Scrap Cost

Turning Scrap Cost into a Daily Management Number

Scrap reported in percent gets shrugged at; scrap reported in dollars per run gets fixed. Here is how to cost it fully and manage it every day.

Scrap reported as a percentage gets a shrug; scrap reported in dollars gets a meeting. A 3 percent scrap rate sounds like background noise until you cost it: 300 units scrapped from a 10,000 piece run at $18 of accumulated cost each is $5,400 per run. Run that job weekly and it is $280,000 a year, on one part number. Most plants running 2 to 4 percent scrap are quietly spending 1 to 3 percent of revenue on material and labor they throw in a bin, and the P&L buries it across six accounts where nobody owns it.

Cost scrap at the point of failure, not at raw material price. A part scrapped at final test carries every dollar spent on it: $6 of material, $7 of machining, $3 of plating, $2 of assembly, $18 total, even though purchasing sees a $6 part. This is also why defect location matters more than defect count: 100 parts scrapped at incoming cost $600, while 100 scrapped at final test cost $1,800. The Scrap Cost calculator builds this accumulated-value math for you; the management move is publishing dollars per run at the operation where the failure surfaced.

Add the two costs most plants forget. Rework: a unit reworked for 12 minutes at a $45 loaded labor and machine rate costs $9, and reworked units fail again at 2 to 5 times the base rate, so rework is often scrap on layaway. Lost capacity: when the line is sold out, a scrapped unit does not just cost $18, it costs the $14 of contribution margin the replacement unit displaced, because the machine hours to remake it came out of sellable production. At a constraint, true scrap cost per unit is accumulated cost plus lost margin, often double the accounting number.

Benchmark honestly. World-class discrete plants run under 0.5 percent scrap by value; good plants run 0.5 to 1.5 percent; 2 to 4 percent is common and expensive; above 5 percent means the process is not in control and no amount of inspection will save it. Measure in dollars and in percent of COGS, monthly, by line and by part family. A plant with $24 million COGS at 2.5 percent scrap by value has a $600,000 annual project sitting in plain sight, and cutting it in half is usually worth more than any single capital proposal on the list.

Attack it with Pareto discipline, because scrap is never evenly spread. Typically 3 to 5 defect modes on 5 to 10 part numbers drive 70 to 80 percent of the dollars. Rank by dollars, not by occurrence count: a frequent $0.40 defect matters less than a weekly $1,800 one. For each of the top modes, assign a named owner, a containment action within 24 hours, and a root-cause countermeasure within 30 days. Plants that work the top five modes to closure typically cut scrap dollars 30 to 50 percent in two quarters, then repeat with the next five.

The failure modes of scrap programs themselves: hiding scrap in rework so the metric looks clean while labor cost climbs; letting operators reclassify borderline parts to protect the number, which is why scrap dollars should reconcile to material variance monthly; and inspecting harder instead of fixing the process, which adds $30 to $60 per hour of inspection cost and still ships escapes, since visual inspection catches only 80 to 90 percent of defects on a good day.

Run the cadence. Daily: yesterday's scrap dollars by line on the tier board, top defect named, containment status stated in under two minutes. Weekly: Pareto review of dollars by defect mode, owner updates on the top five, and any new mode over $500 per week gets a slot. Monthly: reconcile scrap dollars to the ledger, retire closed modes, and reset the Pareto. World-class means scrap under 0.5 percent of COGS, every dollar traceable to a mode and an owner, first-day containment as a reflex, and a floor where operators call scrap in dollars because that is how the board reports it.

Published 2026-07-02.