Capacity Measurement, Utilisation & Effective Capacity
At 7:55 pm, a hospital reception counter has empty chairs, two billing windows open, one doctor running late, and a line forming near diagnostics. On paper, the facility has capacity; in reality, the usable capacity is being throttled by one constrained step.
- Capacity is the maximum output a process can deliver in a defined time period.
- Design capacity is the theoretical maximum under ideal conditions; it is rarely the operating target.
- Effective capacity is the practical maximum after planned losses such as breaks, maintenance, changeovers and staffing constraints.
- Utilisation = actual output / design capacity; it tells you how much of the installed capacity is being used.
- Efficiency = actual output / effective capacity; it tells you how well usable capacity is being converted into output.
- Capacity cushion protects against demand spikes, variability and breakdowns; chasing 100% utilisation often creates queues.
- The interview-safe answer is: define the denominator, identify the bottleneck, compute utilisation and efficiency, then recommend whether to add capacity, shift demand, or debottleneck.
The Big Picture: Capacity Has Layers
Capacity is not one number. A process starts with a theoretical ceiling, loses capacity to real-world constraints, and finally delivers actual output. Good managers do not ask only, “How much can we produce?” They ask, “Which capacity number are we talking about?”
Core Explanation: What Capacity Really Means
Capacity is the rate at which a system can produce output over a specified period. The key words are rate and period. A factory may produce 10,000 units per day, a call centre may handle 800 calls per shift, and a restaurant may serve 120 covers per evening.
The mistake is to treat the highest possible number as the number available every day. Real processes lose capacity because of maintenance, absenteeism, setup time, material shortages, rework, cleaning, batching, machine downtime and coordination gaps.
That is why capacity measurement needs three lenses:
If you already know cycle time, takt time and lead time, capacity becomes easier: capacity is the inverse of cycle time at the constrained step. If the bottleneck takes 5 minutes per unit, the process cannot sustainably produce more than 12 units per hour from that line.
Design Capacity, Effective Capacity and Actual Output
Use these three numbers like a diagnostic ladder.
Think of a movie theatre. Design capacity is the number of seats multiplied by shows. Effective capacity is lower after cleaning time, show scheduling, staffing and screen allocation. Actual output is tickets sold. The same logic applies to manufacturing, hospitals, dark stores, banks and airports.
A quick-commerce dark store may have shelves, pickers and riders, but its effective capacity depends on the slowest combination of picking, packing, billing and dispatch. Its primary capacity driver is process speed inside a compact fulfilment layout, supported by inventory availability, rider allocation and demand forecasting. The so what: adding more riders will not help if picking aisles or packing tables are already the bottleneck.
A Five-Step Way to Measure Capacity Correctly
When a capacity question appears, do not jump straight to one formula. Follow this sequence and name the metric you are using at each step.
Key Capacity Metrics and What Good Looks Like
These are the measures interviewers expect you to know. The “good” value is a rule of thumb, not a universal benchmark - always adapt it to the industry and variability of demand.
Worked Example: A Clinic with Three Consultation Rooms
A clinic has three consultation rooms. Each room can theoretically handle 4 patients per hour, and the clinic runs for 8 hours.
Design capacity = 3 rooms x 4 patients per hour x 8 hours = 96 patients per day.
But the clinic loses 1 hour per room to doctor breaks, cleaning, shift handover and scheduling gaps. So each room is effectively available for 7 hours.
Effective capacity = 3 rooms x 4 patients per hour x 7 hours = 84 patients per day.
If the clinic actually serves 72 patients:
- Utilisation = 72 / 96 x 100 = 75%
- Efficiency = 72 / 84 x 100 = 85.7%
- Capacity cushion versus effective capacity = (84 - 72) / 84 x 100 = 14.3%
The interpretation is simple: installed capacity looks underused at 75%, but usable capacity is being converted reasonably well at 85.7%. The next question is not “Why are we not at 100%?” It is “Where is the remaining loss - demand, doctor availability, diagnostics, billing or patient no-shows?”
How Capacity Decisions Change by Demand Pattern
The right capacity policy depends on demand variability. A stable, repetitive factory line can run with a smaller cushion than an emergency department, airline check-in counter or food delivery operation.
This is where capacity links directly to bottleneck thinking. If one workstation limits total throughput, adding capacity anywhere else only increases waiting before or after the constraint. For that diagnostic method, revise finding the bottleneck and the Theory of Constraints.
Definitions You Should Be Able to Say in One Breath
- Capacity: The maximum output rate a process can deliver in a specified period.
- Design capacity: The maximum theoretical output possible under ideal operating conditions.
- Effective capacity: The maximum realistic output after planned losses and normal operating constraints.
- Utilisation: Actual output divided by design capacity, expressed as a percentage.
- Efficiency: Actual output divided by effective capacity, expressed as a percentage.
- Capacity cushion: Spare capacity kept to absorb variability, uncertainty or demand spikes.
Case Study: Narayana Health and Capacity as a System
Narayana Health shows that service capacity is not just more beds; it is the coordinated availability of doctors, operating theatres, diagnostics, nurses, protocols and patient flow.
Narayana Health is a useful case because healthcare capacity is harder than factory capacity. A hospital cannot simply say, “We have beds, so we have capacity.” A bed without nurses, diagnostics, doctors, pharmacy support, oxygen, ICU backup or discharge coordination is not effective capacity.
Situation: In high-volume hospital operations, demand arrives unevenly, clinical work varies by patient complexity, and many resources are shared. One delayed diagnostic report can hold up a discharge; one unavailable specialist can idle an operating theatre slot.
The move: Narayana Health built its model around high-throughput clinical operations, standardised care pathways where medically appropriate, strong scheduling discipline and better utilisation of expensive assets such as operating theatres and diagnostics. The primary driver is process standardisation around repeatable clinical pathways. Supporting drivers include focused doctor time, coordinated nursing teams, hub-and-spoke patient flows, procurement discipline and scheduling of shared assets.
The lesson: Effective capacity in a service business is a network property. It comes from synchronising constrained resources, not merely from adding physical infrastructure.
The strategic so what: Narayana Health demonstrates that capacity improvement is often a coordination problem before it is a capex problem. A shallow answer says, “Add more beds.” A strong answer asks, “Which constrained resource is preventing beds from becoming usable capacity?”
How AI Changes Capacity Measurement, Utilisation & Effective Capacity
AI changes this topic by making capacity more dynamic, predictive and simulation-led.
- ML demand forecasting: Hospitals, airlines, warehouses and service desks can forecast hourly or daily demand more accurately, then roster staff and open capacity bands accordingly.
- Digital twins and simulation: Teams can test whether adding a counter, changing a shift pattern or redesigning a layout improves throughput before spending capex. This is the practical bridge to using AI and simulation to test a process design.
- Real-time bottleneck alerts: Computer vision, IoT and workflow data can flag queue build-up, machine downtime, picker congestion or delayed discharge before the day is lost.
Use ChatGPT or Claude to practise capacity cases: paste a short process description, ask it to identify design capacity, effective capacity, utilisation, bottleneck load and capacity cushion, then challenge its assumptions. For company-specific prep, load an annual report and your notes into NotebookLM and ask: “Where might this business face capacity constraints, and how would I measure them?”
Interview Relevance
“A plant has an installed capacity of 1,000 units per day but produces only 720 units. Management says utilisation is poor. How would you analyse the problem?”
Always say the denominator out loud: “I am using design capacity for utilisation and effective capacity for efficiency.” That one sentence makes your answer sound structured and prevents formula confusion.
Common Mistake
The most common mistake is treating 100% utilisation as automatically good. It often destroys service levels because there is no cushion for variability, breakdowns or demand spikes. The one-line fix: optimise utilisation at the bottleneck, but protect the system with the right capacity cushion.