A corrective action is complete. Is the quality problem closed?

A replacement shipment, an updated instruction or a retraining record can be useful. None of them, on its own, proves that a recurring defect is under control. The action may be implemented while the mechanism that created the defect remains unchanged.

A practical review separates four questions: what is affected, what causes the failure, what changed, and what later evidence shows.

Contain the immediate risk

Define the failure in observable terms and agree the measurement method. Identify the potentially affected parts, lots or process window as far as the available evidence allows. Set temporary inspection, segregation or other containment appropriate to the issue, with clear authority for disposition decisions.

Containment protects the next operation or shipment while the investigation proceeds. It should have an owner and review criteria. Repeated sorting can become an expensive routine if nobody decides what must happen before it can end.

Investigate occurrence and escape separately

Ask why the defect was created, then ask why the existing controls did not prevent or detect it. These may be different causes requiring different actions.

Use defect samples, process records, revision history and measurement checks to compare explanations. A root-cause label is not a substitute for evidence. Where records are incomplete, describe the uncertainty and define the next observation or controlled check that could distinguish competing explanations.

Avoid treating a supplier’s assurance as proof. Equally, avoid presuming that every failure is caused by operator behavior when the design, tooling, material or measurement system may contribute.

Connect each action to a supported cause

An action register should identify the cause it addresses, the responsible person or team, the due date and evidence of implementation. An instruction change needs to reach the relevant operation. A tooling change needs to be reflected in the parts produced. A new inspection needs a suitable method and acceptance criteria.

Training can support these changes, but it should not conceal a process that still depends on an impractical or ambiguous instruction.

Define effectiveness before declaring closure

Agree what follow-up production will be reviewed, under which conditions, and what result would support closure. The scope depends on the failure and risk; a single replacement sample may not represent normal production variation.

Keep these decisions explicit:

• Has the agreed action been implemented?
• Does subsequent evidence support the intended effect?
• Can temporary containment be reduced or removed?
• Which controls or records should become part of routine work?

If the defect returns, reopen the assumptions and evidence. Repeating the same action without understanding recurrence adds activity without establishing control.

This is a general method, not a claim about a particular supplier or achieved defect reduction. See Supplier Quality Improvement for the problem-led solution and Quality for the broader capability.