Key takeaways
See our roundup of production monitoring systems that often sit alongside an MES.
A manufacturing execution system (MES) is shop-floor software that manages, tracks, and documents production in real time, translating ERP plans into executed work. In the ISA-95 hierarchy it occupies Level 3, sitting between business planning (ERP, Level 4) and machine control such as SCADA and PLCs (Levels 1 to 2).
A manufacturing execution system (MES) is software that manages, monitors, and documents production as it happens on the shop floor. It takes the production plan from the ERP system, turns it into executable instructions for operators and machines, then collects the resulting data, including quantities, quality, scrap, downtime, and genealogy, and feeds it back as a verified production record.
The standards-based way to place MES is the ISA-95 automation hierarchy, which organizes a plant into five levels. According to Siemens, Level 3 "defines the activities of workflow to produce the desired end products" and is where Manufacturing Operations Management systems such as MES reside, while Level 4 holds "business logistics systems (e.g. ERP)." Below MES, Levels 1 to 2 cover sensing and machine control via PLCs and SCADA, and Level 0 is the physical process itself. (Siemens, ISA-95 framework)
In plain terms: ERP decides what and when to make, MES runs and records how it actually gets made, and SCADA/PLCs control the machines that make it. MES is the bridge between the office and the floor.
The most widely cited definition of MES scope is the MESA-11 model, published by the Manufacturing Enterprise Solutions Association in 1996, which defines a healthy MES by 11 core functions. Per MaintainX's standards reference, those functions are:
You do not need all 11 in every plant. A discrete machining shop leans on scheduling, data collection, and performance analysis, while a regulated pharma or food line lives or dies by traceability, quality, and document control. The MESA-11 model is a menu of capabilities, and ISA-95 tells you where those capabilities sit in the wider stack. (MaintainX, MES standards)
The fastest way to understand MES is to compare it with the systems it is constantly confused with. Each operates at a different layer, time horizon, and question.
| System | Layer / question it answers | Time horizon | Typical owner |
|---|---|---|---|
| ERP | Level 4. What should we make, for whom, with what cost and inventory? | Months to days | Finance, planning, supply chain |
| MES | Level 3. How is the order actually being executed and recorded right now? | Days, shifts, minutes | Production / operations |
| Real-time OEE | An MES performance-analysis function. How effectively is each machine running versus its potential? | Live, by cycle | Production / continuous improvement |
| CMMS | Maintenance system of record. What asset failed, what is the work order, and what is the maintenance history? | Live to long-term | Maintenance / reliability |
| SCADA / PLC | Levels 1 to 2. Sense and control the physical machine in real time. | Seconds to milliseconds | Automation / controls |
The key takeaways from this table:
They overlap almost completely on the measurement layer, and that surprises people who shop for them as separate products. An MES collects run time, stop time, cycle counts, and reject counts; an OEE engine consumes exactly those signals to compute availability, performance, and quality.
The practical difference is emphasis. A traditional MES treats OEE as one report among many alongside scheduling, traceability, and document control. A modern real-time OEE platform makes that single number the heartbeat of the shop floor and pairs it with loss analysis such as the six big losses framework. When a vendor sells "real-time OEE," they are usually selling the data-collection and performance-analysis functions of MESA-11, the part of MES that most plants feel the absence of first. The biggest gap in either approach is the same: knowing a machine stopped is not the same as knowing why, and getting it fixed.
You likely need MES-class capability when paper, spreadsheets, and ERP can no longer answer floor-level questions fast or accurately enough. Use this checklist; if you tick several, the case is strong.
Smaller plants often start with the highest-pain function, usually real-time OEE and downtime capture, then expand. That staged path is also the safest way to prove value before a full deployment. For the wider context of where this fits, see our overview of the smart factory.
The benefits all trace back to one thing: replacing after-the-fact paperwork with a live, trustworthy record of production. In practice that means:
The recurring weakness in the classic stack is the seam between systems. The MES (or OEE tool) sees that a machine stopped, but the CMMS that owns the repair is a separate product, so the handoff is manual and the true cause is often lost.
Fabrico unifies real-time OEE and MES-style production tracking with a full CMMS in a single platform. It connects to machine PLCs to capture OEE and cycle times, uses computer vision to record the true cause of a stoppage rather than a generic code, and turns that fault into a prioritized, parts-ready digital work order on the technician's phone with QR-enforced checklists. That is the fault-to-fix loop in one system rather than three.
Because Fabrico is EU-built (headquartered in Bulgaria), it is also a clean fit for teams that need EU data residency. Pairing live production data with maintenance also strengthens reliability metrics like MTBF and MTTR and supports a structured total productive maintenance program. [INSERT VERIFIED PROOF POINT - operator to confirm]
If you want to see real-time OEE, downtime true-cause, and automatic work orders working as one loop, book a Fabrico demo and bring one problem line to walk through.
No. In the ISA-95 hierarchy, ERP is the Level 4 business-planning layer that decides what to make, for whom, and at what cost, while MES is the Level 3 execution layer that runs and records how the work is actually performed on the floor. ERP plans the order; MES executes it and reports back verified production, scrap, and quality data.
SCADA operates at ISA-95 Levels 1 to 2 and controls physical machines in real time through PLCs and sensors, reacting in seconds. MES sits above it at Level 3 and adds context, turning a raw machine signal such as a stopped motor into a production record like a 14-minute downtime event against a specific order, then managing scheduling, quality, and traceability around it.
Yes. Real-time overall equipment effectiveness (OEE) is part of the performance-analysis function in the MESA-11 model. An MES collects run time, stop time, cycle counts, and reject counts, which are exactly the inputs used to calculate availability, performance, and quality. Many real-time OEE products are essentially the data-collection and performance-analysis functions of an MES sold as a focused tool.
The MESA-11 model defines an MES by 11 core functions: operations and detailed scheduling, resource allocation and status, dispatching production units, data collection and acquisition, product tracking and genealogy, performance analysis, quality management, process management, maintenance management, labor management, and document control. Plants typically adopt the functions that address their biggest pain first.
Traditionally yes, because MES manages production execution while a CMMS manages maintenance work orders, spare parts, and asset history, and they are usually separate products. That separation creates a gap: the system that detects a fault is not the one that fixes it. Fabrico closes that gap by unifying real-time OEE and MES-style tracking with a full CMMS so a detected fault becomes a parts-ready work order automatically.
A plant needs MES-class capability when ERP, paper, and spreadsheets can no longer answer floor-level questions fast enough. Strong signals include being unable to see live production versus plan, reconstructing downtime causes after the shift, needing lot or serial traceability for audits or recalls, and OEE numbers that are already stale by the time anyone reviews them. Many plants start with real-time OEE and downtime capture, then expand.