MES vs ERP: a question of boundaries, not competition

MES VS ERP: A QUESTION OF BOUNDARIES, NOT COMPETITION

This article takes a closer look at the fundamental differences between ERP and MES, not just in terms of technical design, but also in how they serve different users, respond to different time scales and support different business needs.

The technology landscape for manufacturing companies can quickly become complex, especially with IT/OT integration. The traditional automation pyramid can be challenged by new technologies and connections. It raises the question of whether MES and ERP still need to be separate systems. We increasingly see customers questioning the boundaries of both and considering embedding MES functions into their ERP. And vice versa, to a lesser extent.

When doing this exercise, it helps to remember that ERP and MES are designed with different audiences, purposes and requirements in mind.

 

TIMING & GRANULARITY OF INFORMATION AVAILABILITY

MES solutions are designed to run in seconds, ERP systems in minutes. If you seek to manage operational disruptions, anticipate and change schedules in real time, and track and prevent issues in production, you need a real-time view of what’s happening. An ERP often works with delays through overnight batch concepts, and its general performance design does not emphasize real-time data availability. This is a problem for shop floor visualization.

Similarly, the granularity of information in an MES can go much deeper. Take, for example, the palletizer stacking plan, which is controlled by the MES. In environments where production has to be flexible, with complicated changeovers and a rich diversity of products and recipes, an MES provides tools and granularity to automate and control these processes. The higher abstraction layer and lower response time of an ERP make this much harder to manage in an ERP system.

A mid-size chocolate factory

A chocolate factory employing 150 people illustrates how this plays out in practice. Their ERP handled financials, but couldn’t provide the responsiveness needed to guarantee efficiency and quality. They implemented MES alongside ERP, linking MES to production machines to register real-time values. This not only gave them visibility into the actual situation on the floor, but also allowed setup times to be shortened by preemptively starting certain actions during the previous production run.

By moving from time-based to count-based quality testing, they also reduced unnecessary testing. Operators were alerted via smartphone when a test was truly needed, aligning quality assurance with actual production activity.

TENTACLES IN THE FIELD

MES systems are designed to interconnect with the field and they often support dozens of protocols such as Modbus, OPC-UA and MQTT. They are built to be flexible, to adapt to and manage the underlying physical process rather than abstract it away. It does the heavy lifting of aggregating information, ironing out various edge cases and ensuring that a consistent, simplified representation can be reported back to ERP or other layers.

Customizing ERP solutions is often more costly and less required. Business processes are more easily adapted, abstracted or virtualized than physical ones. Embedding MES into your ERP will inevitably require painful customization or stark abstraction, often leading to a shadowy middle layer of Excel sheets, paper trails and post-its.

 

DESIGN TO LAST

The criticality of an ERP and an MES is vastly different. Downtime of an MES system usually directly affects production. You don’t measure the loss in thousands of euros, but rather in percentages of revenue — often dwarfing any implementation and licensing costs incurred. An ERP, on the other hand, has a high requirement for data integrity, but the cookies still get baked, even if the ERP is down for two days.

This reflects not only in the technology design principles, but also in its implementation. MES implementation partners are geared up for fast deployments overnight or on weekends to minimize downtime. They have robust 24/7 support mechanisms and extensive testing procedures to ensure uptime. Placing such robustness requirements on an ERP is often unnecessary and partners are often not equipped to meet them.

 

BEYOND THE AUTOMATION PYRAMID

While I believe that in most scenarios a separation of responsibilities between MES and ERP is advised, it’s not a dogma. In some cases a well-scoped investment in a single technology that adopts both ERP and MES functionality can result in low cost of ownership and critically, higher ROI. The simplicity of managing only one heavy business application can be beneficial, in particular for small organizations.

a small cheese producer

A small cheese factory with 15 employees offers a different perspective. They produce a single local product in three sizes. The process is straightforward, with little variation, and the ripening process doesn’t require real-time responsiveness. For them, MES would offer limited added value in daily operations. However, a lightweight data platform allows them to document process compliance and traceability, which is valuable during food safety inspections. It’s a simple but effective step toward digitalization, tailored to their scale and regulatory needs.

Technological evolutions such as IIoT connectivity, Unified Name Spaces (UNS), modular web-based platforms, low-code solutions and data flow tools can bridge some of the gaps that a MES used to close. The traditional automation pyramid can be circumvented. A DCS system could push its counters, asset states and batch information to an UNS and then to an MQTT broker. From there, a simple workflow in Node-RED or Ignition could be triggered to aggregate this information and post-process it before pushing the resulting batch record back into the UNS, which can then be picked up by the ERP.

 

MES, ERP… OR SOMETHING IN BETWEEN?

By now it’s clear that there is no one-size-fits-all. The trade-off on which responsibility to assign to which system must be made carefully. Disruption of the traditional automation pyramid can quickly spiral out of control with a swathe of solutions that become unwieldy to manage.

As with all investments, I’m a firm advocate for a think big, start small, scale fast approach. Companies should first articulate their North Star: where do they want their digital future to lie? From there, you decide which technology choices give you the best ability to achieve this ambition.

Next steps

Regardless of where on the spectrum you land, digitalizing operations is non-optional. The benefits in terms of control, efficiency, safety and compliance are second to none.