How to structure a case in the first two minutes
The opening of a case interview sets the tone for the next half hour. In roughly the first two minutes you listen, confirm the question, ask a few clarifying questions, take a moment to think and then present a structure that fits this problem. Interviewers typically look for clarity and judgement here, not a memorised framework. This guide gives a method you can adapt and a worked example with invented figures. Practice varies by firm and interviewer, so treat it as a default, not a rule.
What happens in the first two minutes?
Five moves, in this order.
- Listen and write. Note the client, the question and every number as it is said.
- Restate. Say the question back in one sentence to check you heard it correctly.
- Clarify. Ask a few questions about the goal, the context and any definitions.
- Pause. Ask for a minute to organise your thoughts. Most interviewers expect this and respect it.
- Structure. Present an issue tree or a short set of areas, with a first hypothesis and where you would start.
How do I listen and restate?
Write the question at the top of the page in the interviewer's own words, then the facts, then the numbers. Restate in one sentence: "So the question is whether Daventree Bakeries should do X, to achieve Y, by when. Is that right?" A restatement catches misunderstandings early, and it buys time to think. Avoid repeating the whole prompt.
Which clarifying questions are worth asking?
Ask two to four, chosen because the answers would change your structure.
- Objective. What does success mean: profit, share, growth, a payback period? Is there a target or a deadline?
- Scope. Which products, regions and time horizon are in or out?
- Definitions. What does a term mean (for example "profit" before or after central costs)?
- Context. Is this a change, and since when? What has the client already tried?
Do not ask for data you can structure without, and avoid a long list of questions that delays your thinking. If the interviewer says "that is for you to decide", state an assumption and move on.
How do I build the structure?
Start from the question, not a framework. Ask: what would I need to believe for the answer to be yes, or what are the main reasons this happened? Then split the problem into parts that do not overlap and together cover everything (mutually exclusive and collectively exhaustive, or MECE). Common starting points can help.
- For a profit question, profit is revenue minus costs, and revenue is price times volume.
- For a "should we enter" question, think about the market, the client's ability to win and the economics.
- For a "how big" question, choose a top-down or bottom-up route.
Adapt the labels to the client. "Customers, competitors, company" is a generic template. "Footfall, basket size, and shop costs" is a structure for a bakery chain. Aim for three or four branches, each with two or three points beneath it, and say why you would start where you start.
How do I state a hypothesis?
A hypothesis is a first answer you are prepared to test and to change. "My initial view is that the fall is cost-driven rather than demand-driven, because sales are growing, so I would start with costs." It shows you are heading somewhere. Say what data would disprove it. If the interviewer's next figures contradict you, say so and revise without embarrassment.
What does it look like in practice?
Prompt: a UK bakery chain, Daventree Bakeries, has seen profit fall over the last year. The interviewer says revenue grew from 150m to 156m, and costs rose from 132m to 144m. The question: why did profit fall, and what should be done?
Computed before writing:
- Revenue growth: 156 / 150 - 1 = 4.0%.
- Cost growth: 144 / 132 - 1 = 9.1%.
- Profit: 18m falling to 12m, a fall of 33.3%; margin from 12.0% to 7.7%.
What you might say: "To confirm, Daventree's profit has fallen, and we want to know why and what to do. May I ask two things? Is profit measured before central costs, and have prices changed? [Answer: after all costs; prices rose slightly.] Thank you. May I take a minute? ... Profit is revenue minus costs. Revenue is up 4% but costs are up about 9%, so the issue looks cost-driven, and my hypothesis is that costs rose faster than volume. I would look at three areas: first, input costs such as ingredients and energy, split into price and volume; second, shop costs such as labour and rent; third, central costs and any one-offs. I would then check whether revenue growth came from price or volume, because if price rose and volume fell, the picture is different. I would like to start with cost by category."
This answer is under a minute, uses the interviewer's figures, states a hypothesis and an order of attack, and leaves room for the interviewer to redirect.
What mistakes should I avoid?
- Reciting a framework with labels that do not fit the client.
- Starting to solve before the question is understood.
- Silence without saying you are thinking.
- Too many branches, so you cannot cover them.
- Splits that overlap, such as "internal factors" and "costs".
- Defending your structure when the interviewer hints it is incomplete. Take the hint, say how it changes your plan and continue.
- Ignoring the objective and the numbers given.
How do I practise?
Use short prompts and give yourself two minutes to produce a structure aloud, then check it for fit, MECE-ness and a clear starting point. Practise with another person who can interrupt, since you need to think aloud while someone watches. Review your recordings for rambling and missing hypotheses. Keep a list of structures you built, and ask for each: what would make it wrong?
What to do this week
- Take three one-line case prompts and build a structure for each in two minutes, aloud.
- For each, write the hypothesis and the one piece of data that would disprove it.
- Check each for overlap and gaps.
- Practise the restatement and two clarifying questions.
- Ask a friend to hand you a "new fact" mid-structure and practise adapting calmly.
About this guide: it is original to Charta Learning and has not yet been reviewed by an experienced practitioner. Names in examples are invented. Interviewers differ in how much structure they want up front, so follow their cues.