Key takeaways
See how first-pass yield feeds the OEE calculation.
First pass yield (FPY) is the percentage of units that move through a manufacturing step correctly on the first try, with no rework, repair, or scrap. The formula is FPY = (good units output minus reworked units) divided by units input. FPY exposes hidden rework that traditional final-yield numbers conceal, and it feeds the Quality factor of OEE.
Most plants quote a final yield number above 95% and still bleed margin on scrap, rework cells, and missed shipments. The reason is almost always the same. The reported figure counts reworked parts as good output, so the cost of getting it wrong the first time disappears into the labor variance. First pass yield is the metric that puts that cost back on the table.
First pass yield is the proportion of units that complete a process step in spec on the first attempt, without rework, touch-up, or rejection. It is sometimes called first time yield (FTY) or right first time. The defining characteristic is simple: a part that needed any human intervention to become sellable does not count as a "pass."
That distinction matters because rework hides waste. A welder who grinds a bad bead and re-runs it produces a good final part, but the time, consumables, inspection, and opportunity cost are real. FPY surfaces that loss. Final yield does not.
According to the standard first-pass yield definition, the metric is calculated as good units output minus reworked units, divided by units input. Strip out the rework correction and you have ordinary throughput yield, which can mask a process that is one operator vacation away from collapse.
The canonical FPY formula for a single process step is:
FPY = (Good Units Out - Reworked Units) / Units In
Express it as a percentage by multiplying by 100. If a CNC cell starts 200 housings, finishes 195 to spec, and 12 of those 195 required a second pass to remove burrs, FPY is (195 - 12) / 200 = 91.5%, not the 97.5% that final yield would report. The 6-point gap is the rework you are paying for.
Single-step FPY is necessary, not sufficient. Most products move through several operations, and yield losses compound. Rolled throughput yield (RTY) chains the FPY of each step:
RTY = FPY1 x FPY2 x ... x FPYn
The iSixSigma reference gives the classic illustration: a four-stage process where every step holds 85% FPY produces an RTY of 0.85 x 0.85 x 0.85 x 0.85 = 52.2%. Only 52 of every 100 raw inputs make it to the dock without a defect at some step, even though each station looks "fine" in isolation.
FPY is a station diagnostic. RTY is a system diagnostic. Confusing them is one of the most common reasons quality-improvement programs target the wrong cell.
| Aspect | First Pass Yield (FPY) | Rolled Throughput Yield (RTY) |
|---|---|---|
| Scope | Single process step | Entire value stream |
| Counts rework? | No, reworked units are subtracted | No, only first-time-good units propagate |
| Reveals | Which station is failing | True probability of a defect-free unit |
| Best used for | Operator coaching, station-level kaizen | End-to-end loss accounting, capacity planning |
| Typical blind spot | Cross-step interactions | Which specific step to fix first |
Consider a line that builds an electromechanical assembly across four stations. The table below uses the structure shown in the Wikipedia FPY reference, with input volumes, good output, and reworked units captured at each step.
| Step | Units In | Good Out | Reworked | FPY |
|---|---|---|---|---|
| A: Stamping | 100 | 90 | 5 | 85.0% |
| B: Welding | 90 | 80 | 0 | 88.9% |
| C: Assembly | 80 | 75 | 10 | 81.3% |
| D: Test & pack | 75 | 70 | 8 | 82.7% |
RTY = 0.850 x 0.889 x 0.813 x 0.827 = 0.508, or about 50.8%. A plant manager looking only at the final 70-of-100 finished-goods number sees a 70% yield. The honest figure, accounting for the rework loops at A, C, and D, is barely half that. The lowest-FPY station, Assembly at 81.3%, is the first place to send the improvement team.
FPY is not an isolated quality metric. It is the operational definition of the Quality factor inside OEE calculation. The reference site oee.com states plainly that "OEE Quality is similar to First Pass Yield, in that it defines Good Parts as parts that successfully pass through the manufacturing process the first time without needing any rework." Reworked parts are quality losses, full stop.
This makes FPY the lever that moves the Q in OEE. The Availability and Performance factors are addressed through downtime and cycle-time work (see our OEE guide and the Six Big Losses framework). Quality is moved by FPY work. Lift FPY from 92% to 97% and OEE rises by the same five points, no extra runtime required.
Low FPY at a station is often a maintenance signal, not just an operator one. Worn tooling, drifting spindle bearings, or a sensor that has lost calibration will quietly push first-pass scrap up before they cause a hard stop.
Tracking FPY alongside MTBF and MTTR and acting on the trend is more useful than waiting for the asset to fail outright. Asset criticality work should weight FPY impact, not just downtime risk.
Every reworked part is a margin leak that final-yield reporting will not show. Treat each FPY point lost as the sum of:
Put a unit cost against each of those lines and a one-point FPY drop at a high-volume station is almost always larger than the same point of availability loss, because it carries scrap and rework labor at the same time.
Improving FPY is process work, not posters. The order of operations below is the one that holds up across discrete and process plants.
You cannot improve what you do not separately measure. Stop reporting only final yield. Capture units in, good units out, and reworked units at each operation. If your MES rolls rework into "good," that is the first thing to fix.
Sort defects by frequency at the lowest-FPY step, then drill into top causes. FMEA is the formal vehicle. The point is to attack the cause, not the symptom.
Apply poka-yoke at the design and fixture level so the wrong move becomes impossible, not just discouraged. This is the heart of kaizen work at the cell.
A station that drifts out of spec is rarely an operator problem. Tie FPY trends to a preventive maintenance program and to condition monitoring on rotating assets.
When vision or sensors detect a defect, the next action should be a work order in the responsible technician's hand, with the right part number and checklist already attached. Without that loop, FPY data becomes a wall poster.
Fabrico is a unified System of Action for industrial manufacturing that combines real-time OEE/MES with a full CMMS. Two pieces of the platform are directly relevant to FPY work.
First, Fabrico connects to machine PLCs to capture cycle time and stop events, and uses computer vision to capture the true cause of downtime and quality events, including the micro-stops and quick rework loops that operators typically under-report. That gives you an FPY number you can defend, not one that depends on honest manual logging.
Second, when a defect signal is detected, Fabrico converts it into a prioritized, parts-ready digital work order on the responsible technician's phone, with a QR-enforced checklist on the asset. The fault-to-fix loop is what stops FPY data from going stale on a dashboard.
If you want to see the loop running on real line data, book a Fabrico demo and walk through your worst-FPY station with our team.
Three failure modes show up almost every time a plant moves from final yield to first pass yield.
Hourly at the station, daily at the tier-1 meeting, weekly at the area review, monthly at the plant level. FPY is most useful as a short-cycle signal. By the time it shows up in a monthly quality report, the cause is two shifts cold and the team cannot reconstruct what happened.
A live FPY chart at the cell, fed by sensors and vision rather than after-the-fact paperwork, is what separates plants that move the number from plants that just report it.
Treat FPY as the operational truth-teller it is. Pair it with RTY at the value-stream level, fold it into your OEE Quality factor, and wire every drop to a work order that lands in the right hands. That is how rework loops shrink and how the cost of poor quality comes out of the P&L.
FPY = (Good Units Out minus Reworked Units) divided by Units In, expressed as a percentage. The defining feature is that reworked units are not counted as good output, even if they eventually pass final inspection. This is what distinguishes FPY from ordinary throughput yield.
First pass yield (FPY) measures a single process step. Rolled throughput yield (RTY) multiplies the FPY of every step in the value stream to give the probability of a unit passing the entire process with no defects. Four steps at 85% FPY each produce an RTY of 52.2%, not 85%, which is why RTY exposes losses that station-level FPY can mask.
They are functionally equivalent. The Quality factor of OEE counts only parts that pass through the process the first time without rework, which is the same definition used for first pass yield. Improving FPY directly lifts the Q in OEE, with no extra runtime required.
There is no single industry-wide number because product complexity varies enormously. As a working rule, world-class discrete manufacturing operations target station-level FPY above 99% and value-stream RTY above 95%. The more useful target is your own trend: FPY should be rising shift over shift at the station you are working on, with the cause of every dip identified before the next tier meeting.
Reported or final yield typically counts any unit that ends up sellable, including units that were reworked one or more times. First pass yield subtracts the reworked units, so it is always equal to or lower than final yield. A large gap between the two is a direct signal that hidden rework loops are absorbing labor and capacity.
Every point of FPY loss represents rework labor, scrapped material, re-inspection time, and lost capacity. Tracking FPY at the station level, rather than only final yield, is the most direct way to put those costs on the table and assign owners to them.
Login
Blog
