This article was originally published in March 2021, and has been updated as of June 2026.
Your line is running. Your team is logging data. But by the time your QA check flags a fill-weight drift, the shift is already over and the rework is already stacked on pallets. That's a timing problem, and it's exactly what the SPC vs. SQC distinction is designed to solve.
If you've ever wondered whether you're applying the right statistical method at the right point in your process, this is for you.
SPC vs. SQC: the core difference
Statistical process control (SPC) and statistical quality control (SQC) are related but they're not the same thing, and using them interchangeably will get you into trouble.
According to the American Society for Quality, Statistical Quality Control (SQC) is the application of statistical and analytical tools to monitor process outputs and ensure products meet customer specifications. It involves gathering and analyzing data to detect defects, eliminate variations, and maintain consistent quality. SPC, by contrast, applies those same statistical methods to the process inputs and in-process variables, in real time, before the product is made.
The practical difference: SPC is your early warning system. SQC is your final verification gate. You need both, but they answer different questions.
- SPC asks: Is my process in control right now?
- SQC asks: Does this finished product meet the customer’s expectation or desired spec?
A facility that only runs SQC is constantly playing catch-up. A facility that only runs SPC may miss the downstream verification that a retailer or auditor expects to see. The goal is to run them together and understand what each one is telling you.
How SPC works in practice
SPC uses statistical methods to help you understand and reduce variability in your manufacturing process. When you're monitoring a filling line, a cook temperature, or a moisture target, SPC gives you the tools to see whether the process is stable before it goes out of spec.
The core tools in any SPC program:
- Control charts (Xbar-R, Xbar-S): Real-time charts that show how a process variable changes over time. Xbar-R charts work well for sample sizes of two to nine. For larger sample sets (ten or more), Xbar-S charts using standard deviation give you better sensitivity.
- Control limits: The upper and lower control limits are set at three standard deviations from the mean, creating a six-standard-deviation spread around the target. When a point falls outside those limits, that's a signal, not just noise.
- Run rules: Statistical rules that detect when a process is drifting even before it crosses a control limit. Eight consecutive points trending in one direction is a run rule violation. Your process hasn't failed yet, but it's heading there.
- Process capability (Cpk and Ppk): These measures tell you how well your process is performing relative to specification limits. Cpk reflects short-term capability within subgroups; Ppk reflects overall process performance. A Cpk below 1.33 in most food manufacturing contexts is a conversation worth having with your team.
- Histograms: Show the distribution of your process data. A narrow bell curve with short tails means most of your output falls comfortably within spec. A wide, flat curve means you have variation you haven't explained yet.
When you have digital SPC charting with automated alerts, the value isn't just the chart. It's that the right person gets notified within minutes of a control limit breach, not at the end of the shift. That's when intervention actually prevents rework. Lincoln Premium Poultry made that shift from paper-based process tracking to real-time digital SPC monitoring and eliminated the lag between process drift and corrective action, without adding production headcount.
For a deeper look at putting these tools to work, see our guide on statistical process control in manufacturing.
Why SPC matters for QA managers specifically
Most facilities underuse SPC. They collect the data, plot the charts, and then review them at the end of the week. That's better than nothing, but it is not what SPC is designed to do.
SPC is designed for real-time decision-making. When your Xbar-R chart shows a drift trend, you have a window, maybe 20 minutes, maybe less, to adjust before nonconforming product is made. After the fact, SPC data is still useful for root cause analysis and CAPA documentation. But its primary value is upstream, not in the incident report.
When a control chart shows an out-of-control point, that event should trigger a documented corrective action. This can be documented digitally and the investigation starts immediately rather than waiting for a supervisor to notice a log entry. That closed loop is what root cause analysis informs for CAPA, and it's what auditors are looking for under both SQF and BRCGS schemes.
If you're preparing for a GFSI audit, your SPC records are documentary evidence. Auditors don't just want to see that you have control charts. They want to see that you acted on them. Learn more about what GFSI schemes require for process monitoring documentation.
How SQC works in practice
SQC is applied after the process runs. It answers a different question: did the output meet spec?
The most common SQC method is acceptance sampling. You draw a sample from a finished lot and test it against defined acceptance criteria. If the defect count in the sample falls below a threshold, the lot ships. If it exceeds the threshold, the lot is held, reworked, or rejected. Acceptance sampling doesn't guarantee a perfect lot, it's a probabilistic decision tool, not a 100% inspection. That's why "SQC helps manufacturers demonstrate that finished products consistently meet customer specifications" is the accurate framing, not "SQC ensures exact conformance."
SQC uses the same seven quality control tools described in Dr. Kaoru Ishikawa's Guide to Quality Control, including:
- Cause-and-Effect Diagram: Also known as a fishbone or Ishikawa diagram, used to identify potential causes of a specific problem.
- Check Sheet: A structured, prepared form for collecting and recording data or tallying problem occurrences.
- Control Chart: A graph used to study how a process changes over time and to identify variation.
- Histogram: A graphical representation showing the distribution and frequency of numerical data.
- Pareto Chart: A bar graph that helps prioritize problems based on their frequency, typically following the 80/20 rule.
- Scatter Diagram: A plot used to determine the relationship or correlation between two numerical variables.
- Stratification: A technique that separates data based on different sources or categories (such as shifts, suppliers, or machines) to spot patterns.
SQC data is also increasingly important for regulatory documentation. Under FSMA 204, finished lot records connected to critical tracking events (CTEs) need to be accessible and traceable. FDA's final traceability rule requires specific records for foods on the Food Traceability List. Your SQC logs, lot acceptance records, and in-process quality data aren't just internal documentation anymore. See our FDA Food Traceability Rule resource for specifics on what records need to be captured and retained.
FSMA 204 and 21 CFR Part 117: where SPC and SQC records become regulatory assets
This is the part most people miss until an auditor or an FDA investigator asks for it.
Under 21 CFR Part 117 preventive controls requirements, you're required to document monitoring activities for each process control, including the data used to verify that controls are working. Your SPC charts can be are that documentation. If your preventive control is a cook temperature, the Xbar-R chart for that CCP is your verification record.
Under FSMA 204, critical tracking events, initial packing, shipping, receiving, and transforming foods, require records that connect lot-level production data to traceability information. SQC acceptance records and in-process quality data are exactly the records that can support this chain of custody.
For more on preventive controls documentation requirements, see our guide to FSMA and PCQI responsibilities and key components of food safety compliance.
Using SPC and SQC together
The debate about whether SPC and SQC are interchangeable terms is mostly semantic. In practice, they work at different points in your process and they answer different questions. What matters isn't picking one, it's using both consistently and responding to what the data tells you.
A few patterns worth building into your program:
- Use SPC during production to detect drift before it becomes nonconformance
- Use SQC at the end of production to document lot-level disposition decisions
- Connect both to your CAPA process so that out-of-control events generate investigations, not just notes
- Make sure your records are in a format that supports audit requests under SQF, BRCGS, and FSMA 204 requirements
The cost of not doing this isn't abstract. Retailer chargebacks for out-of-spec product, FDA 483 observations for inadequate monitoring records, and SQF audit failures for missing corrective action documentation are all downstream consequences of gaps in SPC and SQC discipline. For more on quality assurance in the food industry, see our full guide.
What changes when you get this right
When your SPC program is running in real time with automated alerts, and your SQC records are digitally connected to lot disposition and traceability data, a few things shift.
Your QA manager stops chasing paper logs on Friday afternoon before an audit. Fill-weight drift gets caught during the shift, not after. When an FDA investigator asks for records supporting a critical tracking event, pulling a complete package takes an afternoon. And when an out-of-control point shows up on a control chart, a corrective action can be added and start a CAPA which opens immediately instead of surfacing in a weekly review meeting.
That's the operational difference between collecting statistics and using them.
Is still pulling SPC data manually for audits or reviewing control charts after the shift ends?
We can show you what automated, audit-ready charting looks like in a 20-minute conversation.
