Defining good performance indicators is one of the most underestimated exercises in management. The S.M.A.R.T framework structures thinking — but doesn't replace understanding what you actually want to measure.
A KPI (Key Performance Indicator) should focus attention on what truly matters. In practice, most dashboards accumulate metrics that measure activity rather than value: tickets closed, meeting hours, lines of code produced. Measuring activity often creates an illusion of control.
The S.M.A.R.T framework
To structure KPI definition, the S.M.A.R.T framework is a solid starting point:
- Specific: the KPI addresses a precise subject, not a vague notion. "Improve customer satisfaction" is not a KPI. "Reduce monthly churn rate" is.
- Measurable: the data exists, is accessible and reliable. An indicator that cannot be objectively measured is not a KPI.
- Achievable: the associated objective is realistic given available resources. An impossible KPI demotivates.
- Relevant: the KPI aligns with the organisation's strategic goals. A metric relevant to one team may be noise to another.
- Time-bound: a deadline is defined. A KPI without a time horizon is not actionable.
The limits of S.M.A.R.T
S.M.A.R.T doesn't solve everything. It doesn't guarantee you're measuring the right thing. The distinction between leading indicators (predictive, like NPS or number of demos scheduled) and lagging indicators (past results, like revenue or churn rate) is essential. Too many KPIs are lagging indicators: they identify a problem after it has already occurred.
Classic mistakes
- Too many KPIs: beyond 5-7 indicators per team, attention dilutes
- KPI without an owner: who is responsible for achieving it?
- KPI without a baseline: without historical reference, progress cannot be measured
- Gameable KPI: any indicator can be manipulated if it's in the measured team's interest
The golden rule: a good KPI changes behaviour. If it doesn't prompt anyone to act differently, it serves no purpose.