Production yield = good output / input, as a percentage, with both measured in the same unit. Discrete lines count units, batch and continuous processes usually weigh material, and for a batch you often divide by the theoretical output in place of the input.
Yield for discrete, batch, continuous and cutting processes, with the mass balance and a check against OEE quality.
The one formula, in four forms
Yield asks how much of what went in came out as something you can sell. The idea stays the same across a plant, but the thing you count changes with the process.
That is where most errors start. A weight-based yield on a mixing line and a piece-based yield on a press are both correct, but they cannot be compared.
| Process | Formula | Example | Yield |
|---|---|---|---|
| Discrete (units) | Good units / units started | 4,810 / 5,000 | 96.2% |
| Batch | Actual output / theoretical output | 940 kg / 1,000 kg | 94.0% |
| Continuous | Saleable output / material fed | 11,280 kg / 12,000 kg | 94.0% |
| Cutting, nesting | Weight of parts / weight of stock | 60 kg / 80 kg | 75.0% |
Production yield calculator
Pick the process type, then enter its counts or weights. The defaults are the four examples in this guide. Updates as you type.
OEE quality compares good units with total units made. Enter both counts for the same run to set it next to yield.
Where the input goes
Yield vs OEE quality
Yield compares output with material in, and OEE quality compares good units with total units made.
The working
Measure input and output over the same run or period, in one unit. Sum output and input across runs before you divide, so large runs carry their weight.
Need first-time-right or a multi-step route? Use the first pass yield calculator or rolled throughput yield.
Discrete processes: count units
Example: a stamping line starts 5,000 blanks and ends with 4,810 good parts. Yield is 4,810 / 5,000 = 96.2%, and the other 190 are scrap.
Decide whether reworked parts count as good. Final yield usually counts them and first pass yield does not, so report which one you mean.
Across several steps, a final figure hides what happens inside the route. Rolled throughput yield multiplies the yield of each step: 98.0% x 96.5% x 99.0% = 93.6%, and that guide takes the method further.
Batch processes: compare with theory
In a batch, the recipe tells you what you should get, so yield compares the real output with that figure. This is also called percent yield.
Set the theoretical figure on a clear basis. Moisture driven off in drying or baking, or a change in water content, can make a raw-input yield look like a loss when the process is doing what it should.
So many batch plants report on a dry basis, or set a standard yield for the known change. Then the number shows real losses, such as residue left in the vessel, spills and failed samples.
A yield over 100% is a warning sign in batch work. It often means extra water, wrong input weights or scales that need checking.
Continuous processes: run a mass balance
On a continuous line you cannot count units started, so weigh what you feed and what you sell over the same period. Everything that is left over must be explained.
Example: an extruder is fed 12,000 kg of resin in a week. It gives 11,280 kg of saleable product, 480 kg of edge trim and rejects sent to regrind, and 240 kg of start-up waste.
Regrind forces a choice. If the 480 kg goes back in, net material use is only 11,520 kg, so net yield is 11,280 / 11,520 = 97.9%, while the first-pass figure stays at 94.0%.
Report both. Net yield tells finance what material really cost, and first-pass yield tells the line how much it made twice.
Cutting and nesting: material yield
Example: a laser cuts 60 kg of parts from an 80 kg sheet, which is a 75.0% material yield. The 20 kg of skeleton is designed in, and no one on the shift can do much about it.
So keep two measures here. Material yield belongs to the nesting and design work, and the scrap rate covers parts cut wrongly, which is what the shift controls.
How to measure it on the floor
Pieces, kilograms or liters, for the input and output alike. Convert with a stated factor if needed.
Measure input and output over the same run or week, with work in progress accounted for.
Start-up, changeover, rejects, samples, spills. Weigh or count each one.
Input minus output minus known losses should come close to zero. Investigate what is left.
Measure material actually used, and treat material issued from stores with care. If part of an issue returns to stores or stays at the line, your yield will look worse than it was.
Yield vs OEE quality
Yield and the quality factor in OEE both deal with good output, but they have different bases. Yield compares output with material in, and OEE quality compares good units with total units made.
So the two can disagree on the same line. The laser above can run at 100% quality while its material yield stays at 75%, because the skeleton never counted as a unit.
Start-up waste shows the other way. It hits yield, and it also hits OEE, as start-up losses explains, but only if your count includes those units.
Common mistakes
- Mixing units. Pieces in and kilograms out will give a tidy number that means nothing, unless the conversion is stated.
- Counting rework as first-time good. It hides the second round of work, which first time through exposes.
- Ignoring work in progress. A big change in the buffer makes one week look great and the next look poor.
- Averaging percentages across runs. Sum the output and the input, then divide, so that large runs carry their weight.
- Trusting old scales. A small scale error can be larger than the loss you are chasing.
What to do with the result
Rank losses in money, since a kilogram lost at the end of the line carries more cost than a kilogram lost at the start. The scrap rate calculator and cost of poor quality help you to do that.
Then start with the largest loss you can influence. Often that means start-ups and changeovers, and how to reduce scrap and rework has the next steps.
Yield shows what the material lost, and OEE shows what the line's time lost. For the time side, Fabrico, a light MES for OEE, takes stops, speed and counts live from the line and shows OEE per line and shift. See how Fabrico tracks line output.
What people ask next
How do you calculate production yield?
Divide good output by input, in the same unit. On a discrete line, 4,810 good parts from 5,000 started is 96.2%. Batch and continuous processes usually use weight.
What is the difference between yield and first pass yield?
Yield usually counts reworked units as good if they ship. First pass yield counts only units that were right the first time, so it is equal to or lower than final yield.
How do you calculate batch yield?
Divide the actual output by the theoretical output from the recipe. A batch that should give 1,000 kg and gives 940 kg has a 94.0% yield.
Can yield be more than 100%?
In a sound measurement, usually not. A figure over 100% often points to added water, wrong input weights, a scale problem or a period mismatch, so check the measurement first.
How does yield relate to rolled throughput yield?
Yield often looks at one step or at the whole route from start to finish. Rolled throughput yield multiplies the first-time yield of every step, so it also shows the rework that final yield misses.
Is yield the same as OEE quality?
No. Yield compares output with material in, and OEE quality compares good units with total units made. Designed-in material loss, such as a cutting skeleton, affects only yield.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.
Related guides: OEE for Process Manufacturing: A Plant Guide