Which metrics actually change a decision?

Which metrics actually change a decision?

A metric deserves management attention when a defined movement triggers a different action. If nobody can say who responds, at what threshold and with which lever, the number may be informative but it is not yet a decision instrument.

Work backwards from the decision. Pair a lagging outcome�retention, margin, safety or cash�with a small set of leading drivers that teams can influence in time. Add guardrails to prevent local optimisation from damaging quality, risk or customer value, and keep diagnostic measures available for investigation rather than crowding the executive view.

Every material metric needs a contract: business meaning, formula, population, exclusions, grain, time window, currency or unit, source, refresh time and accountable owner. State whether it is a stock or flow and whether revisions restate history. A stable label over a changing definition is more dangerous than an obvious data failure.

Measurement also changes behaviour. Kaplan and Norton�s work on the balanced scorecard stressed that financial outcomes alone are lagging signals and that measurement systems shape employee choices. Test for gaming, substitution and short-termism; pair speed with quality, growth with economics and automation with customer or control outcomes.

Review the portfolio as strategy changes. Track which metrics produced decisions, which alerts were ignored and which interventions improved outcomes. Retire vanity indicators and thresholds that create noise. The goal is not a perfect dashboard�it is a compact feedback system that directs scarce attention toward actions with a credible path to value.