Aglet

Prioritize a Weighted Average Fix

An average correction can change rankings, thresholds, and period comparisons without changing any source row. Prioritize it by the decision at risk, the measures sharing the calculation, and the confidence of the weight definition. A clearer label may be safer than replacing a historical measure whose existing meaning is documented.

Decide where the work belongs

  1. Describe the consequence

    Show the decision that changes when a large group receives one vote instead of weight proportional to its observations. Use a fixed fixture and record both results. Keep the consequence tied to the report's actual audience rather than assigning an unsupported percentage or severity score.

  2. Compare metric options

    Evaluate a weighted measure, an unweighted measure, and a paired display that names both. For each option, list denominator rules, historical comparability, affected exports, and the owner who can approve semantics. Include a clarification path when the source field does not represent a valid weight.

  3. Choose a queue outcome

    Queue one bounded action: confirm the metric, add an explicitly named variant, or revise the calculation with fixtures. Attach the numerator and denominator evidence and a revisit trigger for weight-field changes. Keep unrelated rounding or chart-axis work separate from the measure decision.

What to carry forward

Prioritization is complete when one average action has a named decision, owner, evidence threshold, and revisit trigger. Escalate a shared measure change only after the weight and denominator are agreed; otherwise prioritize definition work so a plausible number is not silently reinterpreted.

Technical background: PostgreSQL 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