BLOG
What Is an MES System? Production Monitoring and OEE for Data-Driven Manufacturing
The sentence we hear most often from production managers is this: “At month end we know what we produced — we just don’t know what stopped us producing more.” An MES system exists to close exactly that gap. This article explains what an MES system is, how it differs from ERP and SCADA, how OEE tracking is set up, and where to start a production monitoring project.
What Is an MES System?
MES stands for Manufacturing Execution System: the software layer that monitors and manages how a production order is actually executed on the shop floor, in real time. It takes the “what to produce” instruction from the planning layer, collects “what happened” data from the floor, and makes the gap between the two visible.
In practice an MES system answers questions like: which work order ran on which machine, with which operator, and when did it start? How long were the stoppages and what caused them? How much scrap came out of which batch? When a customer complaint arrives, which raw material and which process settings produced that lot?
MES vs ERP vs SCADA
These three layers get confused constantly. The shortest possible distinction:
| Layer | Question it answers | Time scale |
|---|---|---|
| ERP | What should be produced, at what cost, against which order? | Days / weeks |
| MES | How is the order actually running, and where are we losing? | Minutes / hours |
| SCADA / PLC | What state is the machine in right now, at what values? | Milliseconds / seconds |
Having an ERP does not mean you do not need MES — rather the opposite. When the ERP plan misses, the ERP cannot tell you why, because it collects no shop-floor data. MES produces that reason and feeds it back.
OEE: The Core Metric of Production Monitoring
OEE (Overall Equipment Effectiveness) is the product of three components, and it should be the first output of any production monitoring system:
- Availability: for how much of the planned production time was the machine actually running? Breakdowns, changeovers, material waiting and shift stoppages land here.
- Performance: while running, what share of the ideal cycle rate was achieved? Slow running and micro-stops land here.
- Quality: how much of the output was right first time? Scrap and rework land here.
The value of OEE is not the single percentage but the breakdown of where the loss sits. If 60% OEE is driven by availability loss, the answer lies in the maintenance plan; if it is quality loss, the answer lies in the process and in inspection. Improvement spending made without that distinction usually goes to the wrong place — we discussed why measurement culture so often fails to take root in our measurable factory article.
Where Does the Data Come From? The Real Difficulty in MES Projects
Buying MES software is easy; building the data that feeds it is hard. The typical sources we work with:
- PLC and CNC controllers: cycle counters, alarm codes, program numbers. The most reliable source, but every brand speaks a different protocol (OPC UA, Modbus, MTConnect).
- Operator terminals: stoppage reason, scrap entry, order start and stop. Because a human types it, accuracy depends on interface quality.
- Sensors and smart meters: energy, vibration, temperature, pressure. Usually the most practical way to connect legacy machines to an MES.
- Cameras and machine vision systems: piece counts and quality results. An inspection system such as MIS-INSPECT feeds the quality component of OEE without any manual logging.
As we covered in industrial IoT data management, the critical design decision is which data is processed at the edge and which is moved to the centre. Streaming every signal to the cloud is both expensive and unnecessary.
MES Data Is Not Just Reporting: Predictive Maintenance and AI
Once a production monitoring system runs, the resulting data set enables more than reporting. Combining stoppage records with sensor data supports predictive maintenance models; matching quality records against process parameters reveals the settings that trigger scrap. Training, monitoring and updating those models in the field is its own discipline — we manage it in the process-intelligence layer with MIS-AGENT.
How to Start an MES Project
The most common mistake we see is starting with an MES project that covers the entire plant at once: rollout stretches over months, data reliability stays disputed, and nobody takes ownership. This sequence works better:
- Pick one bottleneck line. Where the loss is most expensive is where payback shows fastest.
- Automate stoppage and count data first. Leave manual entry only for the stoppage reason.
- Make OEE visible at the line. A metric the shift team cannot see will not change behaviour.
- Fix a short, stable list of stoppage reasons. Nobody fills in a 40-item reason list.
- Add quality and traceability afterwards. Batch-level traceability only means something once the base data is trustworthy.
This order also makes the jump from pilot to plant-wide easier: on the second line you are deploying a proven template.
The Production Data Layer with MIS Otomasyon
In production monitoring projects we build the shop-floor side: machine connectivity, the data collection architecture, OEE dashboards, and integration of the quality data produced by machine vision systems into that flow. The goal is not to replace your ERP but to feed it reliable data from the floor.
To discuss which data can be pulled from which of your machines, and which line should be first, get in touch with us.