Build a useful investigation brief
Define a known-unit fixture
Choose a small set of independently known values and their expected source and display units. Include a value that exposes scaling and a zero or negative value where meaningful. Record the intended aggregation so the comparison can distinguish unit conversion from a separate change in how the measure is summarized.
Trace conversion and formatter boundaries
Inspect the query result, any conversion layer, formatter input, and rendered label. Record the first disagreement with the unit contract. Compare reused formatters across series or cards to see whether a unit-specific label is being applied to another measure whose source or scaling rule differs.
Compare chart and export contracts
Follow the same fixture into a table, tooltip, and supported export. Check whether each surface converts values or merely labels them. Evaluate one implicated change against the full set so fixing a chart suffix does not leave exported values unconverted or cause a shared formatter to scale another report a second time.
What to carry forward
Produce a unit trace, expected transformation, and first incorrect conversion or label. Separate measure-definition ambiguity from implementation evidence. A supported correction makes numeric meaning consistent across the affected surfaces and states any remaining mixed-unit comparison, rather than hiding it behind a shared axis title or a convenient abbreviation.
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