About the plant
I run the production floor at a small contract manufacturer. A fixed set of machines, two shifts, and a weekly schedule that has to hit a string of customer due dates. We built that schedule every Friday on a whiteboard, backed by a spreadsheet. By Friday afternoon it always looked clean.
The problem: the schedule was right until Monday at ten
Then the week started.
A spindle faults at 10 am Monday. A material delivery shows up a day late. A good customer calls with a rush job we can't really say no to.
Any one of those meant rebuilding the schedule by hand. Which jobs were on that machine? Which other machines can run those operations? What gets bumped if we move them? Which customer gets the phone call? Our planner would disappear into that for most of an afternoon. They were working from memory of each job's routing, too, so the new plan was only as good as what they could recall under pressure.
And while they rebuilt, the floor kept moving. Sometimes a second problem landed before the first fix was even finished.
The worst part wasn't the rebuild itself. It was finding out late. We'd learn on Thursday that an order due Friday wasn't going to ship, when the problem had really started on Monday. A customer can work with a heads-up on Monday. A surprise on Thursday is a different conversation.
What we changed: re-planning that takes minutes
We built the schedule on real data instead of memory: every job's routing, which machines can run which operations, setup times, and due dates. When a machine is marked down, the system shows every job it affects and proposes a new sequence across the machines that can take the work. Each job shows its new projected finish against its due date.
The planner reviews that proposal, adjusts it, and publishes it. The floor sees the new sequence at the stations.
We kept the decision with a person on purpose. The software doesn't know that one customer will happily take a partial shipment, or that only one operator really runs a particular fixture well. The planner knows those things. What changed is that they start from a complete proposal instead of a blank whiteboard.
Rush orders got easier too. "Can we take it?" used to be a gut call. Now we drop it into the schedule, see exactly what moves, and answer the customer with something better than a guess.
The result: a bad hour stays a bad hour
A machine going down still costs the hours it's down. It just doesn't cost the planner's whole afternoon on top of that. Orders that are genuinely at risk get flagged the same morning, so the customer call happens early, while there are still options to offer.
It also showed us something uncomfortable. A good share of our "breakdowns" weren't machines at all. They were missing material, which is a problem with its own fix.
If one bad morning at your plant regularly turns into a lost week, the machine probably isn't the expensive part.
The rebuild is.
Tell us about the last time one breakdown cost you more than the repair did. We'll show you which part of that rebuild could have been done for you, and which part should stay with your planner.
Free workflow teardown
Pick the workflow that hurts. We'll map it with you in thirty minutes and show where the hours and the money are leaking.
- Free
- Reply within 24 hours
- No obligation