← All insights

Quality · Supplier improvement · Problem solving

The supplier says it’s fixed. How do you check?

How to follow a recurring quality problem from containment and root-cause analysis to a verified corrective action.

2 min read

A replacement shipment, a new work instruction or a training record may be useful. None of them alone proves that a recurring defect has been resolved.

A useful review asks four questions: what is affected, why did it happen, what changed and what do later production results show?

Protect current production

Define the defect clearly and agree how to measure it. Identify affected parts, batches or process conditions as far as the available records allow. Set temporary controls such as inspection or segregation, and agree who decides what happens to suspect stock.

Containment protects the next operation or shipment while the investigation continues. Give it an owner and a review point. Repeated sorting can become a costly routine if nobody decides what would allow it to stop.

Ask why it happened and why it wasn’t caught

The cause of a defect and the reason it escaped inspection may be different. Both may need attention.

Compare defect samples, process settings, drawing revisions and measurement results. Test likely causes against the observations. If the records are incomplete, identify what information or trial would help distinguish one explanation from another.

Don’t assume every defect is an operator error. Design, tooling, materials and measurement can also contribute.

Record the cause each action addresses, who owns it and when it is due. Check that the change reaches the actual operation: updated instructions are in use, modified tooling produces the intended part, and inspection uses a suitable method.

Training can support a change. It won’t fix a process that still relies on unclear instructions or a difficult assembly method.

Review later production

Agree the follow-up conditions and acceptance criteria before closing the issue. One replacement sample may not represent normal variation.

Check:

  • Was the agreed change implemented?
  • Do later results show the intended improvement?
  • Can temporary controls be reduced or removed?
  • Which controls and records should become part of routine production?

If the problem returns, revisit the cause. Repeating the same action without understanding the recurrence adds work without resolving the defect.

Explore Supplier Quality Improvement for recurring problems, or Quality for inspection, analysis and improvement.