McKinsey’s website says it rejects roughly 99% of applicants. That number is probably a bit high as a marketing flourish, but the reject rate at top consulting firms is genuinely brutal, and the case interview is the single biggest filter in that funnel. If you bomb the case, nothing else matters: not your GPA, not your referral, not your polished cover letter.
I’ve talked to people who spent 47 hours preparing and still froze. I’ve also seen candidates with maybe 12 hours of focused practice walk out with offers from BCG. The difference usually isn’t raw intelligence. It’s whether they learned to think out loud in a specific way that consultants recognize as structured.
This guide covers how to actually prepare, not just what to prepare.
What a case interview is actually testing
The case interview is not a business knowledge test. Most of the cases you’ll encounter involve industries you’ve never worked in. That’s on purpose.
What the interviewer is watching for: can you break down an ambiguous problem into parts, stay calm when new information arrives that changes your answer, and communicate reasoning in real time? Think of it less like an exam and more like a working session where you happen to be new to the problem. They want to see if you’d be useful to have in the room.
The McKinsey case interview basics guide describes this as evaluating “problem-solving skills and business judgment” rather than domain expertise. That framing is accurate. The firms genuinely don’t expect you to know the profit margins of a Brazilian cement manufacturer. They expect you to know how to find out.
The core frameworks, and when to stop memorizing them
Every prep book teaches the same set: Profitability Tree, Porter’s Five Forces, the 4 Cs, McKinsey’s 7S model. Learning these is necessary but not sufficient. The bigger risk is over-relying on them.
A framework is a starting point for structuring your thinking, not a script to read aloud. Interviewers at tier-1 firms have heard “let’s look at revenues and costs” approximately six thousand times. What they haven’t heard enough of is someone who pauses, asks one smart clarifying question, then builds a custom structure for the specific problem in front of them.
My honest opinion, which could be wrong: the MECE framework obsession in prep culture creates a certain kind of candidate who sounds polished but shallow. The most impressive candidates I’ve seen described build frameworks that are actually specific to the case. “Let’s look at revenues and costs” works for every case, which means it fits none of them perfectly.
A reasonable approach: learn the standard frameworks cold so they’re automatic, then practice abandoning them when the case calls for something more specific.
How to structure a realistic practice plan
The Bureau of Labor Statistics projects management analyst roles (which includes consulting) to grow 11% through 2033, faster than most sectors. There are more candidates now than there were five years ago, which means average preparation quality has risen. You need to prepare more seriously than you did reading advice from 2019.
Here’s a rough plan that works for most people with 6 to 8 weeks before interviews:
- Week 1-2: Learn the frameworks. Read Case in Point or Victor Cheng’s Look Over My Shoulder once. Don’t re-read; move to practice.
- Week 3-4: Solo practice. Work through 15 to 20 cases from free sources (Case Coach, Preplounge’s free tier). Time yourself. The goal here is fluency in structuring, not perfection.
- Week 5-6: Partner practice, minimum 3 sessions per week. Partner practice is where the real improvement happens because it forces you to talk through your thinking in real time.
- Week 7-8 (if you have it): Focus on your weak points specifically. Failing consistently on market-sizing estimates? Do 10 market-sizing exercises in a row. Math errors? Daily mental arithmetic drills.
One thing people underestimate: the first partner practice session is almost always terrible. That’s fine. The awkwardness of talking through a business problem with a stranger is a skill you build, not a talent you either have or don’t.
Math you actually need to be able to do
You will do arithmetic in the interview. This is not optional. The case math is usually not hard (basic multiplication, percentages, estimation) but it happens under pressure, in front of someone evaluating you, while you’re also thinking about the structure of your answer.
Practice doing math out loud. The instinct is to go quiet while you calculate, but interviewers find silence uncomfortable and they can’t follow your reasoning. Get used to narrating: “I’m going to estimate the addressable market at 50 million households, assume 10% penetration in year one, that’s 5 million customers, at an average revenue of $200 that’s $1 billion at maturity…”
Common mistake: using nice round numbers in market-sizing that make the final answer look suspiciously tidy. Real estimates are messier. An estimate of $2.3 billion reads as more credible than exactly $2 billion, even if you’re making up both numbers.
What happens in the partner interview portion
Most firms also include a fit interview alongside the case, sometimes called the PEI (Personal Experience Interview). McKinsey in particular weighs this more heavily than most firms admit. The questions are predictable: tell me about a time you led without authority, describe a situation where you had to convince a skeptical stakeholder, when have you failed and what did you learn.
The failure I see most often here is candidates who have a great story but tell it too linearly. They describe what happened, what they did, and what the result was, in that order, as if reading a resume bullet. The interviewer wants to feel the difficulty of the moment and understand specifically what you thought was the right move at the time. Get specific. “I wasn’t sure whether to escalate to the VP or handle it at my level” is more credible than “I faced a difficult decision.”
Prepare 5 or 6 real stories from your experience that can flex across different PEI questions. Don’t try to memorize a different story for each question type. You’ll run out of material and sound robotic.
Using AI tools during preparation
One newer option worth knowing about: tools like Craqly let you run mock interview sessions with real-time AI feedback on your answers. For case prep specifically, having something that can respond to your structure in real time, without needing to coordinate a partner schedule, is genuinely useful for the early weeks when you’re still building fluency. That said, partner practice with a real human is still irreplaceable for the final weeks before your actual interview. AI doesn’t replicate the social pressure of talking to another person who can push back on your reasoning unpredictably.
The week of your actual interview
Stop doing full cases two or three days before the interview. At that point you’re past the window where additional practice improves performance and into the window where fatigue starts hurting it. I don’t know exactly where that line is for every person, but most people who bomb after serious preparation cite going too hard the final 48 hours as a factor.
The night before: review your fit stories once. Remind yourself of the two or three frameworks you use most naturally. Sleep.
Walking into the interview: the main job is to stay curious about the problem rather than anxious about your performance. Easier said than done, obviously. But firms really do hire the candidates who seem like they’d be interesting to work with on a client problem, not just the ones who can reproduce a framework cleanly. Those aren’t always the same person.
If you’ve done 30 to 40 practice cases with real partners, you’re as ready as you can be. The rest is just whether the specific case they give you happens to play to your strengths or not, and that part is genuinely out of your control.