Published:
Classify the effect, not just the visible mark
The same mark can have different consequences on different products. A sharp edge on an accessible handle is different from a small finish mark on a concealed surface. A missing label may be a small presentation issue or a serious traceability or safety issue.
Use the product specification, intended user, expected use, and agreed defect catalogue. If the classification is unclear, hold the decision and ask the responsible quality owner. The inspector should not invent a lower class to keep shipment moving.
Critical defects: act on the safety risk
Examples that may be critical include an exposed live electrical part, a sharp point accessible during normal use, or contamination in a product with a relevant safety requirement. These are examples for discussion, not a universal defect list.
Stop release of the affected material under the agreed containment procedure. Identify the lot and notify the responsible quality owner. Do not let a low overall defect count cancel a critical finding. A sample with no critical finding does not prove the whole lot is safe.
Major defects: check function and buyer acceptance
Examples that may be major include a closure that will not lock, an incorrect component, or a dimension outside the agreed tolerance that prevents assembly. Compare the finding with the approved specification and test method.
Record the measured result or failed function. “Bad part” gives the supplier no usable instruction. “Latch does not retain the lid under the specified test” identifies the failure and the check that must be repeated after correction.
Minor defects: define appearance limits
A small cosmetic mark may be minor when it does not affect function and falls within the buyer definition for minor defects. An appearance issue can still be major on a prominent surface or a product sold mainly for appearance.
Use photographs with consistent lighting, distance, and scale. Agree the viewing method and boundary examples before inspection. Inspectors who use different lighting or viewing distances can disagree even when they follow the same written sentence.
Build a defect record the office can use
- Defect code and a short description.
- Product, order, lot, and unit or carton reference.
- Observed result and the requirement that was not met.
- Proposed class and the approved rule used.
- Photo showing the location, plus a close view with scale when needed.
- Inspector, time, owner, containment action, and decision status.
Keep one defect catalogue with a revision date. When the buyer changes a limit, replace the controlled version and tell the inspection team. A corrected message buried in Zalo chaos is not a reliable specification.
Keep severity, frequency, and release separate
Severity describes the consequence. Frequency describes how often the finding occurs. The lot decision applies the agreed acceptance plan and any mandatory hold rules. A rare major defect remains major. A frequent minor defect may still cause a lot to fail.
For a nonconforming-unit sampling plan, several defects on one unit do not automatically become several nonconforming units. Record all findings, then apply the agreed counting rule. Use the same rule in the calculator, inspection sheet, and final report.
Review disputed examples with the buyer before the next lot. The useful output is an approved rule and a reference image, not another informal opinion.
Key takeaways
- Use product-specific, buyer-approved defect definitions.
- Classify the consequence separately from the number of findings.
- Record the failed requirement and evidence, not only a severity label.
- Keep one controlled defect catalogue for the floor and office.
Sources and related guides
Put this into practice with KaizenQ
KaizenQ gives your team a structured digital inspection workflow: defects logged with photos on the floor, a live report ready to share as soon as inspection is complete. No manual compilation. No end-of-day rework.