Keep the lesson for the next incident
Retain cutoff exemplars
Keep clear ranks, a tie at N, below-cutoff values, null or unknown category, empty category, and filtered recomputation fixtures. Record expected members, Other contents, totals, colors, legend, and export. Preserve the original ambiguous view beside the corrected meaning.
Classify recurring disputes
Review earlier questions for changed N, tie handling, hidden buckets, null grouping, filter timing, and totals that failed to reconcile. Group named cases by missing ranking or display contract. Avoid treating every top-N difference as a query defect.
Assign ranking stewardship
Make the report owner maintain N and tie definitions while the metric owner approves ranking measure and totals. Add a trigger for cutoff, filter, legend, or export changes. State which bucket cases the fixtures detect and which product decisions still require human judgment.
What to carry forward
The learning record is complete when it links top-N confusion to a missing ranking or bucket definition, preserves boundary fixtures, and names an owner and trigger. Keep intentional summary cutoffs labeled so future readers can distinguish aggregation from omitted data.
Technical background: Google Cloud 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