Production scheduling software: build, buy, or fix the spreadsheet

Finite vs infinite capacity scheduling

Infinite capacity scheduling assumes any machine can run any job at any time - it produces a plan, but not necessarily a realistic one. Finite capacity scheduling respects actual limits: how many hours a machine has, how many people are qualified to run it, and what else is already booked in.

Most ERP scheduling modules are infinite capacity by default and are sold as if that's a minor limitation. It isn't. A plan that ignores capacity is a wish list, and the shop floor knows it within the first shift.

The constraints that actually matter

Generic schedulers are built around the constraints that are easy to model: machine hours and due dates. Real plants are usually constrained by something messier that the software vendor never asked about.

• Changeover time and sequence - some product orders minimise cleaning or tooling changes; others don't. A scheduler that ignores sequence-dependent setup will happily plan the worst possible order.
• Skilled labour, not just machine hours - a machine free at 2pm is no use if the only qualified operator is on a different line.
• Shared tooling or fixtures - two machines can't both run a job needing the one mould you own.
• Material readiness - a schedule that assumes the goods-in delivery arrived on time is a schedule for a different plant.
• Downtime and breakdowns - see our piece on the types of manufacturing software for how maintenance data should feed the schedule, not sit apart from it.

Sequencing rules worth naming explicitly

Before evaluating any scheduling system, write down the sequencing rules your planners already apply from memory. These are usually things like 'run light colours before dark' or 'never split a batch across two shifts'. If a rule isn't written down, no software will apply it consistently, however good the algorithm is.

1. 01List every constraint your current planner works around manually.
2. 02Rank them by how often they change the plan versus how rarely they're mentioned.
3. 03Check whether a candidate system can express each rule, not just accept it as a comment.
4. 04Test the system against last month's actual orders, not a clean demo dataset.

When is a better spreadsheet the right answer?

If your plant has a handful of machines, a small number of sequencing rules, and one person who understands them, a well-built spreadsheet with proper constraint logic can outperform an expensive scheduling system for years. The honest failure mode of spreadsheets is not the tool - it's that the logic lives in one person's head and the file breaks when they're off sick.

The trigger to move beyond a spreadsheet is usually scale: more than a few machines with interacting constraints, multiple planners needing the same live view, or a schedule that needs to react automatically when a machine goes down. Below that threshold, spend the money on documenting the logic properly instead.

Build, buy, or fix - how to decide

Off-the-shelf finite-capacity schedulers are a reasonable starting point if your constraints are close to standard: machine hours, due dates, and simple changeover rules. Where they struggle is anything plant-specific - shared tooling, unusual shift patterns, or sequencing rules with exceptions.

That's the point at which a bespoke or heavily customised scheduling tool starts to pay for itself, because you are no longer paying to work around a product's assumptions every week. If you're weighing that trade-off generally, our guide to custom software vs off-the-shelf is a useful starting point before you talk to any vendor.

We build scheduling tools as part of our wider manufacturing software work, usually once a manufacturer has proved to themselves that the spreadsheet has hit its limit. If that's where you are, get in touch with what your current schedule actually looks like - not the tidy version.