An MES proposal can make the future look straightforward: production orders flow to the floor, operators capture the right information, and supervisors see what is happening. The difficult part is deciding exactly how that will work in your plant.
Before approving a manufacturing execution system project, settle the questions that will determine the scope. A good demonstration is useful. It cannot tell you whether the underlying process, interfaces, and responsibilities have been worked out.
1. Which operating problem are we solving?
“We need an MES” names a solution. Be more specific about the need. Are production records taking too long to review? Is material genealogy incomplete? Are operators entering the same information twice? Is scheduling disconnected from actual production?
Choose the priority outcomes and define how you will measure them. Record the baseline, the affected process, and who owns the result. If several departments want different things, resolve the priorities before making every request part of phase one.
2. What belongs in the first phase?
Draw a boundary around the initial release: site, line, product family, shift, and workflows. State what it will include and what it will leave for later. An initial scope that includes production tracking does not automatically include maintenance management, advanced scheduling, quality release, and warehouse execution.
ISA-95 provides terminology and models for enterprise-control system integration. It is a useful reference when separating business-system responsibilities from manufacturing operations and identifying the information that must pass between them.
3. Who owns each interface?
List the information the MES needs from ERP, equipment, quality, warehouse, and other systems. For every interface, define the source, destination, frequency, owner, and failure response.
Ask practical questions: What happens when an order changes after production starts? Who corrects a material master? How does the line continue during a network interruption? Who investigates a transaction that never arrives? These decisions belong in the scope and estimate.
4. Can the plant support the new way of working?
Walk through the proposed workflow with operators, supervisors, maintenance, quality, engineering, and IT/OT. Check the actual devices, network access, user accounts, data definitions, work instructions, and support arrangements.
A screen that looks simple in a meeting can be awkward at a machine. Consider gloves, cleaning requirements, screen location, changeovers, and the time available for data entry. Make sure the project includes process changes and training, not only configuration.
5. Does the estimate include the whole job?
Separate software costs from implementation, interfaces, infrastructure, testing, validation where applicable, training, internal labor, and ongoing support. Identify which costs are fixed, estimated, recurring, or dependent on unresolved assumptions.
Compare proposals against the same requirements. A lower price is hard to interpret when one proposal includes integration and another leaves it to the customer.
6. What must be proven before go-live?
Write acceptance criteria before implementation begins. Test complete workflows, including exceptions, recovery, user permissions, and performance under the expected load. Agree on who accepts each result and which unresolved issues prevent release.
Also define the cutover window, fallback arrangements, support coverage, and the conditions for extending the solution to another line or site. A successful first release should produce evidence for the next decision.