Scrap vs Rework: The Disposition Decision

Scrap versus rework is the disposition decision for a defective unit. Scrap writes the unit off as a total loss; rework brings it back to specification with additional work, material, or cycle time. Both are internal-failure costs, and the choice you make changes both your cost and which quality metric moves, which is why the "obvious" call is often the wrong one.

On the floor it looks like a simple question: can we fix it or not? But the honest answer depends on things the operator staring at the part cannot see, whether the plant is starved for capacity, what the part has already cost, and how much a reworked unit will quietly cost again before it ships. A defect you can rework is not automatically cheaper than one you scrap. Sometimes it is far more expensive. This post lays out the real difference, when each costs more, and a decision that holds up better than gut feel.

What is the difference between scrap and rework?

The difference is recoverability. Scrap is defective product that cannot be repaired, used, or sold, it is discarded and written off. Rework is defective product that failed inspection but can be brought back to specification through additional processing: a re-weld, a re-machine, a re-test, a touch-up. The American Society for Quality classifies both as internal failure costs within the cost of quality, but they behave very differently on the plant floor.

Which costs more, scrap or rework?

It depends on whether the plant has capacity to spare. This is the counterintuitive part, and it flips the usual assumption that rework is always cheaper because "at least you keep the unit." The two cases:

• If the plant is capacity-constrained selling everything it can make, with the bottleneck running flat out, then rework usually costs more than scrap. Reworking a unit consumes machine time on a line that has none to give, and every minute spent on the second pass is a minute not spent making a brand-new, sellable unit. The rework isn't just its own labor; it is a lost sale of whatever that capacity could have produced.
• If the plant has excess capacity the line has idle time, demand isn't the limit, then scrap usually costs more than rework. The machine time to rework was going to sit idle anyway, so recovering the accumulated value of the part for a bit of extra labor beats writing the whole thing off.

The trap is that most plants default to one answer for both situations. A capacity-constrained plant that reflexively reworks everything is starving its own throughput to save units it didn't need to save; a plant with idle lines that reflexively scraps is throwing away value it could cheaply recover. The right call is situational, and it hinges on a question the person holding the part rarely has the data to answer.

Why do reworkable defects cost more than they look?

Because the rework cost people quote is only the visible part. The full cost of a reworked unit stacks up in ways that never make it onto the rework ticket:

• Capacity, spent twice. The unit already consumed machine time on its first pass; rework spends it again. On a constrained line this double-handling is the largest cost by far, and it is invisible in the labor number.
• Handling, tracking, and retest. A reworked unit gets pulled, staged, re-processed, re-inspected, and re-entered. Each touch is labor and each is a chance to lose or damage the part.
• Residual escape risk. A reworked unit has been disturbed. Rework that isn't tightly controlled can introduce new defects or leave the original one partly addressed, so some reworked units ship as latent problems, the most expensive defect of all, the one the customer finds.
• Metric distortion. Because reworked units ship, they never touch the scrap rate. A plant can run a huge rework operation and post a proud, low scrap number, hiding a hidden factory that a look at first pass yield would expose instantly.

This is why "it's reworkable" is not a reason to relax. A reworkable defect is still a first-pass failure, still a symptom of a process that didn't do it right the first time, and still a draw on capacity and quality risk. The rework option makes the loss easier to hide, not smaller.

By the numbers. Scrap and rework together are the heart of internal-failure cost, and internal failure is one of the four buckets the American Society for Quality uses to show the total cost of quality commonly reaching 15–20% of sales revenue. ASQ's guidance to reduce scrap and rework costs treats them as a pair for exactly this reason: both are money already spent on product that generated no revenue, and the disposition decision only chooses how you pay, not whether you do.

How do you decide scrap versus rework?

Make it a rule, not a reflex, and put the capacity question at the center. The framework:

1. Confirm the unit is genuinely recoverable. Rework that only partly fixes the defect is worse than scrap, because it ships risk. If you can't return it fully to spec, scrap it.
2. Value the unit at its accumulated cost. Know what the part has already soaked up (see scrap rate by cost). Rework only makes sense when its cost is well below that value.
3. Ask whether the line is capacity-constrained. If the bottleneck is maxed out, price the rework as the sellable unit it displaces, often the deciding term. If there is idle capacity, weight rework more favorably.
4. Add the hidden costs. Handling, retest, and escape risk go on the rework side of the ledger, not just direct labor. Reworked units that escape as defects are the most expensive outcome.
5. Write the decision into the disposition procedure, with defect codes for both routes feeding defect tracking. A consistent rule beats case-by-case judgment made under production pressure.
6. Attack the root cause regardless of disposition. Scrap or rework, the unit failed first pass. The disposition manages today's loss; root-cause work stops tomorrow's, which is the only move that actually lowers the total cost.

When is rework a warning sign rather than a solution?

Rework earns its keep as an occasional recovery. It becomes a problem when it hardens into a permanent fixture, a bench with a full-time crew, a standing "touch-up" step everyone plans around, a rework loop so normal that no one questions why it exists. At that point rework has stopped being a disposition decision and become a hidden production department, one that consumes labor and capacity while producing nothing new. The tell is when rework hours are budgeted rather than tracked as loss: a plant that staffs for rework has quietly accepted its defect rate.

Two patterns should trigger a hard look. First, when rework becomes the routine way to hit the schedule, building fast-and-loose because the rework bench will "catch it", the plant has traded first pass yield for the illusion of throughput, and the plant KPIs will show it as a low FPY paired with a suspiciously clean scrap rate. Second, when rework is uncontrolled, without its own procedure, records, and re-inspection, it turns into a channel for escapes; disciplined operations treat rework as a controlled process with the same rigor as first-run production, a principle formalized in regulated settings such as rework management in food safety. In both cases the fix is upstream: the disposition decision manages the unit in front of you, but only root-cause work shrinks the rework department itself.

How does the scrap-or-rework choice affect your metrics?

The two dispositions land on different numbers, and knowing which is the whole reason the metrics have to be read together. Scrap raises the scrap rate; rework does not, because reworked units ship. But both are first-pass failures, so both pull down first pass yield and, across a multi-step line, rolled throughput yield. That asymmetry is the tell: a plant with a low scrap rate and a much lower first pass yield is reworking heavily, and the gap is the hidden factory. If you watched only scrap rate, you would never see it.

Both dispositions also feed OEE and the six big losses as reject losses, and both cap throughput scrap by wasting the machine time already spent, rework by spending it a second time. The honest read requires capturing the disposition, the reason, and the capacity state at the moment of decision, which is precisely what paper logs drop. Plants that record disposition and defect codes at the station, the way Harmony turns paper quality logs into live records feeding root-cause analysis (see the platform or the CLS field story), can finally see scrap and rework as one connected cost, and decide each case on data instead of habit.