Capacity and bottleneck analysis answers one question: which step limits the output of the whole process? Compute each step’s capacity in the same unit, find the lowest one, and remember that improving any other step adds nothing until the constraint moves. Then size the gain, check where the bottleneck shifts next, and test whether the investment pays back.
Capacity and bottleneck analysis is the core quantitative skill in operations cases: find the one step that caps the output of the whole process, then decide whether fixing it is worth the money. The logic is simple, but candidates routinely lose points on unit mismatches, ignoring the next bottleneck, or “optimizing” a step that was never the constraint.
What Capacity and Bottleneck Analysis Means
A bottleneck is the step in a process with the lowest effective capacity, and it sets the maximum output of the entire system. Capacity is the maximum amount a step can process in a given time, for example units per hour, patients per day, or calls per shift. A process behaves like a chain: total output equals the capacity of its weakest link.
This is the practical core of the Theory of Constraints, which is covered at framework level in our operations case framework guide. This guide goes one level deeper into the mechanics you need to run the numbers live.
Operations questions are common in operations cases across manufacturing, healthcare, logistics and services. The prompt often sounds like “output has dropped,” “customers wait too long,” or “we want to produce more without building a new plant.” In all three, the bottleneck is the first thing to find.
The Five-Step Method
Work through the same sequence every time. It keeps your structure clean and gives the interviewer an easy path to follow.
flowchart LR
A["1. Map the process steps"] --> B["2. Put capacity in one unit"]
B --> C["3. Find the lowest capacity"]
C --> D["4. Size the gain from fixing it"]
D --> E["5. Check the next bottleneck and the payback"]
- Map the process. Ask the interviewer for the steps in order, from input to finished output. Write them as a simple left-to-right line.
- Convert every step to the same unit. If one step is quoted per hour and another per shift, or one machine makes 2 units per cycle, normalize everything to units per hour before comparing.
- Identify the lowest capacity. That step is the bottleneck. Also compute utilization (actual output divided by capacity) for each step to see where slack sits.
- Size the gain. Estimate how much total output rises if the bottleneck is improved, and convert it to revenue or contribution margin.
- Check what happens next. Once the bottleneck is lifted, the constraint moves. Your gain is capped by the next-lowest step.
Worked Example: A Three-Step Production Line
A manufacturer runs a line with three sequential steps. The client wants more output without building a second line.
| Step | Capacity (units/hour) | Actual output (units/hour) | Utilization |
|---|---|---|---|
| Cutting | 120 | 80 | 67% |
| Assembly | 80 | 80 | 100% |
| Packaging | 100 | 80 | 80% |
Assembly is the bottleneck: the line cannot exceed 80 units per hour, no matter how fast cutting runs. With two 8-hour shifts, daily output is 80 x 16 = 1,280 units.
Now test the options the way an interviewer expects you to.
| Option | New line output | Daily output | Why |
|---|---|---|---|
| Add a second cutting machine | 80 | 1,280 | Cutting was never the constraint |
| Raise assembly to 95 units/hour | 95 | 1,520 | Assembly still limits, now at 95 |
| Raise assembly to 110 units/hour | 100 | 1,600 | Packaging (100) becomes the new bottleneck |
The third row is the insight that separates strong answers. Spending more to lift assembly from 95 to 110 buys only 5 extra units per hour, because packaging now caps the line. State this out loud: “After assembly is fixed, packaging at 100 units per hour is the next constraint.”
Turning Capacity Into Dollars
Suppose each unit carries a contribution margin of $12 and the assembly upgrade to 95 units per hour costs $400,000. The extra output is 240 units per day, so the added contribution is 240 x $12 = $2,880 per day. Payback is $400,000 / $2,880, or roughly 140 operating days. This is where capacity analysis connects to the break-even and payback guide: the bottleneck tells you how many extra units you can sell, and payback tells you whether the investment is worth it.
Four Traps That Cost Candidates Points
Most mistakes in this part of an operations case are mechanical, not conceptual. In our experience reviewing practice sessions, these four come up repeatedly.
| Trap | What it looks like | Fix |
|---|---|---|
| Unit mismatch | Comparing units per hour with units per shift | Normalize everything before comparing |
| Nameplate vs. effective capacity | Using theoretical machine speed and ignoring downtime or defects | Ask for effective capacity after breaks, changeovers and scrap |
| Optimizing a non-bottleneck | Recommending a faster machine at a step with slack | Show utilization first, then pick the constrained step |
| Forgetting demand | Adding capacity when the market cannot absorb more | Compare the new capacity to demand before recommending |
The last trap matters most. If customers only buy 1,000 units a day, lifting capacity beyond that earns nothing. A complete answer checks capacity on the supply side and demand on the market side. For demand-side estimation, the market sizing techniques guide gives you the toolkit.
Beyond Manufacturing: The Same Logic Elsewhere
Bottleneck thinking transfers to any flow of work. Only the unit changes.
| Setting | Flow | Typical bottleneck | Unit |
|---|---|---|---|
| Factory | Cut, assemble, pack | Slowest machine | Units per hour |
| Hospital | Triage, scan, treat, discharge | Imaging or bed turnover | Patients per day |
| Warehouse | Receive, pick, pack, ship | Pick stations | Orders per hour |
| Call center | Queue, answer, resolve | Agent availability | Calls per shift |
For industry-specific versions, see our manufacturing industry cases guide, the healthcare operations cost cases guide, and the distribution network design cases guide. If inventory piles up in front of one step, that is often the visual signal of the bottleneck, which links to the inventory optimization cases guide.
How to Present It Out Loud
Interviewers score the clarity of your reasoning as much as the answer. Use a short script:
- State the structure first: “I will map the steps, compare capacity in one unit, find the constraint, then size the gain.”
- Say the unit every time you give a number.
- Name the bottleneck explicitly, then name the next one.
- Close with a recommendation, the expected gain, and one risk, such as demand or quality.
Keep the arithmetic clean. Round early and check your result against a sanity range. If you need to sharpen speed, the mental math guide covers the shortcuts that work under pressure.
Practice It With CasesCoach
Capacity math is a skill you build through repetition, not reading. CasesCoach gives you three ways to practice it, and each one fits a different stage of preparation:
- Guided Mode lets the interviewer give hints and guidance, which is a good way to learn the sequence of mapping, comparing and sizing. See Guided Mode vs. Expert Mode for when to use each.
- Expert Mode simulates real interview pressure, so you practice finding the bottleneck while the clock is running.
- Chat with Case walks you through a real case in five stages (context, choosing a framework, filling in data, analysis, conclusion), so you can practice turning a messy operations problem into a clear recommendation. Read more in Chat with a Case to Analyze Real Business Problems.
After each session, you get instant AI feedback scored on five dimensions: Structure, Analysis, Communication, Business, and Math. Math is where capacity cases are won or lost, and the breakdown shows you exactly which dimension to work on next. How AI mock interviews score your performance explains each dimension in detail.
Key Takeaways
- The bottleneck is the step with the lowest effective capacity, and it caps the output of the whole process.
- Convert every step to the same unit before comparing, and compute utilization to see where slack sits.
- Improving a non-bottleneck step adds nothing to total output.
- After you fix a bottleneck, the constraint moves. Your gain is capped by the next-lowest step.
- Always connect capacity to demand and to dollars: contribution margin per unit, then payback.
- Present the structure first, name the bottleneck and the next one, and finish with a recommendation and a risk.
Next Step
Sign up for free to try three complete cases and an AI Mock Interview, and see how the five-dimension feedback scores your capacity math. When you are ready to go deeper, upgrade to Pro to unlock 835+ real cases and practice with the AI Mock multiple times.