A finished batch can still leave one uncomfortable question: where did all the planned output go? Some units may be ready to sell, need correction, remain on hold, or be unusable. Material may also change during a normal process.
When every difference becomes one vague number, inventory looks cleaner than the work. A practical batch production yield loss review separates production rework, manufacturing scrap, expected loss, and unexplained variance so batch costing and saleable output stay honest.
One final yield percentage hides several decisions
Suppose a batch was expected to produce 500 units and only 472 become saleable. The 28-unit gap does not explain itself. Some units may be waiting for relabeling. Some may have been damaged. Some planned output may never have formed because actual fill or process yield differed from the standard.
Rework may become saleable after approved correction. Scrap should not return to available inventory. A measured process change may belong in the expected yield standard. An unexplained variance should remain visible until the team understands it. The purpose is not perfection; it is an explainable outcome.
Define five disposition categories before closing the batch
Use operational definitions that match the business’s actual products and quality process.
Released output
Released output has completed the required production, packaging, and quality steps and can enter usable finished-goods inventory. “Made” is not enough if a hold, cure period, label check, or final approval remains open.
Rework
Rework is output held for an approved correction that may return it to the required standard. Record what must be done, the quantity affected, who can approve the work, and the final result. Until that decision is complete, rework is work in progress—not saleable stock.
Scrap
Scrap is material or output that will not be recovered for its intended saleable use. Record the quantity, reason, disposition, and approval required by the business. Do not reduce inventory silently or hide scrap inside an adjustment called “miscellaneous.”
Expected process loss
Expected process loss is a measured change in a stable, defined process: equipment residue, controlled trim, evaporation, test samples, or another product-specific loss established from comparable runs. The amount still belongs in the batch record and costing. Review results outside the demonstrated range.
Unexplained variance
Unexplained variance is the remaining difference after released output, rework, scrap, and recorded process loss are counted in the same unit of measure. Keep it separate. Do not automatically rename it scrap or normal loss just to force the record to close.
Build a batch disposition bridge
Choose one common unit: kilograms, liters, pieces, or another meaningful measure. Do not mix cases, bottles, grams, and saleable units without documented conversions.
At the disposition point, use:
Unexplained variance = measured input basis − (released output + rework or hold + scrap + expected process loss)
Define the input basis as actual material issued, measurable output at a named checkpoint, or planned unit-equivalent output. Do not compare physical input with an unrelated sales forecast.
A small-batch production record should preserve the source quantities, formula or specification version, dates, checks, reasons, and approvals behind the bridge.
Worked example: reconcile a 100-kilogram batch
A business issues a measured 100 kilograms to one batch. At the first disposition review, the record shows:
- released output: 91 kilograms;
- held for approved rework: 5 kilograms;
- scrap: 2 kilograms;
- expected process residue: 1.5 kilograms.
The first bridge is:
100 − (91 + 5 + 2 + 1.5) = 0.5 kilograms unexplained
The half-kilogram variance still needs an owner and review date. It should not be buried inside expected residue.
The five kilograms of rework are then corrected and reviewed. Four kilograms pass and move into released output; one kilogram is scrapped. The final disposition becomes:
- released output: 95 kilograms;
- scrap: 3 kilograms;
- expected process loss: 1.5 kilograms;
- unexplained variance: 0.5 kilograms.
95 + 3 + 1.5 + 0.5 = 100 kilograms
Nothing is counted twice. The 95% final released yield now shows what was recovered, discarded, expected, and still unexplained.
This is a hypothetical bookkeeping example, not a universal acceptable-loss standard. Use the product’s actual measures, established process, and applicable requirements.
Connect disposition to inventory, costing, and quality
Released output should increase usable finished goods. Rework should remain in a controlled non-saleable state until the correction and review are complete. Scrap should leave usable inventory with a reason. Expected process loss and unexplained variance should affect actual yield and the cost spread across saleable output.
If the same batch cost is divided by optimistic planned quantity instead of 95 kilograms released, cost per saleable unit will be understated. The guide to calculating the full cost of rework helps add replacement materials, extra labor, outside costs, and genuinely lost contribution after the physical bridge is clear.
Quality records should also preserve why an item moved from hold to released or scrap. The simple quality checks every product business should document show how standards, timing, ownership, and failure responses can stay connected to production.
Review three comparable batches before changing the standard
One batch can expose a problem; three comparable batches can begin to show a pattern. Keep the product, formula or specification, measurement points, and units consistent. Then compare:
- first-pass released output;
- quantity sent to rework and recovered;
- scrap by reason;
- expected process loss against its demonstrated range;
- unexplained variance; and
- final saleable output.
If rework repeats, improve the earliest practical check. If scrap rises after a supplier or packaging change, investigate that connection. If the standard is unrealistic, update planning and costing deliberately. For growing unexplained variance, inspect timing, conversions, samples, spills, counts, and delayed entries.
For an industry-specific example, the soap yield guide separates cutting yield, curing change, rejected bars, and final saleable mass instead of treating them as one missing quantity.
Make the review useful, not punitive
A disposition bridge should improve the process, not blame operators. Record the event close to the work, use clear reason codes, and allow notes when categories do not explain the outcome. Review recurring causes by product and step, then ask where the difference became detectable. The goal is better evidence—not a suspiciously perfect zero produced by forced categories.
Frequently asked questions
Is all yield loss waste?
No. Some measured change may be normal for a defined process, while some is avoidable scrap or unexplained variance. Separate the categories before judging the result.
When should rework stay out of finished inventory?
Until the approved correction is complete and the required review confirms the output can be released. Physical presence does not make it saleable.
Should scrap and expected process loss use the same reason code?
No. Scrap is unrecovered output or material; expected process loss is a separately measured part of a stable process. Combining them hides whether the process stayed within its established range.
How often should the disposition bridge be reviewed?
Complete it for each meaningful batch, then compare comparable batches on a regular schedule. Review sooner when customer risk, safety, labeling, or a significant unexplained variance is involved.
Practical takeaway
Take one recently completed batch and divide its output into five categories: released, rework, scrap, expected process loss, and unexplained variance. Reconcile all five in one unit of measure, then follow rework through its final decision.
A trustworthy batch record does not make loss disappear. It shows exactly where the output went—and gives the team a specific place to improve next.
