Scrap, Rework and Yield Accounting: Getting the Real Cost of Good Output

No production process delivers 100% good output. Material is wasted, units are spoiled, some pieces need rework. The accounting question is not how to eliminate loss (that is operations) but how to record it correctly so your per-unit cost reflects reality. Get this wrong and the cost of every good unit is misstated. Here is the working approach.

Normal loss vs abnormal loss

The first and most important distinction. Every process has an inherent, predictable level of loss — the normal loss. Anything beyond that is abnormal.

• Normal loss is the expected loss in a process — built into the design. A cutting operation that loses 3% as off-cuts; an evaporation step that loses 2% as vapour. It is the cost of doing business.
• Abnormal loss is loss above the normal — a power failure ruined a batch; an operator error scrapped a stack of pieces. It is not the cost of doing business — it is an exception.

The accounting treatment of each is different.

How normal loss is treated

Normal loss is absorbed into the cost of good units. There is no separate expense. If you start with ₹1,00,000 of material and 3% is normal scrap with no salvage value, the ₹3,000 of "lost" material is added to the remaining good output's cost.

So the cost per good unit goes up — and rightly so. Customers ultimately pay for the inherent yield of the process.

If the scrap has salvage value (sold as scrap metal, for example), reduce normal loss cost by the salvage realisation. The remaining goes into good units' cost.

How abnormal loss is treated

Abnormal loss is a separate expense, hitting the P&L directly. It does not add to the cost of good units, because it should not have happened.

This treatment matters because it isolates the cost of mistakes / disruptions in a visible line on the P&L. Without separation, abnormal losses are silently buried in product cost — and operations cannot see the trend.

How to set the normal loss percentage

This is judgement. Approaches that work:

• Engineering / spec. The supplier or process designer estimates an expected yield.
• Historical average. Last 6–12 months of actual yield, with the worst outliers removed.
• Industry benchmark. Some industries have published norms.

Revisit the normal loss % once a year. If actual loss has consistently come in below your normal-loss standard, the standard is loose (over-absorbing into good units, understating cost). If consistently above, the standard is tight (chronic "abnormal" losses that are not really abnormal).

Rework — same material, more cost

Rework happens when a defective unit is fixed instead of scrapped. The unit ends up as good, but it has cost more — extra labour, possibly extra material, sometimes machine time.

Accounting treatment depends on whether rework is normal or abnormal:

• Normal rework — part of the process, expected level. Costs absorbed into good units (same logic as normal loss).
• Abnormal rework — one-off problem batch, supplier defect, etc. Costs hit a separate "rework expense" line on the P&L.

The discipline is the same as for scrap: have a normal-rework percentage in your standard, and treat anything beyond as abnormal.

Yield as a continuous metric

Beyond the loss-vs-good accounting, yield percentage is one of the most useful operational metrics. It is simply:

Yield % = Good units / Total units started

Tracking yield over time tells you:

• Whether your process is improving, stable, or drifting
• Which products / batches / shifts / operators run high yield vs low
• When something changed (new supplier, new operator, machine issue)

Yield is a far more sensitive indicator than total scrap rupees — it picks up small drifts before they become big losses.

The full cost picture

Putting it together for a batch:

• Input cost — total material, labour and overhead started
• − Salvage value of saleable scrap
• − Abnormal loss cost (charged separately to P&L)
• − Abnormal rework cost (charged separately to P&L)
• = Cost absorbed into good units
• ÷ Good units produced
• = Cost per good unit

That denominator is good units, not started units — a key point. The cost per good unit goes up if yield drops, even with the same input cost.

How Booksmor helps

Booksmor's production module records material issued, good units received, scrap, and rework on every work order — and splits cost between good units absorbed, normal loss absorbed, and abnormal loss recognised separately on the P&L based on the normal-loss percentage you set per item. Yield trends per product, per batch, per period are visible at a glance. Start a 30-day free trial and let your costing reflect reality.