To calculate OEE for a packaging line, take one shift and multiply three ratios: availability (run time ÷ planned production time), performance (packs made ÷ packs the line should have made at its ideal speed during that run time) and quality (good packs ÷ packs made). You need six numbers, and a shift log plus the end-of-line counter already hold all of them: shift length, scheduled breaks, recorded stoppage minutes, the ideal speed for the format, total packs and rejected packs.
OEE for a packaging line is availability multiplied by performance multiplied by quality, each expressed as a decimal, with the product read as a percentage.
OEE = Availability × Performance × Quality
For a line rather than a single machine, take every count and every stoppage at the line’s output — after the last inspection station, the checkweigher or metal detector, where a pack either goes on to the caser or into the reject bin.
Which stops count follows from the same rule. A jam upstream that the buffer absorbs and never stops the output is not a loss at all; one that starves the output is a line loss.
OEE is one of the KPIs defined in ISO 22400-2, and the three-factor form above is the standard one, so a figure calculated this way can be set against a supplier’s acceptance-test figure or another plant’s number.
Calculate availability, performance and quality as three separate ratios before multiplying them, because each one isolates a different loss, and the split is what tells you which loss to work on.
| Factor | Formula | What it measures |
|---|---|---|
| Availability | Run time ÷ Planned production time | Time lost to recorded stoppages |
| Performance | Total packs ÷ (Run time × Ideal packs per minute) | Output lost through reduced speed and short stops |
| Quality | Good packs ÷ Total packs | Packs that meet requirements on the first pass |
Two conventions decide where a loss lands. Changeovers count against availability even though they are planned: the line was scheduled to produce during those minutes and did not, so changeover time goes into recorded stoppages, not into breaks.
Short stops — a film splice, a cleared jam, a reject-gate fault — rarely reach the downtime log, so they land in performance. That is why performance divides by run time and not by planned production time: dividing by planned time would charge the logged stoppages twice, once in availability and again in performance.
The ideal speed is the fastest speed the line holds for this pack format when nothing is wrong, normally the figure signed off at acceptance for that format. The rated speed on the spec sheet usually belongs to a smaller bag or a lighter product, and using it makes performance look worse than it is on every other format.
The example is one 480-minute shift on a line running a single pack format, with the six inputs below taken from the shift log and the end-of-line counter.
| Input | Example value |
|---|---|
| Shift duration | 480 minutes |
| Scheduled breaks when the line is not expected to run | 30 minutes |
| Recorded stoppages, including changeovers | 50 minutes |
| Ideal line speed for this pack format | 100 packs/minute |
| Total packs produced, including defective packs | 36,000 |
| Defective or rework-required packs | 720 |
Availability is 88.9%: the line ran for 400 of the 450 minutes it was planned to run.
Planned production time = 480 − 30 = 450 minutes
Run time = 450 − 50 = 400 minutes
Availability = 400 ÷ 450 = 0.889 (88.9%)
Performance is 90.0%: the line produced 36,000 packs in a run time where 100 packs per minute would have given 40,000.
Ideal output in run time = 400 × 100 = 40,000 packs
Performance = 36,000 ÷ 40,000 = 0.900 (90.0%)
Quality is 98.0%: 35,280 of the 36,000 packs passed on the first pass.
Good packs = 36,000 − 720 = 35,280 packs
Quality = 35,280 ÷ 36,000 = 0.980 (98.0%)
The 720 rejects include packs that will be reworked. A pack that has to be opened and repacked failed its first pass, and first-pass yield is what quality measures.
OEE is 78.4%: the line delivered 78.4% of the good packs it could have made in the 450 planned minutes.
OEE = 0.889 × 0.900 × 0.980 = 0.784 (78.4%)
Against the commonly quoted world-class level of about 85%, this shift is below it, and the factor split shows where: availability lost 11.1 points and performance 10 points, while quality lost only 2.
The direct formula gives the same 78.4% in one step: good packs divided by what the line could have made at ideal speed across the whole planned production time.
OEE = Good packs ÷ (Planned production time × Ideal packs per minute)
OEE = 35,280 ÷ (450 × 100) = 35,280 ÷ 45,000 = 0.784 (78.4%)
It matches because two quantities cancel when the three factors are multiplied: run time is availability’s numerator and performance’s denominator, and total packs is performance’s numerator and quality’s denominator. If your two results disagree, the inputs are inconsistent — most often total packs and good packs counted at different points on the line, or breaks subtracted a second time when working out run time.
When two or more formats with different ideal speeds share a shift, calculate performance from the ideal time each product’s packs should have taken, summed and divided by run time — not from one blended speed.
Performance = (Product A packs ÷ Product A ideal speed + Product B packs ÷ Product B ideal speed + …) ÷ Run time
Availability and quality do not change: stoppage minutes and reject counts do not depend on which format was running. Only the performance denominator changes.
Suppose — hypothetical figures, on the same 400 minutes of run time — that 24,000 of the 36,000 packs were product A at 100 packs per minute and 12,000 were product B at 80 packs per minute. Product A should have taken 240 minutes and product B 150 minutes, 390 minutes of ideal time in total.
| Method | Calculation | Performance |
|---|---|---|
| Ideal time per product | (24,000 ÷ 100 + 12,000 ÷ 80) ÷ 400 = 390 ÷ 400 | 97.5% |
| One flat speed (100 packs/minute) | 36,000 ÷ (400 × 100) | 90.0% |
| Count-weighted average speed (93.33 packs/minute) | 36,000 ÷ (400 × 93.33) | 96.4% |
The flat speed charges product B for 20 packs a minute it can never make, so the line looks 7.5 points worse than it ran. The weighted average is closer but still wrong, because an average of speeds is not the same as a sum of times: the slower product takes a larger share of the minutes than of the packs.
Treat every bag size as its own format even when the product inside is the same. On a VFFS or pre-made bag machine the ideal speed changes with bag length and fill weight, and a shift log that records only the product name cannot be split by speed afterwards — so record the pack count at each format change, not just at the end of the shift.
