A Key challenge for Organizations is to ensure that data metrics refer to the same for all teams. If BI logic is scattered across various tools, SQL and notebooks, a metrics tax is levied (multiple dashboards showing different revenues). Databricks Metric Views resolves it by allowing the creation of semantic metric views in UC.
Metric Views allows defining calculations and the attributes. When a user checks for Regional Margins, Databricks generates the code to aggregate the data correctly for that specific region using info provided in the view. We can provide complex logic — Period growth averages etc and it inherits the security and lineage of the UC. The logic is certified. Organizations can create excellent AI BI dashboards & genie in a workspace powered by Metric Views.
I have watched teams define the most critical BI logic in the wrong areas. Locking BI calculations inside Tableau/ Power BI reports, churn logic in dbt and in ad hoc SQL. The result was always the same. Org dashboard showed one value, the Marketing dashboard said different and no one trusted the bi. That era is over with Metric Views.
Recently migrated KPIs to Metric Views for an enterprise and the shift was great. Instead of writing ad hoc bi code in every BI report, we defined it in metric views. Whether a user queries via SQL, connects a dashboard or checks for the views in AI BI dashboard / Genie a question, Databricks calculates the same logic given in views.
The compatibility with the AI BI stack is great. Stop building BI metrics in the reports layer and move the definitions as Metric Views.
