Two plants can both report a 2% scrap rate and be running completely different quality operations. One of them scraps 2 parts out of every 100 and moves on. The other scraps 2 parts, reworks 15 more by hand, and still calls it a 2% problem, because scrap rate only ever counted the parts that got thrown away. That gap, the parts that needed a second pass but never got labeled a loss, is exactly what first pass yield is built to catch and scrap rate is built to miss.
Both numbers get quoted in the same breath at most plants, often by the same person in the same sentence, as if they’re two ways of saying the same thing. They aren’t. Knowing which one you’re looking at, and what it leaves out, changes what you go fix next.
What each one actually measures
Scrap rate is simple by design: parts scrapped divided by parts attempted. It answers one question cleanly: how much material did we throw away. It answers that question well. What it can’t tell you is anything about parts that weren’t good enough the first time but got fixed instead of thrown out. A part that needed a second drilling pass, a weld that needed a touch-up, a fitting that needed a manual deburr before it could ship, none of that shows up in scrap rate at all, because nothing got scrapped.
First pass yield answers a different question: what share of parts were good enough to ship on the first attempt, with no rework and no correction. A part that eventually shipped fine after somebody fixed it by hand still counts against first pass yield, because it wasn’t right the first time, even though it never touched the scrap bin.
What rework hides
Rework is the quiet cost that scrap rate structurally cannot see. A part that gets fixed by hand still shipped, still counted as production, and never touched a scrap number, but it consumed real labor, real machine time in some cases, and real schedule risk if the rework queue backs up behind the day’s regular work. A plant tracking scrap rate alone can look like it’s running a tight quality operation while actually burning a meaningful chunk of labor hours on fixes that never show up in any quality report anyone reviews.
The risk compounds when rework becomes normalized. Once “we’ll just touch it up” becomes the default response to a marginal part, the pressure to fix the actual process cause drops, because nothing downstream ever looks broken. Scrap rate stays flat, output looks fine, and the underlying process problem, a fixture slightly out of alignment, a parameter drifting, sits there indefinitely because nothing in the numbers is forcing anyone to look at it.
What each number leaves out on its own
Scrap rate leaves out rework, covered above, and it also leaves out material cost per part, which matters because scrapping one part on a machine running $40 of material per unit is a very different event from scrapping one part running $400 of material per unit, and a flat percentage treats them identically. Pricing scrap in dollars, beyond percent alone, is what corrects for that: a rate that resolves through the actual bill of materials for the product running, component cost times quantity, instead of a generic per-unit guess, so a scrap dollar figure means the same thing across every product a machine runs.
First pass yield leaves out something different: it doesn’t tell you what’s actually wrong. A yield number that drops from 96% to 89% tells you something got worse. It doesn’t say what. That’s what a defect Pareto is for, ranking defect types by how often they’re actually occurring, so a yield drop turns into a specific, named cause, a fixture issue, a material defect, an operator error category, instead of a vague number that’s just lower than it used to be.
A clean 100% first pass yield is also worth a second look before celebrating it. A shift with zero defects logged usually means one of two things: quality really was flawless that shift, or nobody’s logging defects at all. Distinguishing “100% verified,” meaning defects have been tracked and none showed up, from “no defects logged,” meaning nothing’s actually being recorded, keeps a suspiciously perfect number from being mistaken for a real one.
Using both numbers together
Neither number replaces the other, and a quality program built on just one of them has a blind spot the other would have caught. Scrap rate, priced in dollars against real material cost, tells you what’s being thrown away and what it costs. First pass yield tells you how much of your output needed a second attempt, rework included, and the defect Pareto behind it tells you why. Read together, a shop can see the full shape of its quality loss: what’s unrecoverable, what’s recoverable but expensive, and what specific cause is driving each one.
The practical habit worth building is checking both numbers side by side on the same cadence, not scrap rate weekly and first pass yield only when someone remembers to ask for it. A gap between a flat scrap rate and a declining first pass yield is itself a signal, it means rework is absorbing a growing problem instead of the problem showing up where everyone would notice it.
A note on where the labor actually goes
Plants that start tracking first pass yield seriously for the first time often discover their rework labor is larger than expected, because nobody had ever added it up before. That labor was always there, a tech spending twenty minutes touching up a weld, an operator running a part back through a second pass, it just lived inside “normal work” instead of a line item anyone reviewed. Once first pass yield makes that gap visible, the conversation shifts from “our scrap rate is fine” to “how much of our labor budget is going to fixing parts that should have been right the first time, without anyone tracking it,” which is a much more useful question to be asking, and usually a more uncomfortable one to answer the first time it gets asked out loud.
Quick recap
- Scrap rate counts only parts thrown away. First pass yield counts anything that wasn’t right the first time, rework included.
- A flat scrap rate can hide a growing rework problem, because rework never touches the scrap number at all.
- Price scrap in dollars against real material cost, not a flat percentage, so one scrapped part means the same thing across every product.
- A yield drop tells you something got worse. A defect Pareto tells you what, so the number turns into an actual cause.
- A suspiciously perfect 100% first pass yield often means nobody’s logging defects, not that nothing ever goes wrong.
- Track both numbers on the same cadence. A gap between them is itself worth investigating.