Entry 0136·August 21, 2026·Automation·Throughput, Variability & Constraints

Count Rework Without Counting Output Twice

A reworked pack can eventually become acceptable product. That does not mean its first pass was good or that the extra work was free.
Truth · decision framework

A reworked pack can eventually become acceptable product. That does not mean its first pass was good or that the extra work was free.

The measurement needs to follow both the unit and the additional processing.

Separate unique output from processing passes

Identify the unit, batch or other traceable quantity that enters rework. Record why it was diverted, what happened next and the final disposition.

Keep unique accepted output separate from total passes through an operation. If 100 packs enter and 10 require one additional pass, the operation may record 110 processing passes while only 100 unique packs are ultimately accepted. The example assumes no scrap and no further rework.

Counting every pass as new finished output inflates production. Counting only final acceptance hides the original failure if no separate first-pass record exists.

Charge the work where it occurs

Rework may use the original line, another station or a manual table. Record the actual labor, material and equipment time used rather than assuming each rework pass costs the same as the first.

Link the work to its source configuration where the evidence supports that link. A defect category is a useful starting point, not proof of cause.

If rework consumes a constrained resource, assess the effect on scheduled good output. If the resource has spare time, the cost may be different. Avoid counting both lost contribution and the same capacity cost without explaining the basis.

Reconcile the quality measure

Write down what the site's quality factor counts and where its inputs originate. A reported 100 percent can be accurate for a limited sample, or it can mean defects are not captured. Test the counting path before declaring either.

Use the same definitions across baseline and comparison periods. Track changes in product mix and rework routing.

A useful review pairs unique good output with first-pass results and extra work. It can show that recovery was effective while still making the cost of recovery visible.

Published August 21, 2026
Related reading in Throughput, Variability & Constraints