We Have Too Much Scrap and Rework
Rework is hidden capacity loss: the hours to build a unit twice are not in the plan. Scrap found at the end of the line is worse, because every step's hours are already in the part. First pass yield by step tells you where to look.
What Is Usually Going On
- Yield is measured at final test only, so the step that makes the defect is unknown.
- Rework loops back without being counted, so yield looks fine.
- The process is not capable: its spread is wider than the tolerance.
- Defects are found far downstream from where they are made.
- Fixes are retraining and more inspection, not mistake-proofing.
How to Fix It
- Measure FPY by stepFPY = units good the first time / units started, at each step. Rework counts as a fail.
- Roll it upRolled throughput yield = the FPY of each step multiplied together.
- Pareto by hoursRank defects by hours lost (rework hours plus hours in scrapped parts), not by count.
- Find the root causeFishbone, then 5 Why on the top defect. Verify with data at the step.
- Check capabilityIf the spread is wider than the tolerance, inspection will not fix it. Use SPC and process capability.
- Mistake-proof the stepPoka-yoke where the defect is made, so it cannot be made or cannot pass.
Start Free
Guides and web tools on this site. No account needed.
When It Moves to the Floor
The same work as an offline file you keep: saved scenarios, print and export, no internet needed on the floor.
Common Questions
What is the difference between FPY and final yield?
Final yield counts units that pass in the end, reworked or not. FPY counts only units that pass the first time. The gap is the rework.
Should rework count as a fail?
Yes, for FPY. Otherwise the rework hours disappear from the numbers.
Is more inspection the answer?
Inspection finds defects; it does not stop them. Use it to protect the customer while you fix the step that makes them.
How do I size rework?
In hours: rework hours per unit times units, plus the hours already in scrapped parts. Show the total as FTE.