The most expensive sentence in manufacturing is “the system will handle it.” I’ve heard it from plant managers, operations directors, and once from a CEO who had never walked the shop floor on a Thursday afternoon.
Here’s what that sentence usually means: we don’t trust our own data, so we’re hoping software will somehow fix that. It won’t.
Your bill of materials is probably fiction
A $32 million job shop we worked with had 1,900 active assemblies. When we checked, roughly 60% had at least one error — a component that hadn’t been used in two years, a quantity that reflected an old design revision, or a phantom part that existed only in the system.
Scheduling on top of that data doesn’t produce a plan. It produces a confident-looking document that everyone ignores by Wednesday.
Fix the 20% that drives the schedule
You don’t need to correct every assembly before you see results. You need the ones that actually constrain your week.
- Pick the top 30 assemblies by revenue or by hours consumed. That’s usually 15–20% of the list.
- Walk each one on the floor with the supervisor and the person who builds it. Ask what they actually take off the shelf.
- Record real scrap percentages instead of the 2% default nobody has questioned since 2019.
- Set lead times from what suppliers actually deliver, not what they promised when you signed the contract.
- Lock routings so two people don’t run the same job two different ways.
Two numbers tell you whether it’s working
Track schedule adherence and unplanned overtime. The job shop started at 61% adherence with 340 hours of unplanned overtime a month. Eight months later: 88% adherence and 190 hours.
Notice what didn’t change. Same system. Same machines. Same headcount. What changed was that the plan was finally built on data that matched the building.
This is where ERP Consulting earns its keep: not in configuring modules, but in forcing the uncomfortable conversation about which numbers are real. A plan built on assumptions is worse than no plan, because it creates false confidence across the whole business.
Pick one assembly tomorrow. Walk it with the supervisor and write down every difference between the paperwork and the parts. Do five of those and you’ll know exactly what needs fixing.