The monthly report confirms that delivery was late. Unfortunately, the client already knew. A lagging measure describes an outcome that has happened. A leading measure tracks a condition or action expected to influence that outcome, giving the team an earlier opportunity to respond.
Photo by Marco Palumbo on Unsplash.
The relationship is a hypothesis to test. Calling something a leading indicator does not prove it predicts the result. Choose a pair that reflects how your own process works and review the connection over time.
Start with a specific outcome
Choose an outcome the team can define consistently, such as the proportion of agreed delivery dates met. Decide what counts as an agreed date, which cases are included and how legitimate changes are recorded.
Then examine what tends to happen before failure. Missing client input, unresolved approvals or an ageing work queue may provide an actionable warning. Select something the team can influence rather than a number that merely describes external conditions.
Use a measure-pair worksheet
Outcome: the result the team wants to improve.
Lagging measure: calculation, source, owner and reporting period.
Suspected driver: the condition believed to influence that outcome.
Leading measure: how that condition will be observed.
Action trigger: what prompts investigation and who responds.
Countercheck: what prevents the measure encouraging poor behaviour.
Review: whether the observed relationship supports the original hypothesis.
Avoid setting a target before checking whether the underlying information is reliable enough to support it.
An illustrative delivery pair
A service team tracks on-time completion as the outcome. It also monitors upcoming jobs missing essential client input at the agreed readiness check. The response is to obtain the input or renegotiate the commitment before work is blocked.
The team checks that people are not simply moving readiness dates or excluding awkward cases to improve the measure. A cleaner chart is not the same as better delivery.
Retire measures that do not help
If a warning never changes a decision, it may be too late, too noisy or unrelated to the outcome. Adjust the timing or choose another driver. Keep qualitative case reviews alongside the numbers while the relationship is still uncertain.
Related guidance: practical process-mapping guide.
