Check whether the outcome improved
Compare applied schedule variants
Generate supported previews or controlled artifacts for known filter and period variants. Compare values with independent expectations and confirm each artifact identifies its population and window. A shared report title should not make two intentionally different scheduled views appear to represent the same underlying result when their selections materially change the numbers.
Check timestamps and measure context
Inspect generation time, data freshness information where available, units, and relevant measure labels. Confirm each timestamp states what it represents. Exercise an older-data case if supported so a generation timestamp cannot conceal a cached result.
Inspect the linked report destination
Follow a supported destination link and compare its parameters with the artifact’s context. Confirm it preserves the intended view or clearly explains that current data may differ. Reopen a saved artifact after a report default changes where feasible, checking that readers can still identify the original snapshot without reconstructing its schedule from memory.
What to carry forward
Accept when controlled artifacts preserve intended values and visible context, and their destinations accurately explain the relationship to current data. Record delivery verification separately. Keep unavailable freshness or historical-state checks explicit; generating a snapshot successfully does not prove its data is current or that a later interactive view reproduces it exactly.
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