Warehouse native experiments let you analyze results against trusted data already living in Snowflake, BigQuery, Redshift, or Databricks — but locking in your metrics and attributes at setup time doesn't reflect how real experimentation works. Now you can add or update metrics and attributes on a warehouse native experiment at any point during its run, not just when you first create it, so your analysis can evolve as your questions do.

To learn more, read Warehouse native metrics and Creating experiments using warehouse native metrics.