Ontology

An ontology describes how data fits together: the entities you care about (Customer, Order, Product), the attributes they carry (status, price, region), and the relationships between them (Customer places Order, Order contains Product). It also binds business metrics to that model, including names, units, aggregation rules, and time grain.

This structure reduces ambiguity. “Revenue” means the same number everywhere because the ontology fixes its filters, time window, currency, and rollup logic. Dimensions and hierarchies add context, like Region → Country → State, or Channel → Campaign

For AI, an ontology is the map. Natural language systems use the terms, synonyms, and relationships to understand intent, select the right data, and explain answers in business language.

Pro Tip

Pick one owner for each core term. When ownership is clear, naming, definitions, and change control stay tight and trust grows.

Why Ontology Matters

• Consistent meaning: Shared definitions stop duelling dashboards and rework.
• Faster answers: Teams navigate known entities, dimensions, and metrics instead of reverse‑engineering reports.
• Strong governance: Ownership, certification, and lineage keep changes visible and auditable.
• AI readiness: Clean semantics help assistants map questions to the right data and show their work.

Ontology – In Practice

Picture a core “Revenue” metric:

• Definition: Sum of “Order Amount” for completed orders, excluding refunds.
• Constraints: Currency = USD, Time grain = day, Time zone = company default.
• Dimensions: Region, Channel, Customer Segment, Product, Campaign.
• Hierarchies: Region → Country → State; Product Category → Subcategory → SKU.
• Rules: Late-arriving data flagged; fiscal calendar starts on Feb 1.

With this set, any chart, dashboard, or AI assistant draws the same value and can slice it the same way.

Ontology and PowerMetrics

PowerMetrics helps you turn ontology ideas into working analytics:

• Metric catalog: Create and describe metrics once with names, units, default aggregation, and time grain.
• Dimensions and filters: Add dimensions and default filters so slices behave the same in every view.
• Certification and tags: Mark trusted metrics, add owners, and signal which ones are ready for broad use.
• Access control: Use roles and groups to share the right definitions with the right teams.
• Modelling: Connect services, spreadsheets, or warehouses, then model tables and joins to match your business entities.
• Lineage and clarity: Keep descriptions, assumptions, and calculation notes close to the metric so changes stay transparent.
• PMQL and natural language: Use precise metric names and fields for repeatable queries; assistants can tap those definitions to interpret questions.

A lightweight ontology inside PowerMetrics builds the semantic backbone for consistent dashboards today and AI assistance tomorrow.