Aglet

Prioritize a Metric-Grain Correction

A grain problem can make several numbers look plausible while answering different questions. Prioritize it by the decision that could change, the number of report surfaces sharing the measure, and the strength of the metric definition. A local label clarification may outrank a broad query rewrite when the underlying grain is intentional.

Decide where the work belongs

  1. Describe the decision exposure

    Show which choice depends on the measure and whether an event-versus-entity interpretation changes that choice. Use a fixed example with repeated memberships or multiple events. Do not estimate prevalence; record the observed rows, groups, and visible totals instead.

  2. Compare correction paths

    Evaluate a metric-definition change, a separate measure at the needed grain, and a clearer report label. For each option, list affected visuals, historical comparability, fixture requirements, and the owner who can approve the meaning. Keep an explicit defer path when policy is unsettled.

  3. Set the queue boundary

    Choose one next action and its evidence threshold: clarify the metric, split the visual, or revise aggregation. State what remains out of scope, such as unrelated filter behavior or export formatting. Attach the grain fixture and a trigger for revisiting the priority.

What to carry forward

Prioritization is complete when one grain action has a named decision, owner, evidence threshold, and revisit trigger. Escalate a shared measure change only when the intended row unit is documented; otherwise queue definition work before altering numbers that readers already compare.

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