Yield Loss

A factory buys raw material, but not every unit becomes saleable product, because some material is trimmed away and some units fail inspection, and yield loss puts a number on that gap. Define the production boundary first, whether one machine step, a whole batch or an entire plant, since each gives a different result.

Keep units consistent, comparing kilograms of input with kilograms of good output, or units with units, because mixing material weight with finished item count needs a conversion. An illustrative yield-loss rate is input quantity minus good output quantity, divided by input quantity, times 100, so if 1,000 kilograms enter a process and 850 kilograms leave as good output, the measured loss is 15%.

That 150-kilogram gap needs explanation, because it may include expected trimming, unavoidable moisture change, rejects, spills or material still in process, so do not label the whole difference as scrap without tracing it. Oracle describes operation yield as usable output divided by input at a production step, and on the same measured boundary yield loss can be viewed as 100% minus that yield, a simple complement that assumes compatible quantities and classifications.

Separate expected and unexpected loss: a recipe may plan for normal evaporation or trimming, while a sudden increase above that standard suggests equipment, training or input-quality issues. Record scrap reasons, since a total loss percentage tells the size of the problem, not why it happened, and defect codes and shift notes can point to causes.

Track rework distinctly, because a defective unit that is successfully repaired may eventually become good output, and distinguish yield from first-pass yield, which asks how much output passed without rework at a particular step while final yield can improve after rework and extra labour and cost remain. Trace material flows by adding beginning work in process, current inputs, good output, scrap and ending work in process as the process requires, because ignoring inventory still inside a machine can exaggerate loss.

Compare batches on a fair basis, since product mix, recipe, raw-material grade and machine settings can change expected yield and a simple average across unrelated products may mislead managers. Check whether lost material has value, because offcuts can be reused or sold, although counting recovered material as fully saleable output can overstate yield, and monitor process steps separately, since a 15% overall loss could hide one stable step and one deteriorating step.

Look at equipment condition, since worn blades, poor calibration or unstable temperatures may increase loss, and compare measurements before and after maintenance rather than assuming causation. Check input quality, because moisture, size variation or supplier defects can change usable output even with the same machinery, and record lots so a loss spike can be traced.

Train on measurement, as scales, counters and inspection rules must be consistent across shifts, otherwise apparent improvement may be a recording change. Do not improve the percentage by relaxing the quality standard, because shipping borderline goods can reduce reported loss and increase customer complaints, so the definition of acceptable output must remain stable.

Use improvements that address causes, such as a new cutting pattern to reduce trim or a calibration fix to reduce rejects, testing one change at a time when possible and verifying that quality remains sound. For owners, yield loss links process performance to cost, and a trustworthy measure begins with a clear boundary, accurate flows and a stable definition of good output.