A control chart isn't a pass/fail tolerance check — it answers a different question than "is this part in spec?" The common-cause vs. special-cause distinction that's the foundation of SPC, and why it's a required automotive core tool.
A control chart isn’t a pass/fail tolerance check — it’s a
running record of a process talking to you. Statistical Process Control
was built to answer a different question than “is this part in spec?”
It answers “is this process still behaving the way it did yesterday?”
— and those are not the same question.
Statistical Process Control was pioneered by Walter A. Shewhart at
Bell Labs in the early 1920s, who developed the control chart in 1924
as a way to distinguish normal process variation from a genuine signal
that something had changed. That distinction — common cause variation
(the everyday noise every process has) versus special cause variation
(a specific, identifiable disruption) — is still the entire foundation
of SPC a century later.
| Common cause variation | Special cause variation | |
|---|---|---|
| Source | Inherent to the process itself | An identifiable, specific disruption |
| Predictability | Statistically predictable within limits | Not part of the normal pattern |
| Correct response | Improve the process itself if the range isn’t good enough | Investigate and fix the specific cause |
| Example | Slight, normal variation in every machined part | A worn tool suddenly shifting the average dimension |
Reacting to common cause variation as if it were special cause —
adjusting the process every time a measurement drifts slightly — is a
well-documented way to actually make a process worse, a phenomenon
sometimes called “tampering.” SPC exists specifically to prevent that
overreaction.
SPC is one of the five automotive “core tools” required under IATF
16949, alongside APQP, PPAP, FMEA, and MSA — and it shows up just as
often in aerospace and general manufacturing quality systems. The
reason it’s treated as mandatory rather than optional is that it’s the
ongoing monitoring layer: PPAP proves a process can make good parts
once, SPC proves it keeps making good parts over time, and catches
drift before it produces a batch of nonconforming product.
A well-run SPC program puts the control chart in front of the
operator running the process, not just in a quality engineer’s report
reviewed after the fact. A point trending toward a control limit, or a
run of points on one side of the average, is a signal to investigate
before parts actually go out of tolerance — that’s the entire point of
“control” limits sitting inside the specification limits, giving a
warning window before a real defect happens.
SPC itself isn’t a standalone course in SCMEP’s current catalog —
it’s part of the quality systems foundation covered inside our
ISO 9001:2015 Overview and AS9100
training. As a NIST Manufacturing Extension
Partnership affiliate serving South Carolina manufacturers since
1989, our focus is connecting SPC’s statistical thinking to real
process data your team already collects, not abstract charting theory.
If your plant is building or improving an SPC program, you can
browse the Quality and Inspection
training catalog or email the training team.
SPC control limits are only meaningful if the measurement system producing the data can be trusted. See our related guide on Measurement System Analysis and Gage R&R for why MSA has to happen before SPC data means anything.
SPC is a method for monitoring a manufacturing process over time using control charts, developed by Walter Shewhart at Bell Labs in the 1920s. It distinguishes normal process variation from signals that something has genuinely changed.
Common cause variation is the normal, predictable noise inherent to a process. Special cause variation is an identifiable, specific disruption outside the normal pattern. Reacting to common cause variation as if it were special cause can actually make a process worse.
Yes — SPC is one of five automotive “core tools” required under IATF 16949, alongside APQP, PPAP, FMEA, and MSA.
The operator running the process, in real time — not just a quality engineer reviewing a report afterward. A well-run SPC program gives operators a warning signal before a process actually produces out-of-tolerance parts.
