Your PPAP package is nearly complete. Element 10 contains your measurement system analysis report, and the %GRR came back at 24.3%. Your SQE is asking whether it passes. Your quality manager says it depends. The AIAG manual sits on your desk, open to Chapter 4, and you are trying to figure out which number actually matters here.
This is not an unusual situation. Gauge R&R results sit in a gray zone for more submissions than most quality engineers want to admit, and the AIAG MSA 4th Edition criteria have conditions attached that are easy to misread. This guide explains exactly what %GRR and NDC mean, how to apply the AIAG thresholds correctly, and what your options are when a measurement system does not pass.
What a Gauge R&R Study Actually Measures
Gauge R&R stands for Gauge Repeatability and Reproducibility. It is a type of Measurement System Analysis (MSA) designed to quantify how much of the observed variation in your measurements comes from the measurement system itself, rather than from actual part variation.
There are two components:
- Repeatability: Variation from the gauge when the same operator measures the same part multiple times under the same conditions. Also called Equipment Variation (EV). This captures gauge-to-gauge consistency, fixture stability, and environmental noise.
- Reproducibility: Variation from different operators measuring the same part with the same gauge. Also called Appraiser Variation (AV). This captures whether two operators get different readings from the same piece.
The AIAG standard MSA study uses at least 2 appraisers, 10 parts, and 2 replications per appraiser per part (minimum configuration). Most automotive customers require 3 appraisers and 3 replications, particularly for critical safety characteristics. The combined Gauge R&R (GRR) is the total measurement system variation from both sources.
The Two Variance Components You Need to Understand
Gauge R&R results report variation, not measurements. Two variance components matter for acceptance decisions:
Part-to-Part Variation (PV): The actual spread in your parts. If all 10 study parts came from the same production lot and look nearly identical, PV will be low, which inflates your %GRR. This is a study design problem, not a measurement system problem. Parts for a Gauge R&R study must span the full expected production range.
Total Variation (TV): The sum of GRR variance and PV variance. This is the denominator in the %GRR calculation.
The formulas:
%GRR = (GRR Std Dev / Total Std Dev) x 100
%GRR = sqrt(EV² + AV²) / sqrt(GRR² + PV²) x 100
This means %GRR is a ratio: how large is the measurement system variation relative to total observed variation. If part-to-part variation is wide (good study design), GRR looks small as a percentage. If all your study parts are nearly identical, GRR looks large as a percentage even if the gauge itself is fine.
%GRR: The Primary Acceptance Criterion
The AIAG MSA 4th Edition specifies three zones for %GRR:
| %GRR | AIAG Classification | Default Action |
|---|---|---|
| < 10% | Acceptable | Measurement system accepted for most applications |
| 10% to 30% | Conditional | May be acceptable depending on application, cost, and risk |
| > 30% | Not Acceptable | Measurement system should be rejected or improved |
What "Acceptable" Means at < 10%
A %GRR below 10% means the measurement system contributes less than 10% of the total observed variation. For most production measurement systems, this is a strong result. The gauge is adding minimal noise to your capability data and control charts. This zone passes for PPAP Element 10 without condition.
The Conditional Zone: 10% to 30%
This is where most interpretation errors happen. The AIAG manual does not call a 24% GRR a failure. It calls it conditional, meaning the customer and quality team must evaluate several factors before accepting or rejecting it:
- Importance of the characteristic: A 24% GRR on a non-critical aesthetic dimension is lower risk than a 24% GRR on a safety-critical interference fit. If the characteristic is on the Control Plan as a Critical Characteristic (CC) or Significant Characteristic (SC), customers will push harder for improvement.
- Cost of improving the measurement system: If reducing %GRR to below 10% requires a $40,000 CMM upgrade for a gauge measuring a $3 formed bracket, the customer may accept a 22% with containment measures. This is a documented business decision, not a blind pass.
- Process capability relative to tolerance: If your Ppk is 2.1 on a wide-tolerance dimension, a 24% GRR is not materially affecting your ability to make a good/bad decision. If your Ppk is 1.15 on a tight-tolerance dimension, a 24% GRR is consuming a significant slice of that tolerance band in measurement noise.
When you submit a PPAP with a conditional %GRR, attach a note explaining the evaluation. Sending a raw number with no context leaves the interpretation entirely to the SQE reviewer, which usually results in a rejection and a request for more information.
Above 30%: Not Acceptable
A %GRR above 30% means more than 30% of observed variation in the study came from the measurement system. This masks process behavior, corrupts capability indices, and makes control charts unreliable. The AIAG manual treats this as a rejection. You need to either improve the measurement system before PPAP submission or get explicit customer engineering disposition to proceed under containment.
Do not submit a 35% GRR without customer sign-off and a corrective action plan. It will be rejected.
Number of Distinct Categories (NDC)
The second acceptance criterion from the AIAG MSA 4th Edition is the Number of Distinct Categories (NDC). This measures how many statistically distinct groups the measurement system can separate within the part variation.
The formula:
NDC = 1.41 x (PV / GRR)
The AIAG minimum is NDC ≥ 5.
An NDC of 1 means the measurement system cannot reliably distinguish any parts from each other. An NDC of 2 means it can roughly separate parts into high and low. NDC ≥ 5 means the system has enough resolution to support meaningful process control decisions.
Why does this matter? %GRR can look good while NDC is low when part variation is very narrow. Conversely, %GRR can look marginal while NDC is high when part variation is wide. Both criteria must be evaluated together.
A measurement system with %GRR = 18% and NDC = 2 has a problem. A measurement system with %GRR = 25% and NDC = 8 is worth a closer look at whether improvement is worth the cost.
NDC Failure When %GRR Passes
NDC below 5 while %GRR is below 10% signals that your study parts were too homogeneous. The gauge may be perfectly capable, but the study did not sample enough variation to demonstrate it. The fix is to rerun the study with parts that span the full production tolerance range. Do not adjust the gauge.
Bias and Linearity: The Rest of MSA
%GRR and NDC cover the Gauge R&R portion of MSA. A complete PPAP MSA package often requires additional studies depending on the characteristic and customer requirements:
Bias Study: Measures whether the gauge reads consistently above or below the true (reference) value. A gauge that reads 0.15 mm high on every measurement has high bias. Bias is corrected by calibration, not by adjusting process settings. The AIAG standard for acceptable bias is typically evaluated using a t-test, with the null hypothesis that bias equals zero.
Linearity Study: Checks whether bias is consistent across the measurement range. A gauge that reads accurately at the low end of its range but drifts high at the upper end has a linearity problem. This requires multiple reference standards spanning the operating range and regression analysis. Linearity is more common in cases where the gauge is used across a wide measurement span.
Stability Study: Tracks whether the measurement system changes over time. Usually evaluated by running a control chart on a reference standard over weeks or months. IATF 16949 Clause 7.1.5.1 requires that calibration results be retained and that measurement system stability be verified.
For PPAP, the SQE will specify which studies are required. At minimum, a standard Gauge R&R study is expected for all characteristics that have a capability study at Element 11. Bias and linearity are more commonly required for high-precision CMM fixtures and manual measurement systems on critical characteristics.
MSA in PPAP and IATF 16949
PPAP 4th Edition addresses MSA at Element 10 (Measurement System Analysis Studies). The requirement is to conduct MSA studies for all gauges, measurement equipment, and test equipment used to measure critical characteristics on the part.
IATF 16949:2016 Clause 7.1.5.1.1 (Measurement System Analysis) requires the organization to conduct appropriate statistical studies to analyze the variation present in measurement and test equipment. It references the AIAG MSA manual as the prescribed methodology. Customer-specific requirements may override or supplement this, particularly for OEMs with additional MSA requirements in their CSRS (Customer-Specific Requirements Supplement).
Key IATF compliance points:
- MSA must be conducted before PPAP submission, not after
- Results must reference the specific characteristic, gauge ID, and calibration record
- If a measurement system falls in the conditional zone (10-30%), the documented rationale for acceptance must be retained
- Gauge calibration records must be linked to the MSA report
A common audit finding under IATF 16949 Clause 7.1.5.1.1 is an MSA study that was conducted on a different gauge than the one currently in production, or a study that references a part revision that no longer matches the current design record. Both trigger nonconformities.
When %GRR Is Too High: What to Do Next
Before accepting a failing GRR and hoping no one looks closely, work through the root cause in this order:
1. Check the study design first. If parts were all pulled from the same production lot, PV will be artificially narrow and %GRR will be inflated. Rerun with parts selected to cover the full tolerance range. This alone fixes many marginal results without touching the gauge.
2. Examine the appraiser contribution (AV). If reproducibility is the dominant contributor, the problem is operator technique, fixture use, or measurement procedure, not the gauge itself. Rewrite the measurement instruction, train all appraisers to the same technique, and rerun.
3. Examine the repeatability contribution (EV). High EV points to gauge instability: worn fixture, temperature sensitivity, part holding issues, or a gauge that is simply not capable of measuring to the required resolution for this tolerance. Repeatability problems require a hardware fix or a different gauge.
4. Check for interactions. If the GRR ANOVA table shows a significant operator-by-part interaction, different operators are reading different things on specific parts. This points to a fixturing or part-loading issue, where part orientation changes between operators.
5. Evaluate the risk-adjusted decision. For a characteristic at the boundary (20-28% GRR), document the risk assessment explicitly: what happens if this measurement system passes an out-of-spec part? How often does that pass/fail decision get made per shift? What is the downstream consequence? A 25% GRR on a cosmetic gap is lower risk than a 25% GRR on a dimensional constraint that drives assembly interference.
Running Gauge R&R Studies in QualityEngineer.ai
The Analyze module handles Gauge R&R using the AIAG crossed study method. Paste in your appraiser-part-measurement matrix, set the tolerance, and the module outputs:
- %GRR against study variation and tolerance
- %EV and %AV (repeatability and reproducibility breakdown)
- NDC calculation
- ANOVA table with operator-by-part interaction test
- Pass/conditional/fail classification against AIAG thresholds
- A graphical output showing appraiser consistency
The output is structured for direct inclusion in a PPAP Element 10 package. If %GRR falls in the conditional zone, the system flags it and prompts for risk documentation before marking the element complete.
For teams managing multiple active PPAP submissions, the Package module tracks MSA status across all elements for each project, with flags when a study is missing, expired, or conditional without documentation.
Summary
Gauge R&R acceptance in AIAG MSA 4th Edition comes down to two numbers: %GRR and NDC.
%GRR below 10% is acceptable without condition. Between 10% and 30% is conditional, requiring documented evaluation of characteristic criticality, process capability, and cost of improvement. Above 30% is not acceptable without customer engineering disposition.
NDC must be at least 5. NDC below 5 with low %GRR usually means the study parts were too similar. Rerun with parts that cover the full tolerance range.
When %GRR is high, diagnose root cause before deciding whether to rerun the study, improve the gauge, retrain operators, or negotiate a conditional acceptance. Submitting a failing MSA result without documentation or a corrective action plan is the fastest way to get PPAP rejected at Element 10.
The Analyze module at QualityEngineer.ai runs Gauge R&R with AIAG-compliant calculations, ANOVA breakdown, NDC, and PPAP-ready output. Start free, no credit card required.
