Process Control: Definition, Types & Role in Modern MES

What is process control?

Process control is the monitoring and regulation of manufacturing processes so that they stay within defined parameters and produce consistent output. The term sounds narrow but is used for three quite different disciplines, and confusing them is the single most common mistake I see in plant-level discussions: advanced process control (APC) inside continuous processes, statistical process control (SPC) for discrete quality management, and in-process monitoring — the real-time capture of process parameters from machines. A modern MES is almost always talking about the third, often in combination with SPC, and very rarely about APC in the classical sense.

This matters because the tooling, the audience and the expected result differ sharply between the three. A chemical engineer, a Six Sigma Black Belt and a production manager can all say "we need process control" in the same meeting and mean three different things. Sorting that out in the first ten minutes is worth more than any vendor comparison that follows.

The three things "process control" actually means

APC lives inside the machine — it is a control-engineering discipline that existed long before MES and will continue to exist alongside it. SPC is a statistical method, historically paper-based, now almost always running inside a quality module or MES. In-process monitoring is the newest of the three and the one with the largest practical impact on discrete manufacturing in the last decade: for the first time, the actual process parameters that ran when a specific part was produced can be stored, correlated with reject events and retrieved years later for a customer complaint.

Process control vs. production control vs. quality control

Three adjacent terms that are routinely mixed up:

• Process control — keeps the process parameters inside their defined window (injection pressure, welding current, oven temperature).
• Production control — keeps the order flow on track (dispatching, sequencing, reacting to disruptions). This is the classical MES execution layer.
• Quality control — judges the output against specification (inspection, SPC, release decisions).

A well-run plant does all three and connects them. A poorly-run plant optimises one in isolation and is then surprised that the others fail. The common pattern I see: tight APC on the machines, almost no in-process monitoring at MES level, sporadic SPC on paper, and almost no link between a reject event and the process parameters that were running when it happened. That missing link is the single biggest opportunity.

Why in-process monitoring is the part MES buyers actually care about

For a discrete or batch manufacturer evaluating an MES, "process control" is almost always shorthand for in-process monitoring plus SPC — the ability to capture process parameters continuously, correlate them with part outcomes, and flag drift before it becomes scrap. The APC layer is already solved by the machine vendor; no MES replaces a PID loop. What the MES adds is three things the machine alone cannot provide:

1. Persistence across parts and batches. The PLC knows the current pressure. Only the MES remembers which pressure ran on part serial 482193 at 03:14 last Tuesday.
2. Correlation across machines. A quality escape is rarely caused by one machine alone. Correlating process data across upstream and downstream steps is how root causes actually get found.
3. Rule-based triggers. SPC rules (Western Electric, Nelson) running live against process data, raising alarms before a control limit is breached, not after.

Concretely, this is what SYMESTIC's Process Data module does: it captures parameters via OPC UA or digital I/O, stores them against the production order and part, and makes them queryable alongside stop reasons, reject counts and OEE. The common realisation after the first month of live capture is the same in almost every plant — the process was never as stable as the team thought, and the correlation between one specific parameter and a specific reject type is much stronger than anyone had suspected.

Common pitfalls

FAQ

Is process control the same as SPC?
No. SPC is one specific method within process control — a statistical technique for judging whether a process is stable over time. Process control as an umbrella term also covers closed-loop regulation inside machines (APC) and continuous in-process monitoring via an MES. In most discrete-manufacturing discussions, "process control" used without qualification means in-process monitoring combined with SPC, which is also what a modern MES delivers.

Does an MES replace DCS or PLC-level control?
No, and it shouldn't try. The control-engineering layer — PID loops, safety interlocks, real-time regulation — belongs inside the machine and runs at cycle times an MES cannot reach. What the MES adds sits one layer above: capturing the parameters the PLC is already using, storing them against the production order, correlating them with outcomes, and making them queryable for root-cause analysis and traceability.

Which process parameters should we capture first?
Start with the parameters that operators already watch and that show up in reject discussions — typically five to ten per process step. Injection pressure and cycle time for moulding, welding current and duration for joining, oven temperature and belt speed for curing. Capturing everything the PLC exposes on day one produces volume without insight; capturing the parameters with a known quality link produces immediate return and creates the habit of looking at the data.

How does SYMESTIC support process control?
Through the Process Data module, which captures parameters directly from the PLC via OPC UA or digital I/O, stores them against production order and part, and integrates with SPC rules, reject capture and OEE in one platform. The typical deployment pattern across the 15,000+ machines SYMESTIC has connected: start with OEE and stop-reason capture, add process data on the two or three most reject-sensitive parameters per process, and expand from there. Go-live is measured in days, not months, and the first correlations between a specific parameter and a specific defect mode usually emerge within the first few weeks of live data.

Related: SPC · Cp/Cpk · Traceability · OEE · MES · Process Data Product

About the author

Christian Fieg

Head of Sales at SYMESTIC. 25+ years in manufacturing. Six Sigma Black Belt, former global MES & traceability lead at Visteon (900+ machines, 30+ processes). Author of "OEE: One Number, Many Lies" (2025). · LinkedIn