In this lesson, you will learn how to link your dbt tests to downstream dashboard exposures to gain visibility into how data quality issues impact your business reporting.
Bridging the Gap Between Data and Dashboards
Up until now, we’ve focused on running tests against our dbt models to catch issues like null values or primary key violations. However, when a test fails, it can be difficult to know exactly which stakeholders are seeing incorrect data. An exposure in dbt allows you to document downstream tools—like a Looker dashboard or a Tableau workbook—that consume your dbt models.
By defining exposures, you create a line of sight between your raw data transformations and the final "product" delivered to your business users.
Linking Tests to Exposures
When you define an exposure in a yml file, you explicitly list the models it depends on using the depends_on key. dbt then creates a graph connecting your tested models to your business dashboards.
# models/schema.yml
exposures:
- name: executive_sales_dashboard
type: dashboard
depends_on:
- ref('fct_sales_daily')
owner:
name: Data Team
Because your fct_sales_daily model likely has tests (like not_null or unique) defined in its own yml file, dbt now understands that if those tests fail, your executive_sales_dashboard is officially at risk.
When you run your tests via dbt test, the lineage graph becomes your monitoring hub. If a test fails on fct_sales_daily, you no longer just have a generic "data quality issue"; you have an "impacted business report." This allows you to prioritize fixing models that feed high-visibility dashboards, ensuring you resolve the most critical issues first before your stakeholders report discrepancies.
By keeping these relationships documented, you transform data quality from a technical chore into a business-aligned monitoring strategy, ensuring that when data is broken, you know exactly whose dashboard is reflecting it.
Key Takeaway: By defining exposures in dbt, you link data quality tests to the specific business dashboards they support, allowing you to prioritize fixes based on their real-world impact.