Check whether the outcome improved
Exercise the intended unit transformation
Render known values using the documented source and display units. Compare exact expected magnitudes and relevant zero or negative cases. Keep aggregation fixed so the verification does not confuse a unit repair with a separate change to totals, averages, or the population included in the report.
Check all visible unit references
Inspect axis labels, legends, tooltips, table headings, and any explanatory definition. Confirm each refers to the actual value being shown. Where series use different units, check that the comparison remains explicit rather than sharing a label that incorrectly suggests the magnitudes can be compared directly without conversion or context.
Compare supported exports and reuse
Export the controlled result where supported and inspect both value and unit. Check a neighboring report that reuses the formatter or measure. Confirm the correction neither double-scales its values nor changes a unit-specific label on a surface that already followed the intended contract before the fix.
What to carry forward
Accept when known magnitudes and labels agree across the affected chart, detail, and export paths. Record surfaces not exercised and any unresolved mixed-unit interpretation. A visually plausible number remains incomplete evidence unless its source unit, conversion, and displayed meaning have all been compared with the independent fixture.
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