Manufacturing — statistical process control inside ISO 9001
How statistical process control (SPC) sits inside ISO 9001 clauses 9.1.1 and 8.5.1 — control charts, capability indices, and the manufacturing-shop pitfalls.
- ISO 9001
ISO 9001 does not require statistical process control. It does require “determination of monitoring and measurement” sufficient to ensure valid results (clause 9.1.1) and “implementation of monitoring and measurement activities at appropriate stages” of production (8.5.1.c). For manufacturers with critical-to-quality characteristics, SPC is the most defensible way to meet that obligation. This article covers how SPC fits inside an ISO 9001 QMS, practical placement, charts to use, and the shop-floor pitfalls.
What SPC is, and what it is not
Statistical process control is a discipline for understanding and managing process variation. It uses sample data to distinguish common- cause variation (inherent in the process) from special-cause variation (assignable to a specific root cause). The classical toolkit:
- Control charts, track sample statistics over time and signal out-of-control conditions.
- Process capability indices, Cp, Cpk, Pp, Ppk, relate process variation to specification.
- Run rules, patterns that signal special cause earlier than a single point exceeds the control limit.
SPC is not:
- Goal-post checking. “All parts within spec” is not SPC; it is inspection.
- A spreadsheet. SPC is a discipline; a chart is a tool.
- A substitute for design control. SPC catches process drift, not design defects.
Where SPC fits in the ISO 9001 QMS
Clause 4.4, process approach
Identify the processes whose output materially affects product quality. SPC applies to those processes; it is rarely justified everywhere.
Clause 6.1, risk and opportunity
The risk register identifies which characteristics are critical-to- quality. Critical characteristics get SPC; non-critical do not. Document the rationale.
Clause 7.1.5, monitoring and measurement resources
The measurement system itself must be capable. Gauge R&R studies precede SPC, if your measurement variation is comparable to your process variation, your control chart is measuring the gauge, not the process.
Clause 8.5.1.c, production monitoring
This is where SPC operationally lives. Define the characteristic, the sample plan, the chart type, the control limits, the reaction rules.
Clause 9.1, monitoring and measurement
SPC outputs become QMS evidence. Out-of-control signals enter the nonconforming output workflow under 8.7 and the corrective action workflow under 10.2.
Clause 9.3, management review
Capability trends and out-of-control rates are management review inputs.
Choosing the right control chart
| Data type | Chart | When to use |
|---|---|---|
| Variable, individual measurements | I-MR (Individuals and Moving Range) | Low-volume processes; expensive measurement |
| Variable, subgrouped | X-bar and R | Subgroups of 2 to 10; standard production |
| Variable, subgrouped, large subgroups | X-bar and S | Subgroups above 10 |
| Variable, exponentially weighted | EWMA | Detecting small shifts |
| Variable, cumulative | CUSUM | Detecting persistent shifts |
| Attribute, defectives per sample | p chart | Variable sample size |
| Attribute, defectives per fixed sample | np chart | Fixed sample size |
| Attribute, defects per unit | u chart | Variable unit size |
| Attribute, defects per fixed unit | c chart | Fixed unit size |
Chart selection is driven by the data type and the production cadence — not by tradition.
Process capability, Cp, Cpk, Pp, Ppk
| Index | Meaning | When to use |
|---|---|---|
| Cp | Process capability based on within-subgroup variation | Comparing process spread to specification width |
| Cpk | Cp adjusted for centring | Most common; reflects centring relative to specification |
| Pp | Process performance based on overall variation | Reflects total variation including between-subgroup |
| Ppk | Pp adjusted for centring | Reflects total performance including drift |
Rule-of-thumb thresholds: Cpk ≥ 1.33 for new processes; Cpk ≥ 1.67 for critical safety characteristics. Always pair Cp with Cpk, Cp alone hides centring problems.
Reaction rules
When a chart signals out-of-control:
- Stop or hold. Do not continue making nonconforming product.
- Assess the lot. Quarantine; assess against specification.
- Investigate. Root cause, often via 5-why or fishbone.
- Correct. Address the special cause; verify the chart returns to control.
- Document. Nonconformity register; corrective action under 10.2 if recurrence is likely.
- Update. Risk register; control plan; FMEA; documented information.
Common shop-floor pitfalls
- Operator-adjusted control limits. Limits are set by the process, not by negotiation. Drift over time? That is signal, not noise.
- Overuse of run rules. Western Electric run rules increase false- alarm rates. Use them deliberately, not as defaults.
- Charts on the wall, untouched. A chart not used in the next decision is decoration. Pair the chart with a documented reaction protocol.
- Inspecting in quality. SPC focuses upstream; final inspection is post-hoc and expensive. The QMS audit will look for upstream evidence.
- SPC on uncontrolled measurement systems. Gauge R&R first, SPC second.
- Mixing data sources. Different gauges, different operators, different times, treat them as strata; do not pool blindly.
SPC in regulated manufacturing
- Automotive. IATF 16949 makes SPC effectively mandatory for special characteristics; AIAG SPC reference manual is the working text.
- Aerospace. AS9100 picks up SPC via the production-control clause; customer-flow-down often requires capability evidence per characteristic.
- Food. FSSC 22000 leans on PRPs and HACCP; SPC applies where the process is statistically meaningful (filling weights, pasteurisation temperatures).
- Medical devices. ISO 13485 requires process validation where output cannot be fully verified, SPC is often part of the validation evidence.
Documented information
For each SPC application:
- The control plan or equivalent, characteristic, sample plan, chart type, limits, reaction rules.
- Gauge R&R study reference.
- Capability study evidence at process qualification and at periodic re-qualification.
- The chart history.
- Out-of-control event records and corrective actions.