First steps
- Analyse returned parts (no-fault-found share too).
- Weibull by production month.
- Verify the fix with accelerated tests.
How to measure success
Warranty incidents per thousand vehicles and cost. Record the value before you start, so the effect of every action can be shown.
When these methods are the wrong choice
- Weibull analysis: Defects found at the end of line.
- 8D report: Small internal issues where one person can fix the cause in a day.
- PFMEA / DFMEA: Fixing a problem that has already happened; update it after the 8D instead.
- Pareto chart: A single defect type with one known cause.
- Hypothesis tests: Exploring causes; use it to confirm them.
- Cost of poor quality (COPQ): Root cause analysis.
The methods in one line
| Method | What it is |
|---|---|
| Weibull analysis | Models failures over time or mileage: early failures, random failures or wear-out. |
| 8D report | Team-based method for customer complaints: containment, root cause for occurrence and non-detection, verified corrective action, prevention. |
| PFMEA / DFMEA | Analyses how a process or design can fail, rates the risk (AP) and defines prevention and detection. |
| Pareto chart | Ranks defect types or causes so the team works on the vital few first. |
| Hypothesis tests | Proves whether a difference (before / after, machine A / B) is real or chance. |
| Cost of poor quality (COPQ) | Puts a money value on scrap, rework, sorting, warranty and complaints to set priorities. |