Ontologyorgim

What are the layers of an organizational ontology?

An organizational ontology has four layers: entities, types, relationships and facts. Entities are the things the company is made of: the company itself, each brand, each product line, each office or location, and each person who represents the company in public. Types say what kind of thing each entity is, taken from a shared vocabulary such as schema.org (826 types and 1,540 properties in version 30.1, 16 September 2026): Organization, LocalBusiness, Brand, Product, Person. Relationships connect the entities with schema.org properties: worksFor links an employee to the company, parentOrganization links a subsidiary to its parent, brand links an organization to its brands, location links it to its places. Facts are the attributes: names, addresses, phone numbers, founding dates, prices. In JSON-LD, each entity gets one stable identifier with @id, which the JSON-LD 1.1 Recommendation (W3C, 16 July 2020) describes as the keyword "used to uniquely identify node objects". Map the layers in order: entities first, then types, then relationships, then facts, and publish them together as one connected graph.

What are the four layers of an organizational ontology?

The four layers of an organizational ontology are entities, types, relationships and facts. Entities are the things, types are their categories, relationships are how they connect, and facts are their attributes. Each layer is mapped separately and the four are published together as one structured-data graph (publishing).

What belongs in the entity layer?

The entity layer of an organizational ontology lists the things the company is made of: the company itself, each brand, each product line, each office, branch or location, and each person who represents the company in public, such as founders, executives and named authors. Each entity gets one canonical name and one canonical page that defines it.

Where should the types come from?

The types in an organizational ontology should come from a shared vocabulary, normally schema.org, which has 826 types in version 30.1 (16 September 2026, checked 9 November 2026). Types invented in-house can't be read by search engines or other software that only knows the shared vocabulary. The common organizational types are Organization, LocalBusiness, Brand, Product, Service and Person.

Which schema.org type should a single shop or branch use?

A single shop, office or branch that customers visit should use LocalBusiness, or its most specific subtype. Google's Organization structured data guide (checked 9 November 2026) recommends "using the most specific schema.org subtype of Organization that matches your organization", and says local businesses should use the most specific LocalBusiness subtypes. The company as a whole stays an Organization (or a more specific subtype).

Which properties make up the relationship layer?

The relationship layer of an organizational ontology uses schema.org properties that point from one entity to another. Each relationship should read as a sentence you would say about the company:

All of these are defined on schema.org (version 30.1, checked 9 November 2026).

How do I give each entity a stable identifier?

Give each entity one stable identifier with the JSON-LD @id keyword, usually a URL on your own site with a fragment, such as https://example.com/#organization. The JSON-LD 1.1 Recommendation (W3C, 16 July 2020) defines @id as the keyword "used to uniquely identify node objects" with IRIs. Use the same @id for the same entity on every page and every site you run, so relationships on one page point to the entity defined on another.

What belongs in the facts layer?

The facts layer of an organizational ontology holds the attributes of each entity: names and alternate names, legal name, addresses, phone numbers, email, founding dates, prices and links to official profiles (sameAs). Google's Organization structured data guide (checked 9 November 2026) lists name, url, logo, sameAs, address, telephone, foundingDate, legalName and alternateName among its recommended properties, and says no property is required.

Which layer changes most often?

Facts are the layer that changes most often: prices, phone numbers, addresses and opening hours change with ordinary business, while the entities and the relationships between them (who owns which brand, which branch belongs to which company) change only with events such as a merger, a rebrand or a new location. Map the stable layers first so that changing a fact never means rebuilding the graph.

In what order should I map the layers?

Map entities first, then types, then relationships, then facts. Mapping projects that start with facts, the easiest layer, often never finish the relationships, and without relationships the facts are a list of attributes with no structure. The mapping guide walks through each step.

Sources

Next step

Map your own company: follow the mapping guide.