Performance Indicators and Metrics
Measuring what matters
A metric is a measurement or set of measurements taken over time. A key performance indicator (KPI) is a metric that is important enough to be used to measure achievement of a critical success factor. A critical success factor (CSF) is what must be done to meet an intended objective.
The hierarchy to remember: objectives → critical success factors → KPIs → metrics. Confusing CSF with KPI is a standard distractor.
Good metrics are relevant to the objective, specific in what they measure, achievable with the data available, and timely - close to real time, because a metric reported weeks later cannot support decisions. Prefer end-to-end, business-level measures over component-level ones that are easy to collect but say little about value. Leading indicators predict; lagging indicators confirm what already happened. SMART (specific, measurable, achievable, relevant, time-bound) is a useful sanity check for targets.
Worked example: an SLA promises 99.9% monthly availability. Relevant metric: percentage of agreed service hours available. Supporting metrics: mean time to restore service, percentage of incidents resolved within target, user satisfaction score. Volume of tickets opened is not a KPI for availability - it is context, and it is easy to inflate.
Traps: collecting many easy metrics instead of a few meaningful ones; treating every metric as a KPI; measuring activity (tickets closed) rather than outcome (service restored, satisfaction); using lagging metrics alone when a service is degrading; forgetting that measurement exists to support decisions, not to fill reports.
Rule of thumb: before accepting a metric, ask "what decision would change if this number moved?" If the answer is none, it is not a KPI.