dbt · Testing & Documentation

What are exposures?

Mediumexposureslineagebiimpact analysis

Interview question

What are dbt exposures, and why would you define them?

How to answer

The short answer

An exposure declares a downstream use of your dbt models outside the dbt project: a dashboard, an ML feature table consumer, a reverse-ETL sync, or an executive report.

Example

YAML
exposures:
  - name: executive_revenue_dashboard
    type: dashboard
    maturity: high
    url: https://bi.example.com/dashboards/revenue
    description: Daily revenue by region for leadership.
    depends_on:
      - ref('fct_orders')
      - ref('dim_regions')
    owner:
      name: Analytics
      email: [email protected]

Why exposures matter

  • Lineage reaches real consumers, not only other SQL models
  • Impact analysis: "changing fct_orders breaks this Looker dashboard"
  • Ownership and URLs live next to the data product
  • Selection: dbt build --select +exposure:executive_revenue_dashboard

Mental model

Models are producers. Exposures are named dependents that dbt does not build, but must not break silently.

Interview tip: "Exposures close the gap between warehouse models and business-facing artifacts so lineage and ownership are explicit."