Check whether the outcome improved
Define propagation assertions
Write each control's expected effect, card population, measure, default, exclusions, and value for the fixture. State which controls are intentionally independent. A pass requires the card label or nearby explanation to identify its scope and aggregation.
Exercise selected and excluded rows
Run no selection, single selection, multi-selection, excluded value, and independent-control cases. Compare included identifiers, card totals, neighboring visuals, and export results. Record filter state and report revision after each action rather than relying on a final screenshot.
Review the headline meaning
Ask a reviewer to explain what population the card shows after each filter and why an independent control does not change it. Preserve source rows, calculation trace, and visible labels. Record a bounded result when propagation is correct but scope depends on hidden model rules.
What to carry forward
Verification passes when intended controls produce the documented card populations and independent controls remain understandable. Report a bounded pass for uncovered model paths. Keep verification inconclusive when the card changes without a visible reason or its label cannot state which filters apply.
Technical background: Metabase documentation.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow