Production schedule: sequencing the work you can actually do

A production schedule says what will be made, on which equipment, in what order, and when. It is the point where a list of customer orders meets the number of machines, people and hours that actually exist. Most schedules fail not because the sequencing was wrong but because they were built on capacity nobody had verified — a line assumed to run three shifts when it runs two, or a changeover assumed to take twenty minutes when it takes ninety.

What a production schedule needs to exist

• Firm demand separated from forecast, because scheduling against a forecast produces stock rather than sales.
• Real capacity per resource: hours available after planned maintenance, breaks and known absence, not nameplate capacity.
• Changeover times between products, which is where the sequence stops being arbitrary — grouping similar products can free a shift a week.
• Material availability with lead times, since a schedule that assumes components arrive on time is a wish list.
• Batch and lot sizes, including any minimum imposed by the process rather than by the order.
• Constraints that are not equipment: a single qualified operator, a shared tool, a curing time that cannot be compressed.

Sequencing decisions that actually matter

1. Identify the bottleneck resource first and schedule it to full utilisation; scheduling everything else at once is how nothing gets scheduled.
2. Group by changeover characteristics — colour, material, tooling — rather than strictly by due date, then check no order breaches its promise.
3. Leave slack deliberately, typically ten to twenty per cent, because a fully loaded schedule cannot absorb a single breakdown.
4. Freeze the near horizon: the next few days are fixed, the following weeks are firm, beyond that is planned. Rescheduling tomorrow every morning destroys the shop floor’s trust in the plan.
5. Publish it where the people running it can see it, in the form they will actually use.

A schedule nobody can achieve is worse than no schedule, because it stops being a commitment and becomes a suggestion — and once that happens the sequence gets decided informally by whoever shouts loudest. The test is simple: if the schedule was met last week without heroics, it is a plan. If it required overtime and expediting, the capacity assumptions are wrong, not the team.

When a spreadsheet stops coping

• One or two production lines with a handful of products: a spreadsheet is genuinely the right tool, and most manufacturers of this size run one successfully for years.
• Multi-level bills of materials where sub-assemblies have to be scheduled before the parent — this is the point where MRP earns its cost.
• Frequent rescheduling driven by material shortages, which needs live stock and open purchase orders in the same system.
• Many shared resources with sequence-dependent changeovers, where the optimisation is genuinely beyond manual work.
• Traceability requirements linking each batch to materials and equipment, which belongs in a batch record rather than a schedule.

Sequencing by hand works until the constraints interact. Production scheduling software models finite capacity, sequence-dependent changeovers and shared tooling, which is where a plan on paper stops matching what the floor can run.

Not every late task threatens the finish date. The critical path is the chain with no spare time in it, and knowing which tasks are on it decides where attention is worth spending.

Where the schedule lives

Ettex Sheets is the right home while the schedule is a table: one row per order with resource, sequence, start and finish, alongside a capacity view that shows the load per line per week. Kept next to the inventory management data and the planned preventive maintenance calendar, it stops the two most common causes of a broken schedule — missing material and unplanned downtime on a machine that was actually due for service. Ettex is not an MRP or APS system: there is no material requirements explosion, no automatic rescheduling and no finite capacity optimiser. If you need those, buy them — and be honest that most companies buying MRP would have got further by verifying their capacity numbers first.