Lean Metrics and Measurement

Yield

How much of what you start finishes as something good.

Updated
·
4
min read
Definition

What is Yield?

Manufacturing yield is the share of input that becomes good output from a process, expressed as a percentage. A line that consumes 100 units of material and produces 94 good parts has a 94 percent yield. Yield is the umbrella metric covering first-pass yield, rolled throughput yield, and final yield, each of which measures a different cut of the same underlying idea: how much of the work survived as something the customer will accept.

Yield is the most important quality concept on a manufacturing shop floor and the most loosely defined. People use the word to mean different things, sometimes within the same conversation. Final yield, first-pass yield, rolled throughput yield, and quality rate all live under the umbrella, and the differences matter. A shop that reports "yield" without specifying which cut is reporting a number that nobody can act on confidently. The lean answer is to know which cut you mean and to use the breakdown to drive improvement.

"Yield is not one number. It is a family of numbers, and the family argues."

How yield works

The simplest yield calculation divides good output by units started. The arithmetic is always the same. The disagreements live in what counts as "good" and what counts as "started." Different definitions answer different questions.

The yield family

The most useful members of the family for a small shop:

  • Final yield. Good final parts divided by units started, regardless of rework. The customer-facing view. The flattering view.
  • First-pass yield. Units that completed each step on the first attempt, divided by units that started. The single-process honest view, before any rework.
  • Rolled throughput yield. The probability a unit passes every step in the stream defect-free. The product of every step's first-pass yield. The value-stream honest view.
  • Quality rate. A specific OEE component focused on usable output of a single machine. Closely tied to first-pass yield but framed for equipment.

Each cut answers a different question. Final yield answers "what shipped." First-pass yield answers "what worked the first time." RTY answers "what made it through clean." Quality rate answers "how often does this machine produce good parts." A serious lean shop tracks more than one.

Where yield fits on the shop floor

Imagine a 35-person machine shop running CNC parts for two industrial customers. Final yield has been 97 percent for years and everyone considers quality strong. The shop is preparing to qualify a new high-volume contract that requires per-process first-pass yield reporting. The team installs at-operation tracking for the first time.

The picture is more complicated than the 97 percent number suggested. First-pass yield at the main mill is 89 percent. First-pass yield at the lathe is 92 percent. A small rework cell sees about 8 percent of all production. The 97 percent final yield is real, but it is delivered through a rework loop that consumes about 200 labor-hours a month and adds two to three days of lead time on any reworked part. None of that was visible in the final yield number.

The improvement project focuses on the mill. A fixture redesign and a tool-change-frequency review lift mill FPY from 89 to 95 over four months. Rework volume drops by almost two-thirds. Final yield ticks up to 98.5, which sounds modest. The real win is in the labor freed and the lead time recovered, neither of which the final yield number had been measuring. Yield improvement work pays back in places the headline number cannot see.

Common mistakes with yield

  • Using "yield" without specifying which cut. Final, first-pass, and rolled are different numbers. Be specific or be ignored.
  • Counting rework as yield-neutral. Reworked parts ship, but the labor and time costs are real. First-pass yield exists to make this cost visible.
  • Reporting only the headline. A shop-wide yield number is too aggregated to act on. The breakdown by product family and by process step is where decisions get made.
  • Treating yield as a target without changing process. Yield is a result of process control. Pushing for the number without fixing the underlying causes produces short-lived gains and longer-term gaming.
  • Skipping the trend. A single yield snapshot can mislead. The trend over several months is the signal that tells you whether the operating system is getting better or quietly drifting.

Yield and related Lean tools

Yield is the umbrella term. First-pass yield is the single-process cut; rolled throughput yield is the value-stream cut that compounds FPY across every step. The opposites of yield are tracked through scrap rate, the share of unrecoverable output, and rework rate, the share that needs correction. Together these metrics form the operational view of how cleanly a process actually runs.

Common questions

The questions we hear most about this term.

How does yield work as a measurement?
You pick what counts as input, what counts as good output, and the window. A simple yield calculation takes good final parts divided by units started, multiplied by 100. The discipline is in being consistent: parts that started but were scrapped before any process step still count as input; parts that needed rework count as good output only when they finally pass. Most shops report yield at end of line, then break it down by process step using first-pass yield to find where the losses concentrate.
How is yield different from first-pass yield?
Yield is the umbrella; first-pass yield is one specific cut. Plain yield usually means final yield: of everything that started, how much shipped as good. First-pass yield means a stricter version: how much made it through without any rework. A line with a healthy rework loop can show good final yield and poor first-pass yield at the same time. Both numbers matter. Final yield is what the customer feels in volume; first-pass yield is what the operator feels in process control.
What are common mistakes with yield?
The biggest is calling final yield by the unqualified word "yield" and skipping the first-pass and rolled throughput cuts. The final number can be flattering while the per-step and compounded numbers reveal real problems. The second is counting rework as no yield loss. The labor and time spent on rework are real costs even when the part eventually ships. The third is reporting yield across products with very different complexity without normalizing, which lets a simple product's clean yield mask trouble on the complex product.
When should I worry about yield?
Worry when yield is drifting down on a stable product. Process drift is the usual culprit and waiting will not fix it. Worry when final yield looks healthy but first-pass yield is poor: the rework loop is doing real damage to lead time and cost that the final number cannot see. Worry when one product family is dragging the shop average down hard, because focused improvement on that family will pay off more than across-the-board work. Do not over-worry about absolute yield in isolation; the trend and the breakdown matter more.
What does good yield tracking look like?
A final yield number tracked daily by line, broken down by product family weekly. First-pass yield tracked at each cell or operation, hand-updated on a small board near the work. A monthly review pulling everything together into a value-stream rolled throughput yield. A clear top-three list of defect modes contributing to the gap between actual and target. Yield improvements come from process changes; the metric exists to point at where the next change should go.

Ditch the whiteboards and spreadsheets.

Same-day setup. No distributor lock-in. Zero stockouts. Top teams double revenue in 9 months.