Gage R&R (Repeatability and Reproducibility) is a measurement-system analysis that separates true part-to-part variation from variation introduced by the gauge and the operators. Repeatability is variation when the same person re-measures the same part; reproducibility is variation between different people.
You cannot trust capability or SPC conclusions if the measurement system is noisy; a typical rule is that measurement variation should consume less than 10% of total tolerance. Gage R&R is a standard early step in the DMAIC Measure phase.
The standard study: 10 parts spanning the process range, 3 operators, each measuring each part 2 or 3 times in random order. The analysis (ANOVA or average and range) splits total variation into repeatability (equipment), reproducibility (operators) and part-to-part. The result is reported as %GRR of tolerance and of total variation: under 10% is acceptable, 10 to 30% is marginal and depends on the use, over 30% means the gauge cannot be used for that decision.
Example. Tolerance is 0.20 mm and the study gives a measurement standard deviation of 0.012 mm. GRR = 6 x 0.012 = 0.072 mm, or 36% of tolerance: a reading of 0.09 mm from nominal might be a good part or a bad one. Fix the fixture, the method or the gauge before running capability. The metrology guide covers the study setup and the number of distinct categories check.