Good containment produces data you can act on: parts inspected, parts rejected, defect type, and defect PPM by shift. Read together, these numbers show whether the problem is stable, shrinking, or still leaking, and they are the evidence the customer uses to decide when you can exit. Sorting without capturing this data is just moving boxes; the data is what protects you.
Sort yield and defect PPM
Defect PPM is rejected parts divided by inspected parts, times one million. It puts a small-looking reject rate into the language OEMs score you in. Sort yield, the share of parts that pass, is the same story from the other side. Track both by day and by shift, because a rate that looks fine overall can hide a bad shift, a bad machine, or a bad lot.
Trend is the real signal
A single day of clean parts does not prove control; a downward defect-PPM trend across consecutive shifts does. That is why exit criteria are usually written as a number of consecutive clean days rather than a single sample. Plotting the trend also tells you whether your corrective action actually worked or whether you are just getting lucky between escapes.
Turning data into an exit
Package the data the way the customer will read it: defect definition and boundary samples, daily inspected and rejected counts, PPM trend, and the clean-point date when certified stock began. That package is what supports your 8D and your request to come off GP12 or controlled shipping. Clean, independent data carries more weight than your own tally, which is one reason customers ask for a third party.