What is a product ontology?
A product ontology is an operating model of a physical-product business. Where a database stores rows and a dashboard displays them, an ontology models the business as a system: the objects (product, unit, collection, supplier, country, order, passport), the relationships between them (a product belongs to a collection, is made by a supplier, sits in a country), and the actions you can take on them (reclassify, complete, reorder, correct). It captures not just what data you have, but what it means, what you can do with it, and who is allowed to.
Why a table is not an ontology
Ask a spreadsheet "what happens if this supplier fails?" and it has no answer, because a table does not know the supplier is connected to forty products, in three collections, generating a share of your revenue. An ontology does, because the relationships are part of the model. Modelling the links, not just the rows, is what lets a system reason about consequences instead of only reporting numbers.
The four layers of a product ontology
- Objects: the nouns of the business and their properties (a product, its composition, its origin, its passport readiness).
- Relationships: the links that let consequences propagate (product to collection to supplier to country to sales to risk).
- Actions: the verbs you can execute, governed with confirmation and an audit trail.
- Governance: who can do what, isolated per organisation, fully traceable.
Down to the unit
For physical products the ontology goes one level deeper than most, to the unit. Two garments of the same style but different size and colour are distinct objects, each with its own identity. An ontology that models the unit, not just the style, is what makes a genuine digital twin possible: an object you can serialise, trace and give an identity that follows the item from factory to resale.
The Digital Product Passport is the wedge
Regulation is turning parts of this ontology from optional to mandatory. The EU Digital Product Passport requires structured, verifiable information attached to a product identity: composition, origin, care, compliance. A brand that already models its products as an ontology has the passport almost for free; a brand that stores products as flat rows has to build the model under deadline. The passport is the forcing function that makes the ontology worth building now.
Why a product ontology matters
When AI operates over an ontology instead of a flat table, three things become possible that a dashboard or chatbot cannot do: it can reason about scenarios (propagate "what if this supplier fails" through the links), it can act through the same governed flows a person uses, and it compounds as every decision makes the model of the business sharper. The ontology, not the AI, is the durable advantage.