Measures of Capacity Planning Input and Output: Which to Use
Key takeaways
- Input measures count resources you control, such as labor hours, machine hours, and material availability, while output measures count results such as units produced and throughput.
- Measure capacity in inputs when demand is uncertain or work is heterogeneous, and in outputs when work is standardized and you must commit to a delivery rate.
- Convert input measures into output forecasts by dividing available input hours by the cycle time per unit at the constraint.
- Utilization is output divided by input, while efficiency is actual output over standard output, and 100% utilization at a bottleneck is a warning rather than a win.
- Test which measure limits total output by collecting both measure types at every process step before quoting a client engagement.
Measures of capacity planning input and output answer one question: do you count resources or results? Input measures count what you have available: labor hours, machine hours, material supply. Output measures count what comes out: units produced, throughput, utilization rates. Service operations and knowledge work are usually measured in inputs; manufacturing in outputs.
What are input measures in capacity planning?
Input measures describe the capacity you can supply: the resources you control before any work starts. They answer "how much could we do?" rather than "how much did we do?" Capacity planning commonly uses input measures when output is hard to count or demand is uncertain.
Labor hours and staffed hours
Staffed hours are the baseline for most service and knowledge-work settings. Take headcount, multiply by productive hours per person, subtract leave and admin overhead, and you get available labor hours per week. For a solo consultant scoping a client, this number is usually the first thing worth verifying, because it is the one clients most often overstate.
Machine hours and equipment availability
In manufacturing, machine hours capture capacity better than labor hours when equipment, not people, sets the pace. Planned downtime, changeovers, and maintenance windows all reduce available machine hours, so the useful figure is scheduled hours net of those losses.
Material and information availability as an input constraint
Inputs are not only labor and machines. Missing materials idle a fully staffed line; missing specifications idle a fully staffed team. When you scope a client's constraint, check whether the limiter is supply of work inputs rather than supply of labor or equipment. Our capacity planning tools guide covers how input measures feed into that scoping decision.
What are output measures in capacity planning?
Output measures describe what a process actually delivers: units produced, throughput, and utilization rate. They are counted after the work happens, which makes them the evidence base for scoping a constraint before you quote an engagement.
Units and throughput
Units produced is the raw count of completed work over a period: orders shipped, patients seen, reports closed. Throughput is the same count expressed as a rate, typically units per hour or per shift. On a manufacturing line, throughput is capped by the slowest station, not the average of all stations; a worked breakdown of this appears in our manufacturing-line capacity example.
Utilization rate as an output-side measure
Utilization rate divides actual output by demonstrated capacity, so a line running 620 units against a proven 800-unit shift runs 77.5% utilization. Treat it as an output measure because it depends on what the constraint produced, not on hours scheduled.
Lead time and backlog as lagging output signals
Lead time and backlog trail the constraint. When WIP piles up ahead of one station and lead time stretches, the backlog is telling you where the bottleneck sits before any stopwatch study confirms it.
When should you measure capacity in inputs versus outputs?
The decision rule: measure capacity in inputs when demand is uncertain or work is heterogeneous; measure it in outputs when work is standardized and you need to commit to a delivery rate. Inputs (labor hours, machine hours, available resource time) absorb variation. Outputs (units per week, throughput) are only trustworthy when every unit consumes roughly the same effort.
Measure inputs when work is variable
If each job differs (a consultancy, a workshop doing mixed repair work, an invoice-processing function where case complexity swings), output measures hide the real constraint. Plan in hours of constraint capacity against backlog, then convert to outputs per job. The invoice-processing capacity example shows how input-based scoping reveals the step that limits total throughput before you quote the fix.
Measure outputs when you must commit to a delivery rate
When work is standardized, commit in outputs. Set takt time as available production time divided by customer demand (Lean Enterprise Institute glossary), then size input capacity to hit it.
Which capacity measures apply to service businesses versus manufacturing?
The differentiator is what you can standardize. Manufacturing converts inputs to outputs at a predictable rate, so machine hours forecast units reliably. Service and knowledge work convert at variable rates depending on the request, so you forecast output from input measures with wider error bars.
| Operation type | Input measure to collect | Output measure to forecast | Conversion between them |
|---|---|---|---|
| Service | Staffed hours by role, appointment slots | Customers served per week, utilization rate | Slots × historical show rate ÷ average service cycle time |
| Manufacturing | Machine hours, material availability, labor hours | Units per shift, throughput at the bottleneck | Machine hours ÷ standard cycle time per unit |
| Knowledge work | Available person-hours, WIP limit | Completed tickets or projects per week | Person-hours ÷ average touch time per ticket |
Service example: a customer-onboarding team with 120 staffed hours a week and a 90-minute average cycle time forecasts roughly 80 onboardings, before rework. We walk through that arithmetic in our customer-onboarding capacity example.
The honest caveat: these conversions assume a stable constraint. If onboarding volume spikes or a machine goes down, the constraint moves and yesterday's ratio misleads you. Scope the constraint first, then trust the conversion.
Service operations
Collect staffed hours and appointment slots. Forecast customers served and utilization. Takt time works here only when demand arrives on a predictable cadence.
Manufacturing operations
Collect machine hours and material availability. Forecast units per shift and throughput at the bottleneck. This is the only row where the input-to-output conversion is tight enough to quote against with confidence.
Knowledge-work operations
Collect person-hours and WIP limits. Forecast completed tickets per week. Cycle time varies by ticket type, so segment the conversion or accept a wide range in your scoping report.
How do utilization and efficiency relate to input and output measures?
Utilization: output over input
Utilization is an output divided by an input: actual output hours over available input hours. If a machine runs 38 hours of a 40-hour week, utilization is 95%. The standard formula is available-time-based, as defined in standard operations management references. It tells you how much of an input resource was consumed, not whether that consumption produced good output.
Efficiency: actual over standard
Efficiency compares actual output to standard output. If the standard is 20 units per hour and the line produces 17, efficiency is 85%. Efficiency sits on the output side of the ledger; utilization sits on the input side. You can post 100% utilization at 60% efficiency and ship nothing useful.
Why 100% utilization at a bottleneck is a warning, not a win
A constraint running at 100% utilization has no buffer, so every hiccup downstream becomes lost throughput. That is why scoping should test which resource is the constraint before promising gains: the Constraint Impact Calculator shows whether relieving one resource actually lifts system output or just moves the bottleneck.
How do you convert input measures into output forecasts?
The basic conversion: hours divided by cycle time
Forecast output equals available input hours divided by the cycle time per unit. If a workstation takes 12 minutes per unit and you have 480 staffed minutes, the arithmetic gives 40 units. The same formula works for machine hours, labor hours, or any input measure, as long as cycle time is measured against the constraint.
Where takt time sets the target
Takt time is available production time divided by customer demand, and it tells you the cycle time you need to hit. If takt is 10 minutes and your actual cycle time is 12, the conversion shows a demand gap: you can supply 40 units against a requirement of 48. That gap, not the average utilization figure, is what scopes the engagement.
A worked example with real numbers
Say a team has 3 staff, 7.5 productive hours each, so 1,350 minutes per day. Cycle time at the bottleneck is 9 minutes, giving a forecast of 150 units. Demand is 170, so takt is 1,350 ÷ 170, or about 7.9 minutes. The constraint is 1.1 minutes per unit short, which points to where an improvement would raise total throughput.
What mistakes do teams make when mixing input and output capacity measures?
Comparing utilization across mismatched stations
A station running 85% utilization means nothing until you know its cycle time relative to the rest of the line. A five-minute station at 85% is idle money; a bottleneck at 85% is a queue factory. Always normalize utilization against takt time or demand before comparing stations side by side.
Forecasting output without checking the constraint
Converting input hours into output forecasts assumes every resource converts at the same rate. It does not. Run the forecast through the constraint first: if the bottleneck can only absorb 40 hours of work per week, promising output based on 60 hours of labor elsewhere is fiction. Simulation tools such as ProcessModel exist precisely to test whether a projected improvement moves the whole system or just one station.
Treating backlog as available capacity
Backlog is demand waiting, not capacity ready. A 3-week backlog tells you the demand gap, not how much more you can sell. Before quoting, confirm which measure actually limits total output, then scope the engagement around that constraint.
How do you test which measure limits total output before quoting a client engagement?
Collect both measure types in the first session
Record input measures (labor hours, machine hours, material availability) and output measures (units per day, utilization, throughput) for every process step. One without the other hides the demand gap.
Test the constraint before quoting
The scoping workflow that survives a client conversation has four steps. First, map the process steps and capture both measure types at each one, because the constraint only shows up when you compare capacity against demand at the same point; a station can look busy in isolation yet sit far below demand once you normalize the two measures against each other. Second, identify the step whose output measure sits furthest below demand, since that is your bottleneck candidate and the only station worth quoting against until the test proves otherwise. Third, run a small test: add an hour of labor, shift a machine changeover, or resequence WIP, and check whether system throughput moves or whether output stays flat because the constraint sits elsewhere in the line. Fourth, predict where the constraint migrates once you lift it, because a constraint rarely disappears, it moves to the next-scarcest resource, and your quote should account for that move.
If your test changes total output, you have a quotable engagement. If it does not, you have saved yourself and the client weeks of wasted effort, and you can scope the real constraint instead.
Where tools fit, including the honest limits of each
The Constraint Impact Calculator on Improve First runs steps two through four quickly and flags where the constraint moves next, but it will not simulate detailed shift schedules. DO.Impact, ProcessModel, and FlowVisual cover simulation depth to varying degrees; a spreadsheet capacity model remains the fastest first pass, and KaiNexus suits ongoing improvement programs rather than scoping.
FAQ
Is capacity planning counted in units or hours?
Both, depending on the operation. Manufacturing usually counts outputs: units per shift and throughput at the bottleneck. Service and knowledge work count inputs: staffed hours, machine hours, or available person-hours, because output per hour varies too much by job to forecast reliably.
What is the difference between design capacity and effective capacity?
Design capacity is the theoretical maximum output under ideal conditions: no downtime, no changeovers, no defects. Effective capacity subtracts planned losses like maintenance windows and breaks, so it is always lower. Scope engagements against effective capacity, because design capacity overstates what a constraint can actually deliver.
How do you calculate capacity utilization percentage?
Divide actual output by demonstrated capacity, then multiply by 100. A line producing 620 units against a proven 800-unit shift runs 77.5% utilization. Compare stations only after normalizing against takt time or demand, since a high percentage at a non-bottleneck means idle money, not strength.
Can takt time serve as an input measure?
No. Takt time is available production time divided by customer demand, so it is derived from demand, not from resources you control. Treat it as a target cycle time: compare your actual cycle time at the constraint against takt to size the demand gap before quoting.