Machine capacity planning explained: load vs capacity and % loading

Ask a shop when a new order can ship and the honest answer is often a shrug and a guess. The machines look busy, the planner has a feel for it, and the date that comes out is a hopeful round number. Two weeks later the order is late — not because anyone worked slowly, but because the machine it needed was quietly committed to three other jobs the whole time. Machine capacity planning replaces that shrug with two numbers per machine, and a percentage that makes the answer obvious.

This guide is for planners and production managers who schedule work across shared machines and want to promise dates they can keep. It covers what capacity planning is, how load and capacity are calculated, how to read loading percentage, and how Fast Planning Software shows it. Every number below is illustrative — used to make the arithmetic clear, not drawn from any specific plant. For the wider picture, start with the pillar guide, what is production planning software?

1. What machine capacity planning actually is

Machine capacity planning is the practice of comparing, for each machine or work centre, the work assigned to it against the work it can do in a period — so a plant knows whether its plan actually fits the machines it has, before it commits to dates. It answers the question every scheduler eventually hits: we have the orders and the material, but can the machines physically do this in the time we have?

It rests on two numbers, and everything else is derived from them:

• Load — the total hours of work assigned to a machine for the period. It comes from the operations routed to it: each operation's standard cycle time multiplied by the quantity, plus its setting (changeover) time.
• Capacity — the hours the machine actually has for the period. It comes from the shifts it runs, less planned downtime such as maintenance.

Compare the two — usually as a loading percentage — and the plant can see which machines are the constraint and which have room, while there is still time to do something about it. Skip it, and overloads only reveal themselves as missed dates. See Machine Loading & Capacity.

2. Load vs capacity — the two numbers

Getting these two right is the whole game, and each hides a common mistake.

Two subtleties matter. First, setting time is part of load: a machine that changes over ten times a day can lose hours to setup that never appear if you count only cycle time — which is why grouping similar jobs to share a setup frees real capacity. Second, capacity is not the clock: a single 8-hour shift is not 8 hours of available machining once you subtract planned maintenance, tool changes and breaks. Capacity planning is only as honest as these two inputs; the standard cycle and setting times captured on each process sheet are what make the load figure trustworthy.

3. Calculating machine loading percentage (worked example)

Loading percentage is simply load ÷ capacity × 100. The value is in doing it per machine, so the comparison is like-for-like. Here is a worked example across four machines for one week — all figures illustrative:

The percentage does the work a raw hours figure cannot. Turning centre A at 115% is trying to do 46 hours of work in a 40-hour week — it is 6 hours short and will make something late unless the plan changes. Turning centre D at 60% has 16 spare hours. If A and D can run the same operations, moving roughly 6 hours of work from A to D brings A down to about 100% and lifts D to about 75% — the plant's total output is unchanged, but now every machine can keep the plan. That single move is the essence of capacity planning, and we will return to it under load levelling.

4. Finding the bottleneck

A bottleneck is the machine whose capacity constrains the whole plant's output — the one loaded well over 100% while others have room. It matters out of all proportion to its size, for a simple reason: the plant can only ship as fast as its bottleneck runs. Everything about capacity planning eventually points at finding it early.

• The bottleneck sets the realistic date. If turning centre A is the constraint, no promised date is credible unless A has the hours — the other machines waiting on A are not the limit.
• Time lost on the bottleneck is lost for the plant. An hour of breakdown or a slow setup on A is an hour of finished output the whole line never gets back; the same hour lost on an under-loaded machine costs nothing.
• The bottleneck can move. Relieve A and the constraint may shift to milling B at 95%. Capacity planning is continuous because the bottleneck is not fixed.

A machine loading report that ranks machines by loading percentage makes the bottleneck the top row — visible before the schedule is committed, not discovered when the order is already late.

Ranked by loading percentage, the bottleneck is the top bar crossing the 100% line — visible before the week starts. Figures shown are illustrative.

5. Projected availability and pending work

Loading percentage tells you a machine is full; projected availability tells you when it will next be free. It is the forecast built from the pending work already loaded onto a machine — read against a loading report, it is what turns a promised date from a guess into a defensible commitment.

The logic is straightforward. If turning centre A is loaded solid for the next six working days, an operation that needs A cannot start until day seven, however urgent it is — unless something already on A is pushed aside. So when a customer asks for a date, the honest answer is not "the machine is free now" but "the machine has this much pending work, so the earliest this operation can start is here." Reading load against pending work lets a planner:

• Promise a real date — based on when the needed machine actually clears, not on optimism.
• Decide whether to expedite — see exactly what would have to move to pull an urgent order forward, and at what cost to other jobs.
• Spot forward overloads — a machine that is fine this week but solidly booked three weeks out, while there is still time to add capacity.

This is why a machine loading report is read two ways: current load for this week's bottleneck, and pending load for the promise you make on the next order.

6. Load levelling across machines and shifts

Finding an overload is only half the job; load levelling is fixing it — spreading work so no machine drowns while another idles. There are three levers, in rough order of preference:

7. Common traps that hide an overload

Capacity planning fails quietly, in predictable ways. Watch for these:

8. How Fast Planning Software implements capacity planning

Fast Planning Software is the MRP and production-planning product of the Fast Suite, built in Pune by Improsys under the Fast Technology brand, deployable cloud or on-premise for manufacturers across India and worldwide. It turns the arithmetic above into live screens:

9. Frequently asked questions