Key takeaways
Short answer: SPC and MSA are two complementary quality disciplines, and the order matters. SPC — Statistical Process Control — monitors a process over time, typically with control charts, to detect variation and keep the process stable and capable. MSA — Measurement System Analysis — validates the measurement system itself: is the gauge accurate and consistent enough that the data SPC acts on can be trusted? MSA comes first, because SPC built on a bad measurement system is controlling measurement noise rather than the process. For the variation SPC watches, see common cause vs special cause variation.
Statistical Process Control is the practice of monitoring a process over time using statistical methods — most visibly control charts — to detect variation, distinguish normal noise from real signals, and keep the process stable and capable. SPC plots process data against statistically derived control limits so you can see when the process is behaving normally (only common cause variation) and when something has changed (a special cause to investigate). Its purpose is control: keeping the process in a state of statistical control and catching problems early, before they produce large amounts of defective output. SPC is the ongoing, real-time discipline of watching and steering a process based on data. But it has a hidden dependency — every SPC chart is only as trustworthy as the measurements plotted on it, and SPC itself does nothing to check whether those measurements are any good.
Measurement System Analysis is the practice of validating the measurement system itself — confirming that the gauges, methods, and people producing your data are accurate and consistent enough to trust. MSA studies the variation that comes from the act of measuring rather than from the process: it assesses bias (accuracy against a true value), and the repeatability and reproducibility that make up measurement precision, typically through a Gage R&R study. Its purpose is to answer a foundational question: can we believe our data? If the measurement system contributes too much variation, the numbers are partly noise, and any decision based on them is suspect. MSA is the discipline that earns the right to trust your measurements, which is why it underpins everything else — SPC, capability studies, and the quality data behind OEE all assume the measurements are sound.
The clean distinction is what each validates. SPC controls the process — it watches process variation over time and keeps the process stable. MSA validates the measurement — it checks that the data SPC relies on actually reflects reality. They operate on different things: SPC on the process, MSA on the measurement system that observes the process. This is why they are not alternatives but a foundation and a structure built on it. The crucial insight is the dependency: SPC consumes measurements, and if those measurements are untrustworthy, the control chart is plotting noise. A special-cause signal might be a gauge problem, not a process change; a process that looks out of control might just be badly measured. SPC cannot tell the difference — only MSA can, by validating the measurement system first.
The sequence is not arbitrary: MSA should come before, or at least alongside, SPC, because SPC built on an unvalidated measurement system is unreliable from the start. If the gauge is inaccurate or inconsistent, the variation on your control chart is a mix of real process variation and measurement noise, and you cannot tell how much is which. You might chase special causes that are really gauge problems, or miss real process shifts hidden under measurement scatter. Worse, capability indices computed from bad measurements are fiction. So a disciplined quality programme validates the measurement system with MSA first — confirming the gauge is accurate and its repeatability and reproducibility are acceptable — and only then trusts SPC to control the process. Doing SPC without MSA is building a control system on a foundation you have never checked.
A team sets up an SPC chart on a critical dimension and immediately sees alarming variation and frequent out-of-limit points — the process looks out of control. Before tearing the process apart, they run an MSA: a Gage R&R study on the gauge used to take those measurements. It reveals that the measurement system itself contributes the majority of the observed variation — the gauge is imprecise, and much of the scatter on the chart is measurement noise, not process variation. The process was never as unstable as it looked; the measurement was. After fixing the gauge, a fresh SPC chart shows a far more stable process. Had they skipped MSA and trusted the SPC chart, they would have chased phantom process problems for weeks. MSA first revealed that the measurement, not the process, was the real issue.
SPC and MSA together underpin the trustworthiness of the quality factor of OEE. The quality factor counts good versus defective units, a judgement that depends both on the process being controlled (SPC's job) and on the measurements being trustworthy (MSA's job). MSA connects directly to repeatability and reproducibility and to precision and accuracy — the qualities a measurement system needs — while SPC connects to common versus special cause variation. Get either wrong and the OEE quality number becomes unreliable: an uncontrolled process produces real defects, and an unvalidated gauge produces fake ones. Both disciplines are the unglamorous foundation beneath an honest quality factor.
Fabrico consumes the quality outcome that SPC controls and MSA validates, so the integrity of both flows into its OEE. By trending the good-versus-defective result over time, it surfaces the patterns that signal trouble — a reject rate that shifts with a gauge change (an MSA concern) or a process drifting out of control (an SPC concern) — prompting the right investigation. Reliable measurement and a controlled process upstream, honest OEE downstream, go together. Book a demo to see controlled, trustworthy quality data drive your OEE.
SPC (Statistical Process Control) monitors and controls process variation over time, usually with control charts. MSA (Measurement System Analysis) validates that the measurement system itself is accurate and consistent enough to trust. SPC controls the process; MSA checks the measurements SPC relies on.
Because SPC built on an untrustworthy measurement system controls measurement noise rather than the process. If the gauge is inaccurate or inconsistent, the variation on a control chart mixes real process variation with measurement error, and you cannot tell them apart. MSA validates the measurements first.
MSA assesses the variation that comes from the measurement system: bias (accuracy against a true value) and the repeatability and reproducibility that make up measurement precision, usually through a Gage R&R study. It answers whether the data can be trusted.
You can, but it is risky. Without validating the measurement system, an SPC chart may show variation that is actually gauge noise, leading you to chase phantom process problems or miss real ones. Disciplined quality programmes validate measurements with MSA first.
Both underpin the trustworthiness of the OEE quality factor, which counts good versus defective units. SPC keeps the process controlled and MSA ensures the measurements are trustworthy. An uncontrolled process produces real defects; an unvalidated gauge produces fake ones — both corrupt the quality number.
Anmeldung
Blog
