Why Entity Relationship Mapping Is Critical for Brand Visibility
• Entity Relationship Planning Workbook
• AI Search & LLMs: Entity SEO and Knowledge Graph Strategies for Brands (course)