The short version
Product quality improves when a team stops sorting good from bad at the end of the line and starts removing the reasons defects happen. The method is the same in a small workshop and a large plant: measure, rank, find causes, fix, verify.
How to do it, step by step
- Define good. Write down the specification, the visual standard and the customer requirement for the product. A defect nobody has defined cannot be counted consistently.
- Measure where defects occur. Count defects, rework and scrap at the process step that caused them, not only at final inspection. A simple check sheet is enough to start.
- Rank the problems. Use a Pareto chart to find the few defect types that account for most of the loss. Work on those first.
- Find the root cause. Ask why until you reach something the process can change, using the 5 whys or a fishbone diagram. See root cause analysis.
- Fix the process. Prefer changes that make the error hard to make, such as a fixture, a sensor or a poka-yoke, over reminders and retraining alone.
- Verify the fix. Watch the same measure for several weeks. If the defect rate has not moved, the cause was not the real one.
- Standardise and sustain. Update the work instruction, train the team, and keep the measure on a daily board so a slip shows up in a day, not a quarter.
Common mistakes
- Adding more inspection instead of removing the cause. Inspection finds defects; it does not stop them.
- Fixing the symptom, for example reworking parts, and calling it a corrective action.
- Measuring only the final result, which hides which step is responsible.
- Skipping verification, so a fix that did not work stays on the books as done.
Where to start tomorrow
Pick one product line, count defects by type for a week, and chart them. The biggest bar is your first project.