On this page

Define an observable problem

Descriptions such as “lack of care” do not establish what happened. Specify the product or batch, detection point, impact and acceptance criterion. Separate verified facts from hypotheses and collect missing evidence before choosing a cause.

Contain the impact and test causes

Identify affected material, work in progress and deliveries, then apply suitable containment. Containment is temporary control, not root-cause analysis.

Investigate both occurrence and escape: why did the defect arise, and why did existing controls fail to detect it? Test hypotheses across equipment, materials, methods, measurement and information flow.

Link actions to verified causes

State the condition each action changes, its owner, resources and acceptance method. Training may help, but it cannot on its own repair a defective process or equipment condition. Update related standards and controls.

Close on evidence

Require both implementation evidence and effectiveness evidence over an appropriate scope and observation period. Check for new risks and consider similar processes. A few conforming samples do not establish sustained stability.

The illustration is an AI-generated quality-inspection setting.