Aglet

Triage a distinct count that changes across report views

Distinct counts answer how many unique entities appear within a defined population, and that population can change across cards, groups, periods, and joins. Triage should name the identity key and comparison scope, then isolate whether the difference comes from legitimate boundaries or counting logic.

Establish what is happening

  1. Name the counted entity

    Record whether the measure counts customers, orders, devices, cases, or another entity, along with the exact identity field. Capture null or placeholder identifiers separately because they may not represent a unique entity.

  2. Align comparison scope

    Compare the report views using the same period, filters, grouping, and join context. Note whether a grand total is being compared with a sum of group counts; overlapping entities can make those quantities legitimately different.

  3. Classify the discrepancy

    Classify the case as expected overlap, identity-quality ambiguity, filter mismatch, join duplication, or unresolved counting logic. Preserve a small set of repeated and unique identifiers so later work can test the classification.

What to carry forward

Triage is complete when the counted entity, identity field, and comparison population are explicit. A differing grouped sum is not a defect by itself; route it for definition review only when the report promises a non-overlapping or otherwise specified relationship.

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