Production Control: Definition, Methods & MES Role

What is production control?

Production control is the operational discipline that turns a production plan into actual output on the shopfloor. It covers dispatching orders to the right machine at the right time, sequencing work to use capacity efficiently, reacting to breakdowns and missing materials without losing the day, and reporting what actually happened back to planning. It is different from production planning, which decides what to make and when. It is different from ERP scheduling, which sets the calendar. Production control is what happens between the plan and the part.

Over thirty years in manufacturing, including the process control and Manufacturing Execution Systems work I brought into SYMESTIC when I founded it in 1995, the single most common confusion I see is plants believing they have production control because their ERP has a production module. ERP does planning. Production control happens at MES level, because it needs real-time feedback from machines, and that is not what ERP systems are built to deliver. Getting this distinction right is the first step toward a production control system that actually controls anything.

Production control, production planning, PPC: the terms

Five terms are used almost interchangeably in practice and should not be. Clean definitions save a lot of arguments between production, planning and controlling departments.

The practical consequence: if someone asks to improve "production control" and their starting point is an ERP screen that shows planned finish dates, they are not actually working on production control yet. They are working on scheduling. Real control starts when the machine tells you what is happening right now.

Where production control sits: ISA-95 and the MES layer

The ISA-95 standard (IEC 62264) defines production control precisely as part of Level 3 activities, inside Production Operations Management. The named sub-functions are Production Dispatching, Production Execution Management, Production Tracking, Production Data Collection and Production Performance Analysis. Together they are what an MES covers on the production side.

Expressed as a flow, production control does five things in sequence, continuously:

1. Dispatching. The MES receives released production orders from ERP and assigns them to specific work centres, in a specific sequence, with correct work instructions, recipes and tooling.
2. Execution management. While an order runs, the MES enforces the correct process: valid material consumption, authorised operators, approved tools, mandatory checks. Errors are caught at the station, not at final inspection.
3. Tracking. Every cycle, every good part, every scrap, every stop is recorded against the running order in real time. This is the data that makes production control different from production planning.
4. Data collection. Operational, performance and quality data is captured from machines and operators and stored against the production order. Not on paper, not at shift-end, continuously.
5. Performance analysis. Actual versus planned, down to the minute. Variance feeds back into ERP and into the next planning cycle, which is the closed loop that MRP-only environments never quite achieve.

ERP sends orders down. MES executes and sends reality back up. Without the MES layer, production control is a shift log and a phone call.

The four classic methods, and when each actually works

Most textbook articles list Push, Pull, JIT and Kanban as if they were interchangeable menu options. They are not. Each works under specific conditions and fails under others. The choice depends on demand pattern, lead time, setup cost and supplier reliability.

Push (MRP logic). Production is triggered by forecasts and planned order releases. Works when demand is reasonably stable, batch sizes are large relative to takt and setup cost is high. Fails when forecasts drift, because WIP accumulates and obsolete stock grows. Still the dominant approach for most discrete manufacturing in Germany, usually combined with another method for critical parts.

Pull (Kanban). Production is triggered by downstream consumption. Works when consumption is regular, replenishment lead time is short and product variety is bounded. Fails when demand is lumpy or setup cost is high, because either containers stay empty or idle changeovers dominate. Excellent for high-runner SKUs, poor for project-driven production.

Just-in-Time (JIT). A broader philosophy than Kanban, covering synchronisation across suppliers, production and dispatch. Works when supplier reliability is high and the cost of stockout is absorbable (or the cost of inventory is unacceptable). Fragile under disruption, as the 2020 to 2022 supply chain shocks reminded everyone.

Drum-Buffer-Rope (DBR, from Theory of Constraints). Production pace is dictated by the bottleneck, everything else is subordinated to it. Works brilliantly when the bottleneck is stable and identifiable. Adds real value beyond MRP when applied rigorously, but requires the discipline to stop treating local utilisation as a goal.

Real plants use all four in combination. A typical automotive tier-one operation I see has MRP-push for the long horizon, Kanban for fast-moving components, JIT-synchronised sequenced delivery for final assembly, and implicit DBR on the press line. The choice of method is not the battle. The battle is whether production control has the real-time data to enforce whichever method was chosen.

What production control actually does on a real shopfloor

In concrete terms, a day in the life of a modern production control system looks like this:

• An ERP production order is released and appears at the relevant work centre terminal within seconds, with the current revision of the routing, BOM, setup instructions and quality checks.
• The operator logs in, acknowledges the order, and the MES begins counting against it: cycles, good parts, scrap by reason code, machine state changes.
• A stop occurs. The PLC reports the alarm code, the MES categorises it automatically, and the shift supervisor's tablet shows the event within seconds. The supervisor does not find out at shift handover.
• A competing priority comes in. The production planner adjusts the sequence in the scheduling board and the MES re-dispatches the next orders without a paper travel running around the hall.
• The order completes. Actual quantities, actual time, actual scrap and actual material consumption are booked back into ERP automatically. No paper note, no manual rebooking, no reconciliation meeting on Friday.
• At the end of the shift, planned versus actual is on a dashboard, variance is categorised, the top three causes of the variance are named, and the plant manager sees the same number as the controller.

That list is not a vision of the future. It is what functioning production control in a connected factory looks like, and it is what SYMESTIC deploys into plants in weeks, not months.

Where reported production control diverges from reality

Four patterns inflate production-control numbers and hide the real state of the plant. All four are common.

1. ERP progress percentages based on booked labour. "Order 40 % complete" often means operators have booked 40 % of the planned hours, not that 40 % of the parts exist. The two numbers diverge whenever cycle time is off, which is almost always.

2. Setup time overruns booked to the wrong order. A changeover ran 15 minutes long. The excess gets booked into the next order's setup or into the prior order's run time. The variance against either order is contaminated. Over a month, this alone can account for 3 to 5 percentage points of false utilisation.

3. Paper travellers diverging from ERP state. The MES, or ERP, shows one status. The hall shows another. Someone reconciles them at shift end. The reconciliation becomes the record, and the drift is invisible.

4. Weekly production meetings on Monday at 10. By the time the meeting happens, the data is four days old, the plant has moved on, and the decisions being made are about a state that no longer exists. Production control without real-time data is not control, it is narration.

Practical rule: if the answer to "what is happening on line 3 right now?" takes more than two clicks, production control is still a job title, not a system.

What changes with a modern MES

The shift from ERP-only to ERP plus a modern MES is not about adding screens. It is about closing three loops that most plants never had closed:

1. Plan to execution. Production orders land on the floor in real time with the right context, not as printouts that already disagree with the system by the time the operator reads them.
2. Execution to planning. Actuals flow back automatically, which makes variance analysis honest and makes the next plan better rather than just later.
3. Execution to quality. Production control triggers quality actions automatically, for example sampling rules tied to cycles produced, not to a paper plan that the operator may or may not remember.

The measurable effect is predictable. Plants that close these loops typically see 5 to 10 percentage points of availability improvement in the first six months, not because they changed equipment, but because they stopped accepting invisible losses.

A real case: Meleghy Automotive

Meleghy Automotive is a tier-one body-in-white supplier with plants in Germany, Spain, Czechia and Hungary. The starting point was the familiar one: SAP R/3 for planning, a mix of modern and legacy presses and joining lines on the floor, and a gap in between where paper travellers and shift meetings tried to hold the operation together.

The SYMESTIC deployment started at Wilnsdorf with a focused production control scope: OEE capture at the critical process steps, bidirectional SAP R/3 integration via ABAP IDoc mapping machine cycles to production orders, and bidirectional CASQ-it integration to trigger quality sampling directly from production events. The same modular configuration was then rolled into Gera, Brandýs, Bernsbach, Reinsdorf and Miskolc within six months. Meleghy scaled it themselves using the SYMESTIC kit, which is the point.

The measured results across the six plants after the rollout:

• 10 percent reduction in downtime
• 7 percent increase in output
• 5 percent improvement in availability
• Six plants live in six months, with the customer running expansion autonomously

None of those numbers came from new equipment. They came from production control that could actually see what was happening, in real time, against the production orders that SAP was still responsible for releasing. ERP kept its job. MES took over the job ERP was never designed to do.

FAQ

How is production control different from production planning?
Planning decides what and when over a long horizon, typically in ERP or APS. Control executes the plan on the shopfloor in real time, typically in MES. A plant can plan well without controlling well, and the planning quality shows up in on-time delivery. A plant can control well without planning well, and the control quality shows up in responsiveness to disruption. Both are needed.

Does ERP do production control?
Only in a limited sense. ERP releases orders and collects confirmations, but it does not observe machines, does not react in minutes and does not enforce execution rules at the station. What ERP calls "production control" is scheduling plus order progress reporting. The execution layer is MES.

What is the difference between APS and production control?
APS is a detailed finite-capacity scheduling tool. It produces the plan the production control system will then execute. APS and MES are complementary, not alternatives. Many mid-size plants get far with ERP scheduling plus MES execution, and only add APS when product variety and constraint complexity exceed what ERP can schedule realistically.

Can we do production control without an MES?
Technically yes, using paper travellers, Kanban boards, operator calls and spreadsheets. In practice, without real-time machine data, production control degrades into production reporting with a lag. Below roughly 10 connected machines the overhead may not be worth the data. Above that, manual control becomes a liability.

How long does it take to implement digital production control?
With a cloud-native MES and a clear scope, first production lines can be live in weeks. Full plant rollouts usually land between three and six months. The Meleghy example covered six plants in six months. Multi-year MES projects belong to on-premise implementations from the previous generation, and are no longer a prerequisite.

Where does production control end and shopfloor management begin?
Production control delivers the facts: what happened, against what plan, with what variance. Shopfloor management is the human cadence built on top of those facts: daily meetings, problem-solving boards, continuous improvement actions. Control without shopfloor management produces dashboards nobody acts on. Shopfloor management without control produces meetings nobody trusts.

What does SYMESTIC's Production Control offering cover?
Order dispatching, execution enforcement, real-time tracking, exception handling, bidirectional ERP integration (SAP, Infor, Microsoft, proAlpha and others), and connection to quality, maintenance and energy modules. Deployment is cloud-native with go-live typically in weeks. See SYMESTIC Production Control.

Related: MES · Production Planning · Bottleneck · Operating Time · Machine Availability · OEE · SYMESTIC Production Control

About the author

Uwe Kobbert

Founder and CEO of SYMESTIC GmbH since 1995. 30+ years in manufacturing, starting with process control systems and MES for food and beverage at STERIA, then building SYMESTIC from a classic MES consultancy into a cloud-native platform running on 15,000+ machines across 18 countries. Self-funded, no external investors. · LinkedIn