Implementation is where most of the value — and most of the risk — sits. The software decision takes weeks. Living with the consequences takes years.
A workable sequence
- Define the planning process you intend to run, before configuring anything.
- Clean and load item master, BOMs and routings.
- Establish inventory accuracy and a transaction discipline the floor will follow.
- Configure order policies, lead times and safety stock deliberately, not by default.
- Run a conference-room pilot with real part numbers and real orders.
- Train planners on the logic, not just the screens.
- Cut over on a defined date with a fallback plan.
Data preparation is the project
In small plants, sixty to eighty percent of implementation effort is data. Item numbering, unit-of-measure consistency, BOM structure, routing steps, lead times and current inventory all have to be correct at the same moment.
Treat data as a workstream with an owner and a schedule, not as a task that happens alongside everything else.
The first ninety days after go-live
Expect the system to expose problems that were previously absorbed by people. Shortage messages will be noisy, lead times will need correction, and some BOMs will be wrong in ways nobody knew.
Plan for a structured stabilization period: a daily exception review, a fast path to correct data errors, and a named person who owns the answer when the system and the floor disagree.
Common questions
- Should we run parallel with our old process?
- For a short, defined period on a limited scope, yes. Long parallel runs are a warning sign: they usually mean nobody trusts the new data and the old process quietly wins.