Metrics Layer for Analytics Engineers
A metrics layer, also called a semantic layer, defines business metrics such as revenue once, in code, so every tool calculates them the same way. It declares measures, dimensions and filters, and good metric design avoids ambiguous names, mixed grains and duplicated definitions.
Stage: Metrics and Semantic Layers · Part of the Analytics Engineer roadmap · Last reviewed 19 November 2026
What to learn in Metrics Layer
- Define a measure and its allowed dimensions in a semantic model
- Create a ratio metric from two measures
- Write a clear, unambiguous metric name and description
- Check a metric against a hand-calculated value
- Explain what happens when two teams define revenue differently
Metrics Layer interview questions
Why define metrics in a semantic layer instead of in each dashboard?
When each dashboard computes its own logic, numbers drift and teams argue about which is right. One governed definition gives consistent answers across tools and makes changes in a single place.
What is a ratio metric?
A ratio metric divides one measure by another, such as revenue per order. The layer calculates numerator and denominator at the requested grain before dividing, which avoids the error of averaging pre-computed ratios.
Where Metrics Layer fits in the Analytics Engineer roadmap
Metrics Layer belongs to the Metrics and Semantic Layers stage (intermediate level), which takes about 2 hours. One definition of revenue.
Other units in this stage
Back to the full Analytics Engineer roadmap