Defect vs. Defective: What’s the Difference?

A defect is a single failure to meet a requirement. A defective is a unit that contains one or more defects. The distinction sounds pedantic and is anything but: an invoice with three errors carries three defects but counts as one defective invoice, and which of those numbers you track changes what your metrics can tell you.

Why the distinction matters

Counting defects measures intensity — how much is wrong. It feeds metrics like defects per unit and DPMO, and it rewards fixes that remove individual failure modes even when some units still fail. Counting defectives measures reach — how many units are wrong. It feeds first-pass yield and percent-defective figures, and it reflects what a customer actually experiences: a unit either arrived right or it did not.

The two can move independently. A change that eliminates the most common error might cut the defect count sharply while barely moving the defective rate, because most bad units also carried a second error. Teams that track only one of the two numbers regularly misread their own progress.

The distinction also decides which tools apply downstream. Defect counts feed defects-per-unit views and DPMO, and they pair with control charts built for counts. Defective counts feed proportion-based views such as first-pass yield and percent defective. Choosing the wrong frame early means rebuilding the measurement plan later, which is why the definition step in Measure settles this before any data is collected.

A worked mini-example

An illustrative furniture maker inspects 100 chairs. Twelve chairs have at least one problem, and across those twelve the inspectors log twenty findings — scratches, loose joints, missing floor glides. Defectives: 12 of 100. Defects: 20. A month later a fix for the scratching brings the defect count down to eleven, but ten chairs still fail inspection, because most scratched chairs also had a loose joint. The defect count applauds; the defective rate shrugs. Both are telling the truth about different questions.

• Define what counts as a defect, in writing, before anyone starts counting
• Decide the unit of analysis up front — per chair, per order, per invoice line
• Never blend defect counts and defective counts in a single metric
• Use defects per unit to prioritize failure modes; use defective rate to report customer impact
• Keep the definitions stable over time, or your trends become meaningless

Count defects to learn how much is wrong. Count defectives to learn how often.

This vocabulary is foundational — it appears in our free White Belt program and is assumed everywhere after. Yellow and Green Belt build on it directly, since DPMO, yield, and sigma level all depend on counting the right thing in the right way.