Metric Layer

Metric Layer

The metric layer (also called a semantic layer or metrics store) is a centralized system that defines and manages all metrics, dimensions, and time grains for an organization. Think of it as a single source of truth for business metrics. Instead of each dashboard or tool re-implementing its own metric calculations, the metric layer provides a governed API or interface with canonical definitions. In practical terms, a metric layer includes:

Key components:

• A metric registry with canonical names, descriptions, formulas, and owners for each metric.

• Definitions of dimensions (categories like product, region) and default time grains (daily, weekly, monthly, etc.) associated with each metric.

• Business logic like default filters, aggregation methods, and data formats behind each metric.

• An API or SQL interface so that BI tools, notebooks, and applications can query these metrics consistently.

By centralizing this logic, a metric layer ensures consistency. As one expert writes, “A metrics layer (aka semantic model) centralizes business logic — names, formulas, grains, filters, and dimensions — behind a governed API.” For example, modern data platforms (like dbt’s semantic layer or headless BI tools) require you to define the metric formula plus its time grain and dimensions for reuse. This way, everyone uses the exact same definition of “monthly active users” or “total revenue” no matter which tool they query.