First-Pass Yield: The One Metric That Exposes Hidden Rework
Most organizations can tell you how much they produced. Far fewer can tell you how much they produced right the first time. That second number — first-pass yield — is one of the most honest metrics in Lean Six Sigma, because it refuses to hide rework. A factory shipping 1,000 good units a month looks healthy until you learn it had to make 1,300 to get there. The 300 it scrapped or reworked are pure waste, and first-pass yield is the metric that drags that waste into the light.
The definition is simple: FPY is the proportion of units that pass through a step correctly the first time, with no rework, scrap, or correction. If a step starts 100 units and 92 come out right without intervention, its FPY is 92 percent. The discipline is in measuring it honestly — which is exactly where teams trip.
What bad measurement looks like
Bad first-pass yield numbers are almost always too high, and the inflation comes from counting the wrong thing.
Counting final inspection only — a unit reworked three times still 'passes' at the end, so the metric reports near-perfect quality.
Confusing FPY with throughput yield, ignoring units that were quietly fixed mid-process.
Excluding rework that happens 'off the books' because an operator corrected it without logging it.
Reporting a single plant-wide number that averages a strong step together with a failing one, hiding the real problem.
Measuring yield in a remote or distributed operation by self-report, where nobody wants to flag their own defects.
The last point hit hard through the pandemic. With supervisors off-site and supply problems forcing improvised substitutions, a lot of correction happened informally — and a yield number built on self-reporting quietly drifted away from reality.
What good measurement looks like
Good first-pass yield measurement is specific, step-by-step, and unforgiving of hidden fixes. A few practices separate the credible numbers from the comforting ones.
Measure at each step, not just the end. Calculate FPY for every meaningful process step, then multiply them for rolled throughput yield. A five-step process at 95 percent each is only about 77 percent end to end — a number a final-inspection metric would never reveal.
Count any intervention as a miss. If a unit needed touching up, adjusting, or re-running, it did not pass first time. Rework counts against yield even when the customer never sees the defect.
Make the count automatic where you can. Self-reported defects are systematically underreported. Sensors, scans, or system gates that record outcomes without human judgment give you a number you can act on.
Separate yield by step and shift before averaging. A blended figure hides the weak link. The point of FPY is to find the step bleeding the most rework, not to produce a single reassuring headline.
Tie the gap to cost. Translate the units lost between input and first-pass output into dollars of scrap, labour, and delay. That is what turns a quality metric into a business case.
Done this way, first-pass yield becomes a map. A clean overall number that hides a 70-percent step is worse than useless — it tells leadership everything is fine while one workstation quietly absorbs the cost. The step-level view points the improvement team straight at the problem.
Where to start
Pick one process that matters and chart its yield step by step for a few weeks. Count every correction as a miss, however small. The first honest number is usually uncomfortable — and that discomfort is the whole value. It tells you, finally, where the rework is hiding and what it is costing you.
If your quality numbers look fine but rework keeps eating your margins, XNM's strategic advisory can help you measure what is really happening and act on it.