Gauge R&R (Repeatability and Reproducibility) is a statistical study that quantifies how much of the variation in a set of measurements comes from the measurement system itself, rather than from the parts being measured. It is the core technique inside Measurement System Analysis (MSA), and it answers one blunt question: can you trust the numbers your gauges and inspectors produce? If a measurement system contributes too much of the observed variation, every downstream decision built on that data (quality gates, capability studies, control charts) is compromised. Gauge R&R separates measurement error into two components, repeatability and reproducibility, and expresses their combined effect as a percentage so you can accept, improve, or reject the system.
The two halves of Gauge R&R describe different sources of measurement error, and confusing them leads to fixing the wrong problem.
Combined, they form the total measurement system variation, usually written as GRR. The remaining variation in your study should come from genuine part-to-part differences, which is the signal you actually want to detect.
A measurement system is the lens through which you see your process. If the lens is distorted, you cannot trust anything you observe through it. This is why a validated Gauge R&R study must precede statistical process control and any process capability analysis. If your gauge adds 25 percent of the total variation, a control chart may flag false alarms or, worse, mask real special-cause signals. Capability indices like Cp and Cpk become meaningless because part of the spread they measure is pure measurement noise. A structured improvement effort such as DMAIC almost always validates the measurement system in the Measure phase for exactly this reason: you cannot improve what you cannot measure reliably.
Gauge R&R results are reported as a percentage so they can be judged against a standard rule. There are two common bases:
A related metric is the number of distinct categories (ndc), which estimates how many separate groups of parts the system can reliably tell apart. An ndc of 5 or more is generally considered acceptable.
The Automotive Industry Action Group (AIAG) guidelines give a simple decision rule based on the %GRR value:
These thresholds are guidelines, not laws. A safety-critical dimension may demand tighter limits, while a rough sorting gauge might tolerate more.
Imagine a Gauge R&R study on a shaft-diameter caliper. After running the standard deviation calculations, you obtain these values in millimeters:
Because these components combine in quadrature (variances add, not standard deviations), first square each value. Repeatability variance is 0.008 squared = 0.000064. Reproducibility variance is 0.006 squared = 0.000036. The GRR variance is their sum: 0.000064 + 0.000036 = 0.000100, so the GRR standard deviation is the square root, 0.010 mm.
Next, find the total variation. Part-to-part variance is 0.040 squared = 0.001600. Total variance = GRR variance + part variance = 0.000100 + 0.001600 = 0.001700, so total standard deviation is the square root, roughly 0.041231 mm.
Now compute % Study Variation: (GRR standard deviation / total standard deviation) times 100 = (0.010 / 0.041231) times 100 = 24.25 percent. That result falls in the 10 to 30 percent band, meaning the system is marginal and worth improving. Since repeatability (0.008) is larger than reproducibility (0.006), the caliper resolution or condition is the bigger lever, so recalibrating or upgrading the gauge would help more than retraining operators.
When a study fails, the root cause usually falls into a handful of categories. Ranking them with a Pareto analysis helps you attack the biggest contributor first.
Fixing these connects directly to broader reliability discipline. A well-run total productive maintenance program keeps gauges calibrated and in good condition, and a FMEA can flag measurement failure modes before they reach the shop floor.
To be clear, running a formal Gauge R&R study is a dedicated MSA exercise performed in statistical software, and Fabrico is not a Gauge R&R calculator. What Fabrico provides is the reliable production data layer that makes measurement discipline actionable at scale. Fabrico delivers real-time OEE and production monitoring, and its CMMS product tracks the assets, work orders, and preventive schedules that keep your measurement equipment calibrated and in-spec. Fabrico also offers computer-vision monitoring that captures cycle and stoppage data even on machines with no PLC, so quality events are logged consistently rather than by inconsistent manual entry. Good measurement systems reduce the unplanned downtime that stems from misjudged quality, and a shared CMMS ensures every gauge has a documented calibration history. Once your measurements are trustworthy, Fabrico gives that clean data a permanent, EU-hosted home.
The classic AIAG crossed study uses 10 parts, 3 operators, and 3 trials each, for 90 total measurements. The parts should span the full range of process variation, and operators should measure in random order without seeing prior results. Fewer parts or trials reduces the statistical confidence of the estimate, so the 10 by 3 by 3 layout remains the practical standard for most dimensional checks.
No. Calibration compares a gauge against a known reference standard to confirm its accuracy (closeness to the true value), while Gauge R&R measures precision (the spread of repeated readings). A gauge can be perfectly calibrated yet still fail Gauge R&R if its readings scatter, and it can be repeatable yet biased if it is out of calibration. Both are needed, and a proactive maintenance program should schedule both.
Gauge R&R validates the measurement inputs, while reliability metrics such as MTBF and MTTR describe equipment performance over time. They are complementary: if your downtime and failure timestamps are captured inconsistently, your reliability metrics inherit that measurement noise. Trustworthy data collection, which Fabrico automates, is the shared foundation beneath both quality and reliability analytics.
A trustworthy measurement system is the starting line for every quality initiative, and reliable production data is what keeps it honest. See how Fabrico turns real-time OEE, CMMS, and computer-vision monitoring into a clean, EU-hosted data foundation for your quality program by booking a demo.
Login
Blog