OEE vs Quality: Why Scrap and Rework Quietly Halve Your OEE

Key takeaways

• Quality is the third OEE factor — good parts vs total parts produced.
• It looks small (most plants report 95-99%) but compounds with start-up scrap, rework, and first-pass-yield losses that often go uncounted.
• Plants that bucket start-up scrap as "planned waste" instead of Quality loss systematically overstate Quality.
• Rework that eventually passes is not free — it consumes capacity and should be tracked as Performance loss even when Quality looks fine.
• Customer returns are a Quality signal that almost never makes it back to the OEE number.

Short answer: Quality measures the ratio of good parts to total parts produced. It looks small in isolation (most plants report 95-99%) but the true Quality number is usually lower because start-up scrap, rework that eventually passes, and downstream returns are systematically under-counted. Quality is the OEE factor most distorted by accounting choices rather than measurement gaps. See also Quality by Design vs Quality by Inspection.

What Quality measures

Quality is the simplest of the three OEE factors:

Quality = Good parts / Total parts produced

If the line produced 1,000 parts and 940 passed first-pass QC, Quality = 94%.

The arithmetic is easy. The hard question is what counts as a good part and what counts as total parts.

What gets under-counted as Quality loss

Start-up scrap. The first parts off a line after changeover are often out of spec. Many plants bucket these as "planned waste" or "set-up scrap" and exclude them from Quality. The PLC counted them; the spec said they failed; they were Quality losses. Excluding them flatters the number.

Rework that eventually passes. A part that needs grinding, repolishing, or re-test before it passes is technically a good part by Quality's definition. But it consumed capacity to fix. Some plants count rework as Quality loss; some count it as Performance loss; many do not count it at all. The result is a Quality number that looks 99% on lines where rework is a real problem.

Customer returns. A part that passed QC but failed in the field is a Quality loss that the line never sees. Without a feedback loop from RMA back to OEE, Quality stays artificially high.

The first-pass yield trap

First-pass yield (FPY) is the share of parts that pass QC the first time without rework. It is a stricter measure than Quality.

A line with 99% Quality but 85% FPY has a 14-point rework problem masquerading as a Quality win. The Quality factor in OEE looks fine because the parts eventually pass, but real capacity is being consumed reworking them.

The fix is to either count rework as Quality loss or count it as Performance loss (capacity consumed beyond the design cycle). Either way it has to show up somewhere, not vanish.

Why Quality usually looks better than it is

• QC sampling. If you only inspect 1 in 20 parts, you find 1/20th of the actual defect rate.
• Generous spec tolerance. Tightening tolerance reveals defects that were always there.
• Aggregation across SKUs. One bad SKU averaged with five good SKUs looks like everyone is at 97%.
• Time lag. Release tests and customer feedback arrive after the OEE report is locked.

What good Quality measurement looks like

1. Count all scrap. Including start-up scrap. Bucket the cause separately, but include the count.
2. Track rework as a distinct loss. Either inside Quality or as a Performance penalty. Pick one. Stick to it.
3. Feed customer returns back. Even with a 30-60 day lag, the signal matters.
4. Report FPY alongside Quality. Two numbers, not one. The gap between them is the rework cost.

How Quality interacts with the other factors

If Quality is 90% and Performance is 90%, OEE drops to 81% on a perfect Availability line. Quality and Performance compound multiplicatively, so even small Quality losses bite when Performance is also imperfect.

The compounding is also why fixing Quality often unlocks more apparent capacity than fixing Availability. Each point of Quality recovered is a point of saleable output not lost downstream — and it usually does not require equipment changes.

How an OEE platform should handle Quality

A real OEE platform pulls Quality from line-side QC (vision, gauges, manual inspection) and lets you tag every reject with a reason code. It exposes FPY alongside Quality so the rework picture is visible. And it integrates with the LIMS or QC system so release-test results back-fill Quality when they land.

Fabrico's OEE module ties Quality to reason-coded reject capture and exposes both Quality and first-pass yield on the same dashboard — so rework cannot hide as a Quality win.

See how Fabrico captures this automatically — explore OEE for manufacturing or book a demo.

Related reading

• Quality by Design vs Quality by Inspection
• OEE vs Utilization
• OEE and ISO 22400
• OEE for Batch Manufacturing

Frequently asked questions

Is Quality the same as first-pass yield?

No. Quality counts any part that eventually passes as a good part. FPY counts only parts that pass the first time. The gap is rework.

Should start-up scrap count as Quality loss?

Yes, in most definitions. Bucket the cause separately, but the count of failed parts belongs in Quality.

What is a good Quality target?

World-class is around 99%. The 85% OEE world-class benchmark assumes Quality ~99%, Availability ~90%, Performance ~95%.

Should I track rework inside Quality or separately?

Either is defensible. Tracking it separately as a Performance loss is closer to truth because rework consumes capacity. Tracking it inside Quality is simpler.

How do customer returns affect Quality?

They should reduce Quality retroactively if you want a true number. Most plants do not connect RMA back to OEE, so the field-failure signal is lost.