The short answer
Machine loading compares the work assigned to a machine against the hours it actually has, expressed as a percentage. The load is the total operation hours of the work orders routed to a machine — quantity times standard cycle time, plus setting time. The capacity is the realistic hours the machine can run in the period. Load divided by capacity is the loading percentage: above 100% and the machine is overloaded; below and it has room. That single number, read across every machine, is how a planner turns an MRP plan into a schedule the floor can keep.
Capacity planning is the control half of production planning — the reason a plan is loadable and not just a wish. This guide sits under the pillar, what is production planning software?, follows on from how MRP works (which assumes capacity), and leads into production scheduling (which sequences within it). The product page is Machine loading & capacity.
How load is calculated
Load comes from the work orders MRP raised and the process sheet / routing attached to each. A routing lists the operations a part goes through, the machine or work centre each runs on, and the standard times for each: the cycle time per piece and the setting (setup) time per batch. The load a work order places on a machine is straightforward:
Two things make load realistic. First, setting time is counted per batch, not per piece, so splitting an order into small batches inflates load with repeated setups — a real planning trade-off. Second, load lands on the specific machine each operation is routed to, which is why routing accuracy matters: a wrong machine on a routing loads the wrong resource and hides the true bottleneck.
How available capacity is calculated
Capacity is not calendar time. It starts there — shifts per day × hours per shift × working days in the period — and is then discounted for everything that stops a machine from producing:
- Planned downtime. Maintenance windows, tea and meal breaks, and planned changeovers come off the top.
- Availability losses. Breakdowns and material waits mean the machine is not running whenever it is staffed; an availability factor accounts for it.
- Efficiency / performance. A machine rarely runs at exactly its theoretical rate; an efficiency factor scales the theoretical hours to what is really achievable.
The result is effective available capacity — the hours the machine can genuinely produce in the period. Skip the discount and you plan against calendar hours the machine will never deliver, which is precisely how overloads stay hidden until a job is late. The availability and efficiency factors are not guesses when the floor books actual run and downtime; they are measured, which ties capacity planning directly to the efficiency and OEE dashboards.
The loading percentage
With load and capacity in the same units — hours per period per machine — the comparison is one division:
| Machine | Load (hrs) | Capacity (hrs) | Loading % | Read as |
|---|---|---|---|---|
| CNC-1 | 92 | 80 | 115% | Overloaded — bottleneck |
| CNC-2 | 64 | 80 | 80% | Room to absorb work |
| VMC-1 | 78 | 80 | 98% | Full — watch closely |
Read across a week, this table is a plan's early-warning system. CNC-1 at 115% will miss something unless work moves; CNC-2 at 80% can take it. The whole craft of capacity planning is spotting that imbalance before the week starts and acting on it — which is only possible if the loading is finite.
Finite versus infinite loading
The distinction that separates a real capacity plan from a pretty chart is finite versus infinite. Infinite loading assumes a machine can absorb any amount of work — it will happily show 150% and schedule the job anyway, because it treats capacity as unlimited. Useful for a rough forecast; dangerous for a committed week. Finite loading respects the machine's available hours and treats anything over 100% as an exception to be resolved, not a number to be ignored.
The practical consequence is that finite loading forces the overload into the open, where a planner can do something about it, instead of burying it in an optimistic schedule that the floor silently fails to meet. This is the same finite principle that governs Gantt scheduling: the schedule may only place as much work on a machine as the machine can actually run.
Bottlenecks and leveling
Every plant has a bottleneck — the resource whose capacity is lowest relative to the demand on it, the machine that sits at 115% while others have slack. Because throughput is capped by the bottleneck, capacity planning is largely the discipline of finding it, protecting it, and offloading it where you can. A loading report read against pending work surfaces the bottleneck early, so it is managed rather than discovered on the day the job ships late.
Once the overload is visible, load leveling resolves it — smoothing work until every machine is at or under 100% for each period:
- Move work to an alternate machine where an operation can run on more than one resource.
- Shift non-urgent work to an earlier or later period with spare capacity.
- Re-sequence by priority so the tightest due dates get the bottleneck hours first.
- Add capacity — an extra shift, overtime, or an outsourced (OSL) operation — when levelling within the plant is not enough.
Rough-cut versus detailed capacity
Capacity is checked at two grains. Rough-cut capacity planning takes the master plan and validates it against a few key resources — is the plan even plausible for the bottleneck? — before committing to detail. Detailed capacity planning then loads every operation of every released work order onto every machine, period by period. The rough-cut pass catches gross infeasibility cheaply; the detailed pass produces the loading percentages a scheduler acts on. A capable system does both, so planners are not surprised late by an overload a rough-cut check would have flagged early.
Utilisation, honestly measured
It is tempting to chase a utilisation target, but the honest figure is measured, not assumed. Once setups, breakdowns, material waits and idle time are counted, useful utilisation on many Indian discrete-manufacturing machines sits well below the theoretical maximum — and that is not a failure to hide but a map of where the hours go. The value of a loading and utilisation view is not a number to boast about; it is knowing that CNC-1 loses six hours a week to setups and four to material waits, so the biggest loss can be attacked first. Measuring load against real available capacity, and tracking actual availability through shop-floor booking, is what makes utilisation a lever rather than a slogan — which is why loading connects directly to plan-vs-actual and OEE.
How Fast Planning loads machines
Fast Planning Software — built by Improsys in Pune under the Fast Technology brand, cloud or on-premise — loads work orders onto finite machine capacity directly from the routing standard times, so daily load, percentage loading and projected availability are visible per machine before the week is committed. Overloads surface as loading above 100%, and work is levelled by moving operations, re-sequencing by order and resource priority on the DayPilot Gantt board, or adding a shift. Because loading, MRP and the shop-floor booking that feeds availability all run on the same shared platform, the capacity a planner loads against is the capacity the floor actually has — measured through efficiency and OEE, not assumed. The result is a plan whose quantities, from MRP, are guaranteed loadable onto the machines you really run.
Frequently asked questions
What is machine loading?
Machine loading is the comparison of the work assigned to a machine against its available capacity for the same period. The load is the total hours of the operations from the work orders routed to that machine — quantity multiplied by standard cycle time, plus setting time. The capacity is the hours the machine is actually available in the period, from its shifts. Expressed as a percentage, load divided by capacity shows which machines are overloaded (above 100%) and which have room, so a planner can level work before a due date is missed.
How is machine capacity calculated?
Available machine capacity for a period starts from calendar time — shifts per day multiplied by hours per shift multiplied by working days — and is then discounted for reality. Planned downtime for maintenance, setups and breaks reduces it, and an availability or efficiency factor accounts for the machine not running at 100% of its theoretical rate. The result is the realistic hours the machine can actually produce in the period, which is what load is compared against. Using calendar hours without the discount overstates capacity and hides overloads.
What is the difference between finite and infinite loading?
Infinite loading assumes a machine can absorb any amount of work — it schedules as if capacity were unlimited, which is fine for a rough forecast but not for a real week. Finite loading respects each machine's available hours and refuses to load more than that, so an overload shows up as a loading percentage above 100% instead of a silent promise the floor cannot keep. Finite loading is what keeps a schedule realistic and lets planners move work to another machine, add a shift, or re-sequence before the date slips.
What is a bottleneck in capacity planning?
A bottleneck is the resource whose capacity is the lowest relative to the demand placed on it — the machine that runs well above 100% loading while others have room. Because a plant can only produce as fast as its bottleneck, capacity planning is largely about finding it, protecting it and, where possible, offloading work from it. A machine loading report read against pending work is exactly how the bottleneck is spotted early, so it can be managed rather than discovered when the job is already late.
What is load leveling?
Load leveling is smoothing work across machines and time so no single resource is overloaded while others sit idle. When a machine loading report shows one machine above capacity, leveling moves eligible operations to an alternate machine, shifts non-urgent work to a later period, re-sequences by priority, or adds a shift — until every resource is at or under 100% for each period. Leveling is what converts an over-optimistic MRP plan into a schedule the shop floor can actually run to date.
What machine utilisation is realistic?
Realistic utilisation depends on the process and the mix, and it is measured, not assumed. Many Indian discrete-manufacturing shops run useful machine utilisation well below the theoretical maximum once setups, breakdowns, material waits and idle time are counted — which is exactly why measuring load against real available capacity, and tracking actual availability, matters more than quoting a target number. The value of a loading report is not a benchmark to hit but visibility of where the hours actually go, so the biggest losses can be attacked first.