Quality and Inspection KPIs: Benchmarks for FPY, RTY, DPMO, Sigma and Gage R&R
First pass yield, rolled throughput yield, DPMO, sigma level, warranty returns, cost of poor quality and inspection effectiveness: how to measure each KPI and which figures have a cited source.
Use these benchmarks as reference points for spotting a gap, not as targets to hit. Your own weekly trend, measured with a fixed definition, says more about your process than any published range.
Every figure shows its basis. A superscript number links to a public source in the Sources list, and the sentence gives that source's scope and date. A figure marked no cited source is kept exactly as originally published because it has not been matched to a public source that states it, so treat it as illustrative.
The station first pass yield row uses overall equipment effectiveness (OEE) quality rates: the average across machines in 94 Swedish manufacturing companies, from data collected 2006 to 2012, and the world-class threshold in Nakajima's framework1, 2. Other ranges are as originally published, with the original's wording kept in the cell. Not stated means the original gave no figure for that column.
Method: How the figures in this guide are sourced
A figure with a superscript number is stated by the cited source, listed under Sources with its publisher, date and access date; sigma conversions are read from the NIST standard normal table. A figure marked no cited source is kept exactly as originally published because no public source stating it was found. It is illustrative, not measured industry data. Worked examples are arithmetic, not benchmarks.
First pass yield
First pass yield (FPY) is the anchor KPI because it exposes hidden rework that plain yield conceals. The OEE quality rate is a close station-level measure: Vorne describes it as similar to first pass yield, counting good parts as those that pass the first time without rework4.
Across machines in 94 Swedish manufacturing companies, with OEE data collected from 2006 to 2012, the average quality rate was 96.9 percent1. That is an average of company-reported machine data, so read it as a station figure, not a routing yield.
The world-class quality rate in Nakajima's total productive maintenance framework, as tabulated in a 2022 tool and die study, is above 99.9 percent2.
The gap between your reported yield and your FPY is your rework tax. Measure FPY at the station, not just at final, and track the trend weekly.
For example, a station showing 96 percent yield but 88 percent FPY is quietly reworking 8 percent of output.
The lever is error-proofing at the source, which typically lifts station FPY by 3 to 6 points within a quarter no cited source.
How to measure
Divide units that pass the station the first time, with no rework, by units entering it.
Measure at each station, not only at final inspection.
Rolled throughput yield (RTY) is the honest end-to-end number, and it punishes long processes. Typical multi-stage assembly lands at 70 to 85 percent RTY, while world-class operations exceed 95 percent no cited source.
For example, a 5-step line at 97 percent per step yields 0.97 to the fifth power, about 85.9 percent RTY. The same steps at 99 percent yield 95.1 percent.
The improvement lever is reducing step count and stabilizing the weakest station, because RTY is dominated by your worst FPY.
For example, cutting a redundant handling step, taking a line from 8 to 7 stages, can add several points of RTY with zero process change to the remaining steps.
Defects per million opportunities (DPMO) lets you benchmark across product complexity. Typical manufacturing sits around 3,000 to 25,000 DPMO, strong operations run 200 to 1,500 DPMO, and world-class holds under 100 DPMO no cited source.
Automotive and aerospace suppliers often contract to sub-50 DPMO on critical characteristics no cited source. Measure DPMO with a fixed, documented opportunity count so the metric stays comparable month to month.
The levers are mistake-proofing high-frequency defect modes and tightening incoming material control, since supplier defects commonly account for 30 to 50 percent of total DPMO in assembly plants no cited source.
How to measure
Fix and document the number of defect opportunities per unit.
Divide defects by units times opportunities, then multiply by one million.
Sigma level rolls DPMO into a single capability score and sets a clear ladder. Most plants operate between 3.0 and 4.0 sigma no cited source.
Sigma levels conventionally allow a 1.5 sigma long-term shift, so 3.0 to 4.0 sigma is the standard normal tail beyond 1.5 and 2.5: 66,807 down to 6,210 DPMO5. Moving from 3.0 to 4.0 sigma cuts defects roughly ten-fold5.
Best-in-class discrete manufacturing reaches 4.5 to 5.0 sigma no cited source, around 233 to 1,350 DPMO on the same convention5. The 6.0 sigma target of 3.4 DPMO is reserved for the most critical characteristics; the Environmental Protection Agency (EPA) describes a Six Sigma quality level as approximately 3.4 defects per million opportunities6.
Capability ties to defect rate the same way. For a centered normal process between two-sided limits, the NIST/SEMATECH e-Handbook gives about 0.27 percent rejects at a Cp of 1.00, 64 ppm at 1.33 and 0.6 ppm at 1.667. Those are short-term figures with no shift applied.
Track sigma monthly on a stable sample, and set a realistic annual goal of plus 0.3 to 0.5 sigma no cited source. Chasing a full sigma jump in one year usually signals a measurement problem, not a real gain.
How to measure
Convert DPMO to sigma with one stated convention, and say whether it includes the shift.
Use a stable sample of the same product mix each month.
What moves it
Every lever that lowers DPMO.
A trustworthy measurement system, so a jump reflects the process and not the gauge.
Warranty return rate
Warranty return rate is your field-truth KPI, and it validates everything upstream. Consumer durable goods typically run 1 to 3 percent annual return rate, and strong products hold 0.3 to 0.8 percent no cited source. Premium engineered products target under 0.2 percent no cited source.
For scale on the cost side, Warranty Week's compilation of U.S.-based manufacturers' filings puts the average 2025 warranty claims rate at 1.30 percent of product sales, with an average accrual rate of 1.43 percent8. That is a cost ratio, not a unit return rate.
Measure returns by ship cohort, not calendar month, so early-life failures are attributed correctly. A rising early-life return rate points to escaping defects and weak final inspection; a rising late-life rate points to design or wear-out.
How to measure
Count returns against the ship cohort they came from.
Split early-life failures from late-life failures.
What moves it
Closing the gap between internal DPMO and field DPMO, which exposes inspection blind spots.
Cost of poor quality (COPQ), tracked as a percent of revenue, is the executive-facing KPI that ties quality to money. Typical plants run 15 to 25 percent of revenue, and disciplined operations hold 5 to 8 percent no cited source. World-class organizations push below 5 percent no cited source.
The diagnostic ratio inside COPQ is prevention and appraisal spend versus failure spend. Struggling plants spend most of their quality budget catching and fixing defects; mature ones shift toward prevention.
Rebalancing toward prevention is the single most reliable lever for pulling COPQ down two to four points year over year no cited source.
How to measure
Sum internal and external failure costs with appraisal and prevention costs.
Divide the total by revenue for the same period.
Report prevention and appraisal spend against failure spend as a ratio.
Inspection effectiveness is the KPI people forget to measure, yet it gates every other number. Gage repeatability and reproducibility (gage R&R) should consume under 10 percent of tolerance to be acceptable, and under 30 percent to be marginally usable, per the Automotive Industry Action Group (AIAG) criteria as summarized by Minitab3.
Above 30 percent the AIAG treats the measurement system as not acceptable3, and in practice it rejects good parts and passes bad ones. The same guidance adds a resolution test: the study should split process output into 5 or more distinct categories9.
Escape rate, the share of defects that reach the customer despite inspection, should sit under 2 to 5 percent in a healthy operation no cited source. Improve both by qualifying gages before trusting the data, since a bad measurement system inflates apparent defect rates and hides real ones at the same time.
How to measure
Run a gage R&R study on each critical gauge before using its data.
Divide defects found by the customer by all defects found to get the escape rate.
What moves it
Qualifying and repairing gauges.
Inspection placed where defects are created.
Run your defect and escape counts with your own inputs. Defect Rate
How to build your own baseline
Set targets as a linked system, not isolated numbers, because these KPIs move together. Lifting station FPY through error-proofing raises RTY, drops DPMO, moves sigma up, cuts escapes and warranty returns, and pulls COPQ down.
For example, a plant at 88 percent FPY, 78 percent RTY, 12,000 DPMO, 3.7 sigma, 1.2 percent warranty return and 18 percent COPQ has a coherent improvement path. A twelve-month stretch goal of 94 percent FPY, 88 percent RTY, sub-4,000 DPMO, 4.2 sigma, 0.6 percent warranty and 11 percent COPQ is aggressive but achievable when the levers reinforce each other.
Write down each KPI's definition, including the DPMO opportunity count, and keep it fixed.
Qualify your gauges before you trust any baseline.
Collect several months of data before you set a target.
Compare against your own trend first and published ranges second.
Is my measurement system acceptable?. Minitab LLC (citing AIAG Measurement Systems Analysis Reference Manual, 4th ed., 2010). Section 'Gage R&R criteria', table 'Percentage of process variation'. Accessed September 30, 2026.
Lean Thinking and Methods - Six Sigma. U.S. Environmental Protection Agency. 'Introduction' section, second paragraph (page last updated 2026-05-27). Accessed September 30, 2026.