ThingConnect

Quality metrics

Scrap rate

Scrap rate is the share of parts produced that had to be thrown away — scrapped or rejected quantity divided by total quantity produced, as a percentage.

Updated November 20262 min read

← All glossary terms

Scrap rate is the quality number most shops already track, because scrap is visible: it ends up in a bin. The catch is everything it doesn’t count.

The scrap rate formula

Scrap rate = Scrapped or rejected quantity ÷ Total quantity produced × 100%. Both counts should cover the same part and the same period — a shift, a batch or a week — and the rule for what counts as scrap should stay the same from run to run.

A worked example

One batch on one machine (illustrative numbers):

  • Total parts produced

    Value
    1,200
  • Scrapped

    Value
    18
  • Reworked and shipped

    Value
    42

Scrap rate = 18 ÷ 1,200 = 1.5%. Check your own batch in the scrap rate calculator.

Scrap rate vs rework vs FPY

The 42 reworked parts above don’t appear in the 1.5% at all. Counting them, first pass yield is (1,200 − 18 − 42) ÷ 1,200 = 95.0% — a much bigger quality problem than the scrap rate alone suggests.

That is why scrap rate works best next to FPY rather than on its own: scrap rate shows material thrown away, FPY shows how much time went into parts that needed a second attempt.

Frequently asked questions

Yes, if they are thrown away. Many shops record them separately as setup scrap, so a job’s running scrap rate isn’t hidden by first-off pieces and the setup loss stays visible on its own. Decide the rule once and apply it the same way on every run.

Start with the reasons: record why each rejected part failed, then fix the biggest cause first — often tool wear, offsets after a tool change, or first-off pieces. Tracking scrap by machine and by tool shows patterns that a single plant-wide percentage hides.

Related terms

Filed under

Quality metrics

Live demo

See ThingConnect on a real shop floor

A one-on-one walkthrough with our team.

Book a demo
Talk to our team

See total output per machine and shift, counted straight from the controller