Case interviews aren't won by memorizing a list of frameworks, they're won by how you structure, communicate, and adapt your thinking live, in real time, in front of an interviewer. This deck focuses on the process itself: breaking a broad prompt into a MECE issue tree, framing your structure out loud in the first few minutes, leading with a hypothesis instead of exploring everything equally, and recovering cleanly when new information proves your first structure wrong. If you already know your frameworks and want to sharpen how you actually run a case, this is built for that.
18 cardsConsultingFree, no signup
Card 1 of 18
You keep hearing that consulting firms interview you with something called a case. What actually is a case interview, and why do firms run one instead of just asking about your résumé?
A case interview is a short business problem you solve out loud, with the interviewer, in real time.
The interviewer describes a client situation, say a food company whose profits have slipped, and you work it through out loud over roughly twenty to forty minutes: asking for data, doing arithmetic on paper, and finishing with a recommendation. Firms run cases because the job itself is being handed a messy problem, often in an industry you know nothing about, and having to break it apart quickly. A résumé shows what you have already done, while a case shows how you think when you don’t yet know the answer.
Card 2 of 18
A friend tells you the case is a test of business knowledge, so you should spend the week memorizing industry facts. Is that right, and what is the interviewer actually scoring?
They score how you break the problem down, handle numbers, judge, and explain, not what you already know.
Interviewers are watching four things: whether you split the problem into sensible parts, whether you’re comfortable doing arithmetic out loud and saying what the number means, whether your judgement about the business is sensible, and whether you communicate clearly enough to sit in front of a client. Industry knowledge isn’t on that list. The interviewer will hand you every fact you need, because in the actual job consultants get staffed on industries they have never touched and are expected to get up to speed from the data in front of them.
Card 3 of 18
It’s the morning of your first consulting interview and you have no idea how the next forty minutes are meant to go. What are the phases of a case, and roughly how long does each one take?
Prompt, clarify, structure, analysis, recommendation, with the analysis taking most of the clock.
The interviewer reads the prompt while you take notes, about a minute or two. You ask a couple of clarifying questions, take a short silence to build your structure, and present it, three or four minutes in total. Then comes the analysis, the long stretch at roughly fifteen to twenty-five minutes, usually including a chart to read and some arithmetic. You close with a recommendation in a minute or two. Timings vary by firm and by round, but the shape rarely does, and knowing it stops you from spending ten minutes on a structure meant to take three.
Card 4 of 18
You’ve just been given the prompt: Calder Foods is thinking about buying a smaller rival. Before you start structuring, what should you ask, and which questions are actually worth the interviewer’s time?
Ask only what would change the shape of your structure, then repeat the objective back and go quiet.
Ask two or three questions, and only ones whose answers would change how you structure. The three that usually earn their place are what success looks like in numbers, how the client actually makes money, and whether there’s a hard constraint like a budget, a deadline, or a brand they won’t give up. Questions that wouldn’t move your structure, industry trivia or data you’d be handed anyway, read as stalling. Restate the objective in your own words before you take your silent structuring time, so you and the interviewer are solving the same problem.
Card 5 of 18
What does MECE actually mean, and why does it matter for how you structure a case?
MECE means your buckets don't overlap and together cover the whole problem, so nothing's double-counted or missing.
MECE stands for mutually exclusive, collectively exhaustive: each branch of your structure shouldn't overlap with another, and together the branches should cover the entire problem. In a case, that keeps you from counting the same driver twice under two different buckets, and from leaving an entire category of the problem unexamined. The practical test isn't reciting the acronym, it's asking yourself, before you speak, whether a given fact could plausibly belong to two of your buckets at once, or whether there's a whole area your structure doesn't touch. Real interview structures are rarely perfectly MECE, but they should be close enough that the interviewer never has to ask "where would that go?"
Card 6 of 18
NorthPeak Airlines, a regional airline, comes to you because profits fell 20% last year and the CEO wants to know why. How do you build an issue tree to break that question down?
Split the driving question into MECE branches, then keep splitting until each branch is a fact you could go verify.
Start with the driving question at the top of the page, why did NorthPeak's profit fall 20%, and split it into its first MECE layer: since profit equals revenue minus cost, one branch is revenue and the other is cost. Each branch then splits again, revenue into price and volume, cost into fixed and variable, and volume can split further into individual routes or ticket classes if that's where the real story lives. Keep branching until each end node is something you could actually go check with a number, like "average fare on the Denver-Boise route," not a vague theme like "market conditions." The finished tree becomes your map for deciding which branches are worth digging into first, not just a diagram you draw and set aside.
Card 7 of 18
You've just heard the prompt and taken your silent structuring time. How should you actually lay out and walk the interviewer through your structure in the next few minutes?
State the objective, name your buckets top-down, then say which branch you'd start with and why.
Open by restating the objective in your own words so the interviewer knows you understood the question, then walk through your structure top-down: name the two or three main buckets before diving into any sub-branch, so the interviewer can follow the shape of your thinking before the detail. Say each bucket out loud as a short label, not a paragraph, drawing it on paper as you talk if you have one in front of you. Close by telling the interviewer which branch you'd want to start with and why, since a structure that never turns into a starting point is just a list. The whole moment should take well under two minutes, structure is meant to set up the conversation, not replace it.
Card 8 of 18
Someone hands you a list of case frameworks: profitability, the 3Cs, the 4Ps, Porter’s Five Forces. What is each one actually for?
Each named framework is a checklist for one kind of question, not a general purpose case answer.
A profitability tree splits profit into revenue and cost and keeps splitting to find which piece moved. The 3Cs, customer, competition and company, size up where a client stands in a market. The 4Ps, product, price, place and promotion, cover how a product gets taken to market. Porter’s Five Forces, rivalry, new entrants, substitutes, buyer power and supplier power, judge how structurally attractive an industry is. Treat them the way a pilot treats a checklist, as a quick test of whether your own structure has a hole in it, rather than as the structure itself.
Card 9 of 18
What's the essential difference between a profitability case and a market-entry case, and how does it change the driving question you're answering?
Profitability cases diagnose why a number moved; market-entry cases decide whether to make a move at all.
A profitability case starts from a problem that's already happened, profit or margin has moved in the wrong direction, and the driving question is diagnostic: which specific lever, price, volume, fixed cost, or variable cost, actually caused the change, and what should the client do about it. A market-entry case starts from a decision the client hasn't made yet, whether to enter a new market, launch a product, or make an acquisition, and the driving question is evaluative: is this worth doing, compared to the alternative of not doing it. That distinction should reshape your structure, not just your opening line: a profitability tree branches through revenue and cost hunting for the delta, while a market-entry tree branches through market attractiveness, the client's ability to win in it, and the financial return, because you're building a case for or against a decision rather than diagnosing what already broke.
Card 10 of 18
What's the difference between reciting a memorized framework and building a structure live for the prompt in front of you, and why can interviewers usually tell?
A memorized framework ignores the specific prompt; a live structure is built from the actual facts you were given.
A scripted framework is a generic template, the classic profitability tree or 4Ps, recited the same way regardless of what the prompt actually said, and it shows up as buckets that don't quite fit: a market-entry structure applied to a cost problem, or category names that don't match the client in the prompt. A structure built live starts from the specific facts and objective you were just given and only borrows the shape of a familiar framework where it genuinely applies, which is why it sounds tailored, using the client's actual situation, numbers, and constraints instead of generic labels. Interviewers can tell the difference almost immediately because a scripted structure tends to arrive too fast, sound identical to the last several candidates, and fall apart the moment a follow-up question pushes on a branch that doesn't actually apply here. The fix isn't memorizing more frameworks, it's practicing the underlying skill, breaking a novel question into MECE parts, so you can build the right structure for whatever prompt shows up.
Card 11 of 18
Your first interview was a back-and-forth where the interviewer drove every step, and your second let you run the case yourself. What’s the difference between those two styles, and what changes for you?
Interviewer-led cases hand you one question at a time, candidate-led ones make you pick the next move.
In an interviewer-led case the interviewer sets the sequence and hands you one question at a time, so the skill is answering each one crisply and completely, then stopping rather than wandering ahead. In a candidate-led case you drive: after your structure you decide what to ask for next, when a branch is settled, and when to move on, so the risk is stalling or drifting. Firms differ, some lean consistently one way and many sit in between, and the style can change between rounds. Your structure has to be just as good either way, what changes is who chooses the next step.
Card 12 of 18
What does it mean to be "hypothesis-driven" in a case, and how is it different from just working through your structure branch by branch in order?
Lead with your best guess at the answer, then test it first, instead of exploring every branch equally.
Being hypothesis-driven means forming an early, informed guess about what the answer probably is, based on the prompt and the first facts you're given, and then using your structure to go test that guess first rather than working through every branch in the order you listed them. It's different from exhaustive branch-by-branch exploration, which treats every part of the tree as equally likely to hold the answer and burns limited case time confirming things that were never really in question. A hypothesis isn't a guess you're married to, if the data doesn't support it, you drop it and form a new one, but leading with one gives you a reason to ask for the data that actually matters instead of every data point available. It mirrors how real consulting engagements run: you rarely have time to check everything, so you prioritize based on what's most likely to be true and most valuable to confirm.
Card 13 of 18
You’ve laid out four branches for Northbay Logistics and the interviewer says, “where would you like to start?” How do you choose, and what do you actually say?
Start where the money is biggest and the answer is quickest to check, and say that reason out loud.
Two tests settle it: which branch could plausibly account for most of the gap, and which one you could close fastest with data the interviewer is likely holding. Then say the reason aloud, for example “labor is roughly seventy per cent of Northbay’s cost base, so a small move there outweighs anything on price, I’d start there.” Naming the criterion is half of what’s being graded, and it gives the interviewer a cheap way to redirect you if they’d rather begin elsewhere. Picking a branch silently looks like guessing, even when the pick is right.
Card 14 of 18
Fifteen minutes into a case for Bramwell & Co., new data makes it clear your original structure is aimed at the wrong problem. What do you do?
Name the mismatch out loud, adjust the structure on the spot, and keep going, don't pretend nothing changed.
The worst move is plowing ahead on a structure you already know doesn't fit, or quietly abandoning it without saying anything, both leave the interviewer unsure whether you've noticed the problem at all. Instead, name it directly: "this new data suggests the issue is really about X, not Y, let me adjust my structure," then briefly restate the one or two branches that change. You don't need to rebuild the whole tree from scratch, usually only the branch that's now wrong needs to move, the rest of your structure often still holds. Interviewers read a clean recovery as a strong signal, real client work is full of information that overturns your first hypothesis, and handling it calmly is closer to the actual job than never being wrong in the first place.
Card 15 of 18
You've spent 25 minutes working through Ferrow Retail Group's declining store profitability and found the real driver is rising labor cost at a handful of underperforming stores, not weak sales. How do you wrap up the case?
Lead with the recommendation, back it with your top two or three findings, then flag the key risk.
Synthesis means answering the original question directly, not narrating everything you did to get there, so lead with the recommendation itself: Ferrow should address labor cost at the specific underperforming stores rather than launch a broad sales initiative. Back it up with the two or three findings that actually drove that answer, in this case that labor cost per store rose faster than revenue at a specific subset of locations while sales held roughly flat elsewhere, skipping the dead ends you explored along the way. Name the biggest risk or open question, for example whether the labor cost spike is a one-off staffing issue or a structural wage trend, and what you'd want to check next to be sure. A good close takes well under a minute and could stand alone as the answer, even if the interviewer heard nothing else you said.
Card 16 of 18
The prompt is about falling margins at a specialty chemicals distributor and you have never heard of that industry in your life. How do you structure it without faking knowledge you don’t have?
Every business buys something, adds something, and sells it, so structure the economics, not the industry.
Say plainly that you’re not familiar with the industry and ask one or two questions about how the client makes money: what it sells, who buys it, and what its largest costs are. Then structure the economics rather than the sector, revenue as price times volume, cost as what the client buys plus what it costs to operate, and only add industry-specific branches once the interviewer has given you the background. Interviewers expect this and volunteer the context, because their own consultants land on unfamiliar industries constantly. Guessing at jargon you don’t understand costs far more than admitting the gap in one sentence.
Card 17 of 18
Meridian Health, a hospital group, wants to cut the time patients wait in its emergency department. There’s no profit in the objective at all. How do you structure a case like that?
When profit isn’t the goal, structure the process the metric travels through, step by step.
A profit tree is one shape of structure, not the only one. Turn the goal into a measurable number first, here the average minutes between a patient arriving and being seen, then break the problem along the path that number travels: arrival, triage, waiting for a bed, treatment, and discharge, asking at each step where the queue builds. Splitting demand against capacity is a second workable cut. Public sector, non-profit and operations cases come up regularly, and they are exactly where a memorized profit framework falls apart while a structure built from the objective still holds.
Card 18 of 18
Halverson Industries is losing money on its core manufacturing business and separately weighing whether to enter a new overseas market. How do you build a structure that handles both a profitability problem and a market-entry decision in the same case?
Treat it as two connected trees: diagnose the core problem first, then use that answer to inform the entry decision.
Resist the urge to mash both prompts into one tree, build two separate MECE structures, a profitability tree for the manufacturing decline and a market-entry tree for the overseas decision, and connect them explicitly rather than blending their branches together. Sequence matters: diagnosing the core profitability problem first usually matters more, since a company bleeding cash domestically has a weaker case for funding an expansion, and what you learn there, for instance whether the problem is temporary or structural, directly changes how attractive the new market decision looks. State that sequencing out loud to the interviewer before you start, for example "I'd want to understand the manufacturing issue first, since it affects how much capacity Halverson has to fund entering a new market," so the connection between the two trees is clear rather than left implicit. The goal isn't a single unified diagram, it's showing you can hold two structures in mind at once and reason about how the answer to one changes the other.
Reading is not the same as remembering. Study this deck with scheduled review so the cards come back before you forget them.