The formula is simple. Reliable results depend on a fixed counting rule, a clear process boundary, and a consistent time period. This guide works through those choices before showing a calculation.
Set the process and boundary first
Choose one process step or a defined sequence, such as final assembly, sewing line inspection, or a packing check. State where the measurement starts and ends. Do not combine results from different operations unless the metric is explicitly defined for the whole route.
Choose the cohort before counting. A practical choice is every unit assigned to the named step during one shift, day, or production order. Count a unit in the FPY denominator only after its first assessment is complete. Units that entered the step but have no first-attempt result by the cut-off are work in progress (WIP), not completed attempts; report them separately and include them in a later cohort or period only under a consistent rule. Do not silently change the boundary between periods.
Count each physical unit once. If the process checks a unit at several stations, decide whether FPY means passing the first attempt at the named step or passing every step in a fixed sequence. Record the denominator and boundary alongside the result.
The FPY formula
FPY = units that pass the defined process on the first attempt ÷ total units in the defined cohort with a completed first attempt × 100%. The denominator is units, not individual defects. A unit with three defects remains one unit that failed first pass for a unit-based FPY calculation. Once a unit has a failed first result, later correction or retesting cannot change that first-pass result.
Worked example: 500 units complete their first assessment at a defined assembly step during the day. Of those, 480 pass and 20 fail. The first-pass result is 480 ÷ 500 × 100% = 96%. Fifteen of the 20 failed units are later reworked and pass a subsequent check; five remain unresolved at the cut-off. All 20 remain first-pass failures. The five unresolved units have a known first-attempt result, so they stay in the denominator. A unit still waiting for its first assessment would be WIP and would not enter this denominator yet.
See this workflow in KaizenQ
Log findings on the floor and share the report from one record.
Request a demoKeep FPY separate from final yield
Final yield answers a different question: how many units ultimately meet the release requirement after allowed rework or repair? In this example, 495 units have passed by the cut-off, so the observed final yield at that time is 495 ÷ 500 = 99%. Five units remain unresolved. Report the cut-off and open quantity with that figure.
Do not replace FPY with final yield when reviewing first-time process performance. Rework can raise the final number while hiding the amount of correction needed. A useful report shows first-pass units, units reworked, scrapped or rejected units, WIP with no first-attempt result, unresolved failed units, FPY, final yield at a stated time, and the exact cohort.
Units, defects, and repeat failures
Track defects separately from products. If one unit has two cosmetic issues and one functional issue, it may have three recorded defects while counting as one failed unit for unit-based FPY. Keep defect counts by type for root-cause analysis; the FPY denominator counts products.
Count an ordinary planned sequence of required checks as one first assessment when the defined process requires those checks before recording a result. If a unit fails and needs correction or a repeat test to pass, record a first-pass failure even when the correction happens before the unit leaves the station. Count that unit once in the cohort denominator and keep it out of the first-pass numerator. Retain each assessment and correction in the process record.
Make the number useful
Compare like with like: same process, entry criteria, unit definition, cohort boundary, and time cut-off. Show the numerator and denominator with the percentage. A trend without those details can hide changes in product mix or unfinished work.
Pair the rate with the count and defect categories. Review repeat issues, station, shift, product, and relevant specification changes. FPY shows where first-time performance changes; it does not prove the cause. Verify suspected causes with production records and measured evidence before changing a work instruction. Record the assessed cohort, numerator, denominator, time boundary, and results from the defined production record.
Key takeaways
- Define the process, cohort, and time boundary before counting.
- Count each unit with a completed first attempt in the denominator. Count only units that pass that attempt in the numerator; later rework or retesting does not change the first-pass result.
- Count a unit once even when it has several defects.
- Keep reworked or retested units in the denominator after their first attempt is assessed, but out of the first-pass numerator. Show final yield separately with its cut-off.
- Report counts and rules with the percentage so the result can be repeated.