By measuring the thing that failed, over a period long enough for it to fail again, against a criterion defined before the corrective action was implemented. Confirming that training was delivered is not an effectiveness check.
Define the check before the action
If the effectiveness criterion is written after the corrective action is complete, it will be written to be satisfiable. Defining it at the point of action planning forces the question of whether the action can plausibly work.
Measure the failure, not the activity
- Complaint rate for the specific failure mode, not complaint volume overall
- Recurrence of the same nonconformance code within a defined window
- In-process reject rate at the operation that failed
- Audit re-check of the specific practice, performed by someone independent
Give it enough time
A thirty-day window on a failure mode that historically appears quarterly proves nothing. The observation period should be set against the historical frequency of the event.
Reviewed by the named consultant. Last updated July 14, 2026. General information only — not regulatory or legal advice for a specific situation.