When metrics argue with each other
Conflicting dashboards are a training opportunity, not a reason to pick the friendlier chart.
Two reports can both be accurate and still point in opposite directions. That usually means different windows, different filters, or different definitions of the same word.
In the Data Signal Lab we pause teams before they reconcile numbers. First they write the definition each metric claims to measure. Often the conflict dissolves once naming is honest.
When definitions truly diverge, the briefing should say so. Stakeholders can choose which definition fits the decision; they cannot choose if the conflict is hidden.
A calm analyst earns trust by showing the disagreement early, not by averaging it away.