Case Interview Tips for People Who Hate Memorising Frameworks
A case interview is not a puzzle with a right answer. It is a thirty-minute demonstration of how you break a messy problem apart, and whether anyone could follow you while you did it.
The Five Phases of Almost Every Case
Cases vary enormously in subject and barely at all in shape. Knowing the shape is most of the preparation.
They describe a situation in a few sentences. You play it back in your own words and ask two or three clarifying questions — about the objective, the timeframe, and what success looks like in numbers.
You take sixty to ninety seconds of silence, write down a breakdown of the problem, then present it. This is the single most heavily weighted moment in the case.
You work through branches of your structure, usually with some maths and often with a chart or a table they hand you. They will interrupt, add constraints, and change the facts.
At some point the numbers say something. Saying what it means — not just what it is — is the difference between a candidate who computes and one who advises.
Sixty to ninety seconds: what you would do, the two reasons why, the biggest risk, and what you would check next. Delivered as a decision, not as a summary of your own work.
Interviewers score structure, quantitative comfort, business judgement, and communication — usually on a sheet with those four headings. Note that arriving at the correct answer is not one of them. A candidate who reasons cleanly to a defensible wrong number beats one who guesses the right number silently.
Eight Ways to Prepare That Actually Transfer
Most case preparation goes wrong by optimising for recall. Cases reward construction, and construction has to be practised out loud.
Cases are a spoken performance. Reading case books silently builds recognition, not delivery, and the gap between the two is obvious in the room.
Ten minutes of percentages, ratios, and multiplying awkward numbers. Not because the maths is hard, but because arithmetic under observation is a different skill from arithmetic alone.
Practise breaking unfamiliar problems into three or four mutually exclusive branches. A tailored imperfect structure scores better than a textbook one bolted onto the wrong problem.
Axes, units, timeframe, then the one thing that is odd about it. Interviewers hand you exhibits specifically to see whether you orient before you talk.
You are allowed to say you need a minute to think, and you should. Candidates who talk continuously while thinking produce structures nobody can follow.
How profit decomposes, what fixed and variable costs do at volume, why market share and market growth are different problems. Not an MBA — just enough not to be lost.
Painful and unusually effective. You will hear the filler, the trailing sentences, and the point where your structure stopped being a structure.
Nearly every case round includes fifteen minutes of behavioural questions, and candidates who spent all their preparation on cases lose ground there for no reason.
Eight Prompts and How to Open Them
The opening ninety seconds sets the whole case. Here is where to take each of these first.
"Our client's profits have fallen for three quarters. Why?"
Decompose before you speculate. Profit is revenue minus cost, revenue is price times volume, cost splits fixed and variable. Ask which side moved before you generate a single hypothesis — half of candidates start guessing here and never recover the structure.
"Should our client enter the Indian market?"
Four branches: is the market attractive, can the client win in it, what would entry cost and return, and what are the alternatives to entering at all. That last branch is the one most candidates omit and interviewers are listening for.
"How many electric buses does this city need?"
A sizing question. Say your approach before any number, pick a route — demand-side from ridership, or supply-side from fleet and utilisation — and state every assumption as you make it. Being roughly right transparently beats being precisely right silently.
"Our client wants to cut costs by fifteen percent."
Split the cost base first, then sort by size and by ease of change. Never open with headcount. The good answer identifies which costs are actually addressable in the timeframe rather than listing everything expensive.
"A competitor just cut prices by twenty percent. What now?"
Do not react. Ask why they did it, whether they can sustain it, and what happens to your client's volume and margin under two or three responses. Reactive price-matching with no analysis is the failure mode here.
"Should our client acquire this company?"
Value standalone, then value the synergies separately, then discount the synergies because most of them do not arrive. Then integration risk. Candidates who treat synergies as certain get pushed on it every time.
"Revenue is up but margins are down. What is happening?"
Something in the mix changed. Product mix, customer mix, channel mix, discounting, or input costs. Name mix explicitly as a possibility — it is the intended insight and many candidates never reach it.
"What is your recommendation?"
Answer first, then two supporting reasons, then the biggest risk, then the next step. Do not narrate your journey through the case. They watched it happen and they are now scoring whether you can be put in front of a client.
What to Do While the Case Is Running
Everything here is about being followable. An interviewer who loses your thread scores what they understood, not what you meant.
Silence to build a structure is expected and is not held against you. Ninety seconds of quiet producing a clean breakdown beats thirty seconds producing a mess.
Say which branch you are entering and why you chose it. Interviewers cannot score reasoning they cannot hear, and they will not ask you to repeat it.
Every number you invent should arrive with a reason attached. Unexplained numbers read as guesses even when they are good ones.
Talk through the calculation as you write it. If you slip, the interviewer sees where and can correct you, rather than watching a wrong number appear from nowhere.
Working in round numbers is expected. Announce it once and nobody will mind that you turned 4.7 million into 5 — they will mind if you silently lose a factor of ten.
When an interviewer repeats something or asks whether you have considered X, that is not conversation. It is a course correction, and ignoring it twice is usually fatal.
The number is not the answer. What it implies for the client's decision is. Add one sentence of so-what after every calculation.
Know what you would recommend if time ran out right now. Cases are frequently cut short deliberately to see whether you have been forming a view or just processing.
Four Questions Worth Asking
Case rounds still end with your questions, and candidates who spent all their energy on the case often waste them.
"What kind of cases does this office actually run?"
Practice cases are generic; real work is not. The answer tells you whether the job resembles the interview at all.
"How much client contact does someone get in the first year?"
The single biggest difference between firms at entry level, and rarely stated accurately in recruiting material.
"What separates people who thrive here from people who leave at two years?"
Usually produces the most candid answer of the round, and often names the real working conditions.
"What would you have wanted to know before joining?"
Personal rather than institutional, so the answer tends to be honest rather than rehearsed.
"I would not enter directly. Two reasons: the addressable segment is roughly a third of what the client assumed, and their cost position is around twenty percent behind the incumbent. The biggest risk to that view is my share assumption, which is the softest number here — so before committing I would test it against actual channel data over about two weeks."
Six Ways Cases Are Lost
Interviewers see the same four frameworks all day and can tell instantly. Build a structure from the actual problem, even if it is less elegant.
Unspoken arithmetic cannot be scored and cannot be rescued when it goes wrong. Narrate the calculation as you write it.
A repeated question is a redirection. Candidates who push on with their own plan after two hints almost always fail the round.
Every figure needs a so-what. A calculator produces numbers; the job is producing decisions.
The close is not a recap of your process. Lead with the decision, then the reasons, then the risk.
Behavioural questions are scored in the same round with the same weight for many firms. Preparing only cases leaves points on the table.