Measurement System Analysis (MSA): A Beginner's Guide
In Lean Six Sigma, the saying is: you cannot improve what you cannot measure. But there is an equally important corollary -- you cannot trust what you cannot measure accurately. Before you analyse process data and draw conclusions, you need to know whether your measurement system itself is trustworthy. That is what Measurement System Analysis (MSA) is for. MSA is a structured set of methods for evaluating the adequacy of a measurement system: the instruments, the people using them, the procedures they follow, and the environment in which measurements are taken.
What MSA Evaluates
A measurement system can fail in several distinct ways. MSA evaluates five key properties:
Repeatability. When the same person measures the same part with the same instrument multiple times under the same conditions, do they get the same result? Repeatability failure means the instrument itself is inconsistent -- it introduces variation even when nothing else changes.
Reproducibility. When different people measure the same part with the same instrument, do they get the same result? Reproducibility failure means different appraisers are applying different technique, interpretation, or judgement -- the measurement depends on who is doing it.
Bias. Is the average measurement systematically higher or lower than the true (reference) value? A biased measurement system gives consistently wrong answers even if it is highly consistent. Bias is detected by comparing measurements against a known standard.
Linearity. Is bias consistent across the full operating range of the instrument? A gauge may be accurate at low values but biased at high values. Linearity analysis checks whether bias is stable across the range you actually use the instrument.
Stability. Does the measurement system remain consistent over time? A gauge that is accurate today may drift as it wears, as calibration lapses, or as environmental conditions change. Stability is evaluated by monitoring the same reference standard over time.
Gauge R&R Explained
The most commonly performed MSA study is the Gauge Repeatability and Reproducibility study, commonly called a Gauge R&R. It simultaneously quantifies repeatability and reproducibility variation and compares their combined magnitude to the total observed process variation. The result is expressed as a percentage -- the proportion of total variation that comes from the measurement system rather than from the process itself.
The interpretation benchmarks are widely accepted across industries:
Under 10% GR&R: Excellent. The measurement system is acceptable for its intended use.
10% to 30% GR&R: Marginal. The measurement system may be acceptable depending on the application. Improvement should be pursued.
Over 30% GR&R: Unacceptable. The measurement system contributes too much variation. Conclusions drawn from data collected with this system are unreliable. Improvement is required before the data can be used to drive decisions.
A Gauge R&R study typically involves having two or three appraisers each measure the same set of ten parts (chosen to represent the full range of process variation) two or three times, in a randomised order. Statistical software then decomposes the observed variation into its repeatability and reproducibility components.
A Non-Manufacturing Example: Permit Application Review
MSA is not only for manufacturing environments. Consider a municipal permitting office that is trying to reduce inconsistency in the time it takes to approve permit applications. Before analysing what drives approval time, it is worth asking: is the measurement of approval time itself consistent? Are different staff members recording the start date the same way (application receipt date vs. date of completeness review)? Is the end date the completion of technical review or the issuance of the permit? These definitional questions are the analogue of reproducibility in a manufacturing Gauge R&R. If two reviewers would record different approval times for the same application, the measurement system has a reproducibility problem. Resolving it requires standardising the definition and the recording process before any process analysis begins.
What to Do When Your Measurement System Fails MSA
If a Gauge R&R study reveals an unacceptable result, the path forward depends on the source of the problem. If repeatability is the dominant contributor, the instrument itself needs attention -- recalibration, maintenance, or replacement. If reproducibility is the dominant contributor, the problem is in how people use the instrument -- operator training, clearer procedures, or more explicit operational definitions can help. In service or knowledge-work contexts, reproducibility problems often stem from ambiguous definitions of what is being measured and need to be resolved by standardising the measurement process before anything else changes.
XNM applies Lean Six Sigma measurement and process improvement methods to public-sector and capital-project environments. Reach out to XNM's strategic advisory team to discuss process measurement and improvement for your organisation.