Blog / Quality
Quality · Guide

How to Calculate OEE Quality (Formula and Example)

OL By OEE Lab Editorial |Updated 1 November 2026 |6 min read
Short answer

OEE quality = good units ÷ total units made. Good means it meets spec the first time, with no rework. Count scrap and rework as not good, and count them once, so a quality of 96% means 4 of every 100 units were lost.

What OEE quality measures

Quality is the third factor of OEE. It asks: of everything the machine made while it ran, how much was good the first time?

Stops belong to availability. Slow running belongs to performance. Quality looks only at the units.

Quality = Good units ÷ Total units made
Total units = good + scrap + rework
Quality loss (min) = (Total − Good) ÷ Ideal rate

This follows Overall Equipment Effectiveness (R. C. Hansen, 2001), where quality rate is good units divided by total units produced.

Worked example

Example: a line makes 14,700 units in a shift. 380 are scrapped and 208 need rework. The rest, 14,112, pass first time.

StepCalculationResult
Not good units380 scrap + 208 rework588
Good units14,700 − 58814,112
Quality14,112 ÷ 14,70096.0%
Quality loss in time588 ÷ 40 units per min (ideal rate)14.7 min
Share of 480 min scheduled14.7 ÷ 4803.1%
Good units, first pass14,112 · 96.0%
Rework208 · 1.4%
Scrap380 · 2.6%
Where the 14,700 units went. Only the teal share counts as good.

Notice the two views. Quality is 96%, but the time lost to bad units is 3.1% of the shift. The second number is the one that sits in OEE next to stops and slow running.

Put it together with the other factors from the same shift: availability 87.5% (420 of 480 minutes), performance 87.5%, quality 96%. OEE = 0.875 × 0.875 × 0.96 = 73.5%. Check: (14,112 ÷ 40) ÷ 480 = 352.8 ÷ 480 = 73.5%.

The shift adds up to 100%: 352.8 min of good output, 14.7 min of quality loss, 52.5 min of speed loss and 60 min of stops make 480 minutes.

How rework counts

A unit that fails the first pass is not good, even if you fix it later. Hansen treats the original time as lost. He then gives three cases for the rework itself:

  • Rework offline on other equipment or by hand: the bottleneck OEE does not change. Record the quantity so you can reconcile with first pass yield.
  • Rework on the same machine outside scheduled time (weekend, overtime): it is a cost decision, and it shows in TEEP, not OEE.
  • Rework on the same machine inside scheduled time: treat it as new input. Its time, speed and quality count in current OEE.

Hansen gives a clean example. The ideal output is 100 units an hour and normal running is 75% OEE. In week 1, eight of 168 hours produce 800 units in the wrong boxes. Those hours score 0%, so the week is (160 × 0.75) ÷ 168 = 71.4%.

In week 2 the 800 units are fed back over 12 hours and 780 come out good. That is 780 ÷ (12 × 100) = 65% for those hours. Week 2 had 144 scheduled hours, so it is (12 × 0.65 + 132 × 0.75) ÷ 144 = 74.2%. Over both weeks: (292 × 0.75 + 8 × 0 + 12 × 0.65) ÷ 312 = 72.7%.

Normal running (reference)75.0%
Week 1: 8 h of wrong-box units score 0%71.4%
Week 2: 800 units fed back, 780 good74.2%
Both weeks together72.7%
Hansen's rework example. The bad week takes its hit, the rework hours earn their own score, and the two-week result is 72.7%.

TEEP is the calendar-time cousin of OEE. If you need it, see OEE vs TEEP.

What counts as a good unit

  • Meets spec at the point you count it. Count at the end of the step or at transfer, and keep that point fixed.
  • Not on hold. Product waiting for a test is not good yet.
  • Found bad later? It goes into waste, with the root cause written down. Do not quietly remove it from last week.

If the area passes everything on without checking, Hansen says to adjust the count with your historical quality level. It is better than assuming 100%.

Where bad units come from

A quality number says how much was lost. It does not say why. Group each reject by cause as soon as you count it. The usual groups are:

  • Start-up and changeover. Units made while the machine settles. These often repeat every run, so they add up.
  • Process drift. Set points move during the run and the units slip out of spec.
  • Material. The raw material or the packaging is out of spec.
  • Handling. Damage after the machine, on the conveyor or in the pack-off.

Hansen asks for structural waste not to be hidden: the waste that every normal run makes still counts as a loss, even if "everyone knows" about it. Our advice is to track the amount per run, because a number that shows up every run is a design problem, not bad luck.

Also remember that OEE counts time, not money. A scrapped unit costs its material and its labour, and OEE only shows the time. Report the cost of poor quality next to OEE when you argue for a fix. A small quality loss can still be an expensive one.

Quick sense check before you publish a quality figure: good units can never be more than total units, so quality can never be more than 100%. And the units you count as good must match what you handed to the next step.

How to measure quality on the floor

You need two counts, taken at the same place for the same period: total units out and rejected units. The difference is good units.

  • Count rejects when they happen, at the reject bin, with a reason code. A guess at shift end is always low.
  • Separate scrap from rework, and keep the reason for each.
  • Reconcile weekly: units transferred to the next step should match your good count. Hansen advises comparing the transferred count with the sum of crew reports.

Method and tools are covered in OEE data collection methods. For scrap on its own, see how to calculate scrap rate and the scrap rate calculator.

Common mistakes

  • Counting reworked units as good. Quality then reads 100% while the line fixes parts all day.
  • Counting scrap in only one place. Rejects at a downstream station never reach the machine that made them.
  • Leaving out start-up scrap or changeover scrap.
  • Mixing units. Good in kilograms, total in pieces.
  • Ignoring returns. Product that comes back from the customer is waste with a root cause. Log it.

The wider list is in OEE calculation mistakes. For ways to cut scrap and rework once you can see them, read how to reduce scrap and rework. To see how quality fits the rest of the formula, go to how to calculate OEE.

See it live in Fabrico

Quality losses are easy to argue about and hard to see by shift. Fabrico shows OEE per line and shift, with the cause tagged by the operator at the line, so a bad run is traced to its shift and reason. See OEE per line and shift in Fabrico.

Book a Fabrico demo

What people ask next

Does rework count as good units?

Not on the first pass. A unit that needed rework did not come out right the first time, so it is not good for that run. If you rework it later, the extra time counts where it was done. The rules are in the rework section above.

Is OEE quality the same as first pass yield?

They are close, but not identical. OEE quality is good units ÷ total units on one machine or line, and it feeds OEE as lost time. First pass yield is a quality metric for one process step, and rolled throughput yield multiplies it across every step. See how to calculate first pass yield and rolled throughput yield.

Do start-up rejects count in quality?

Yes. Units made while the machine settles are scrap, so they lower quality. Many plants hide them by excluding the start. Our advice: count them, then work on them as a loss. See start-up losses.

How do I measure quality when I make liquid or rolls, not pieces?

Use any unit that fits the process: kilograms, metres, litres, square feet. Good quantity divided by total quantity gives the same quality rate. Keep the unit the same for the ideal rate.

What about units on hold or waiting for inspection?

Leave them out of good units until they pass. Hansen says to leave out product on hold or likely to be condemned. If it is later found no good, it moves to waste with the root cause noted.

Related guides: How to Calculate Production Yield by Process · First Time Through: Formula and FPY Difference · Throughput Time: Meaning, Formula and Example

← More OEE Lab guides · Open the OEE calculator →

OEE essentials: OEE calculator · What is OEE · Six big losses · Availability · TEEP · Cost of downtime · OEE benchmark · Free OEE tracker · Free OEE software · How to track downtime · How to implement OEE · Light MES vs full MES