Scrap Rate

Scrap rate is the most expensive metric on the shop floor because every unit it counts is fully consumed and fully wasted. The material, the labor hours, the machine time, the energy, the queue space, all of it is gone. A 2 percent scrap rate on a constraint operation is roughly equivalent to running the constraint 2 percent fewer hours, which the shop would never voluntarily do. Most shops underreport scrap because the recording habits are casual, and underreporting protects the comfortable illusion that quality is fine.

"Every scrapped part is a part you paid to make twice."

How scrap rate works

The calculation is straightforward: scrapped units divided by total produced units, multiplied by 100. The arithmetic is easy. The challenge is making the count honest.

Where scrap rates lie

Most underreported scrap comes from one of these patterns:

• Quiet binning. Operators toss a bad part without logging it because logging is inconvenient. The defect mode is invisible to the metric.
• Rework masquerading as scrap-free. A part that needed substantial rework before passing inspection is counted as good. The yield loss is real even though the unit eventually shipped.
• Mis-categorization. Parts scrapped due to a process problem get recorded as material defects to keep the operation's number clean.
• Aggregating across mix. A new product in ramp-up scrapping at 6 percent and an established product scrapping at 1 percent average to a misleading 3 percent that hides both stories.

The fix is structural. Visible scrap bins, a simple recording habit (a hash mark on a tally sheet at each cell), and a defect category on every entry. The whole system takes a minute per part. The honest data it produces is worth far more than the minute.

Where scrap rate fits on the shop floor

Imagine a 30-person plastics injection molding shop running parts for two consumer goods brands. The reported scrap rate has been a steady 2.5 percent for years and management considers it fine. A new operations lead asks for a week of careful tracking with bins at each press and a hash-mark tally. The honest number comes back at 4.8 percent.

The breakdown is more useful than the total. Two presses account for most of the gap. One has a worn gate producing intermittent flash that operators have been quietly trimming or rejecting without logging. The other has a temperature controller that drifts on long runs, producing short shots in the last hour of each shift. Neither problem is exotic. Both have been there long enough to feel normal.

The fixes are unglamorous. A new gate insert on press one. A controller replacement on press two. Bin-and-tally tracking made permanent. Six weeks later the scrap rate is 1.8 percent and the cost-of-poor-quality estimate drops by about $80,000 a year. The metric did its job once it was measured honestly. The improvement followed naturally from the data being visible.

Common mistakes with scrap rate

• Casual recording. A scrap rate that depends on operator-initiated paperwork will always look better than reality. Make recording physically easier than not recording.
• Averaging across product families. Different products have different baseline scrap. Mixing them obscures which product is the problem.
• Treating it as a financial line only. The cash cost of scrap is the smaller part of the loss. The constraint time spent producing the scrapped unit is the bigger cost, and it never appears on the income statement.
• No defect category on the entry. A count of scrap without modes is a number you cannot act on. Three categories per cell is usually enough to surface where the improvement work belongs.
• Treating it as the operator's problem. Scrap is almost always a process problem. Operators see it first; fixing it usually requires changing a tool, fixture, material, or upstream condition.

Scrap rate and related Lean tools

Scrap rate is one half of the yield story; the other half is rework rate, which covers recoverable defects. Both feed into first-pass yield and the broader category of defects the lean wastes call out. Scrap is the most direct input to cost of poor quality, since the material and labor are unrecoverable. Honest scrap data is foundational to any quality improvement program.