Published 7 min read
How to Practice Case Interviews: Drills, Partners, and Reps
To practice case interviews well, work in this order: solo structuring drills first, then partner cases once your structures hold up under pressure, then full mock interviews in the final two weeks before your real ones. Most candidates skip straight to full partner cases, run ten of them badly, and wonder why they aren't improving. What actually moves the needle is the right case, in the right order, reviewed properly.
I conducted many interviews at McKinsey, and I could tell within the first two minutes which candidates had drilled the fundamentals and which ones had just done a lot of cases. The first group built structure like it was muscle memory. The second group had memorized the shape of five case types and hoped the real interview would look like one of them. It usually didn't.
Why most practice hours don't transfer
Doing a case with a friend feels productive. You talk for 30 minutes, you land on a recommendation, everyone feels good. But if you can't point to a specific skill that got better, that hour didn't build anything. It just burned a case you can no longer use.
The candidates who improve fastest treat practice like a skill-builder, not a rehearsal. They isolate one weakness, drill it in a short rep, get feedback, then move on. The candidates who plateau treat practice like a performance. They run the whole case start to finish every time, feel good about a smooth ending, and never notice that their structuring in minute two was actually the weak point.
If you don't know your own weak point yet, that's the first problem to solve, not case volume. Read what a case interview actually is if you're still fuzzy on the format, because you can't drill a skill you can't name.
Build a practice ladder before you drill
Before you touch a single case, decide what stage you're in. There are three, and skipping stages is the single biggest reason smart candidates waste weeks.
Stage one is isolated skill drills. You are not doing full cases yet. You are practicing one component, structuring, math, or synthesis, in short bursts by yourself.
Stage two is partner cases with a pause after every phase. You do a full case, but you stop after the opening, after the structure, after the math, and after the recommendation, and get feedback on each before moving to the next.
Stage three is full, uninterrupted mock interviews under time pressure, exactly like the real thing, with feedback saved for the end.
Most candidates start at stage three because it feels closest to the real interview. It's the last stage for a reason. Running full cases before your structure is solid just reinforces bad habits at speed.
Solo practice: the reps a partner can't give you
You don't need another person for the two things that break most candidates: structure and math.
For structure, take a prompt (you can pull one from any case book or write your own from a news headline about a company), and give yourself two minutes to build a full issue tree on paper. No talking, no partner, just the tree. Then check it against two questions. Does it fully cover the problem, and do the branches overlap anywhere. If you want the full logic behind building a tree from scratch instead of memorizing one, that's the whole subject of how to structure a case interview.
For math, drill without a calculator, out loud, narrating your approach before you compute. Case math rewards clean setup more than fast arithmetic. A candidate who says "I'll estimate this as 40,000 units at $12 average margin, so roughly 480,000 dollars" before doing the multiplication looks more in control than one who mutters and produces a number with no context.
Do fifteen minutes of tree-building and fifteen minutes of math drills most days for the first two weeks. That's the unglamorous part nobody posts about, and it's the part that makes the partner cases actually useful once you get there.
Partner practice: how to not waste the session
A partner session is only as good as the feedback that comes out of it, and most feedback is useless because it's vague. "Good job" and "that was a little messy" don't tell anyone what to fix.
Before you start, agree on what you're scoring. Structure, math, synthesis at the end, and communication throughout are the four buckets I used as an interviewer, and they still work as a scorecard between two candidates.
Give feedback in this order every time:
- One thing that worked, stated specifically, not "good energy" but "you signposted three branches before diving in."
- One thing to fix, with a concrete replacement, not "be clearer" but "restate the numbers before you set up your tree so I know you heard them."
- One thing to try next time, framed as an experiment, not a verdict.
If you don't have a case partner yet, consulting clubs, MBA case clubs, and online case-practice communities are the fastest ways to find one. Look for someone at a similar prep stage. Pairing with someone far ahead of you means you spend the session watching them think instead of building your own reps.
The four drills worth repeating
Not every practice format earns its time. These four consistently move candidates forward, and they map to the stages above.
| Drill | Stage | What it builds | Time |
|---|---|---|---|
| Timed tree-building, solo, no talking | Solo | Structure speed and MECE instinct | 10-15 min |
| Narrated math, solo, out loud | Solo | Clean setup, confident delivery | 15 min |
| Phase-paused partner case | Partner | Feedback on each stage before errors compound | 30-40 min |
| Full mock, uninterrupted, timed | Partner or solo-recorded | Stamina, composure under pressure | 45-60 min |
Rotate through these rather than defaulting to full cases every time. A week that's mostly tree-building and math drills, with two or three phase-paused partner cases, builds more real skill than five full cases run back to back with no structured feedback.
Common mistakes when practicing
The most common one is doing too many cases too early, before the fundamentals are automatic. Volume without a target skill just repeats whatever bad habit you already have.
The second is only practicing case types you're already comfortable with. If profitability cases feel easy and market sizing feels shaky, you should be doing more market sizing, not more profitability. Comfort during practice is a warning sign, not a good sign.
The third is skipping the review. If you finish a case and move straight to the next one without writing down what went wrong, you'll make the same structuring mistake in your fifth case that you made in your first. Keep a running list, even three bullet points after each session, of what to fix next time.
The fourth is practicing only with friends who are too kind to give real feedback. A partner who tells you everything was great isn't preparing you for an interviewer who's grading you against everyone else they've seen that season.
How many reps you actually need
There's no fixed number that works for everyone, and anyone giving you an exact case count is guessing. What matters more is the mix. A candidate who's done 15 well-structured, well-reviewed cases across different types is better prepared than one who's done 40 rushed ones without feedback.
As a rough shape: two to three weeks of daily solo drills to build the fundamentals, then two to three weeks layering in partner cases with paused feedback, then a final week or two of full uninterrupted mocks that simulate real interview pressure. If your timeline is shorter than that, compress the stages, don't skip them. For a full week-by-week plan that ties practice into the rest of your prep, the case interview prep roadmap walks through the whole sequence, and the frameworks guide is worth reviewing before your solo drills if you're still shaky on the standard starting points.
The bottom line
Practice that works has a shape: solo drills to build the fundamentals, phase-paused partner cases to catch mistakes early, then full mocks to build composure under pressure. Skip the shape and you'll rack up hours without moving your score, no matter how many cases you log. If you want the order laid out day by day instead of piecing it together yourself, the case interview prep roadmap is where I'd start.
Go deeper
The CaseMap method inside the course teaches the structuring approach your solo drills should be building toward, so your practice hours actually compound instead of just repeating.
Get the complete Cut to the Case course →
One payment, lifetime access, 30-day guarantee.
Frequently Asked Questions
How should I start practicing case interviews?
Start with solo drills, not partner cases. Spend the first couple of weeks on timed tree-building and narrated math practice by yourself, so your fundamentals are solid before you add the pressure of a partner and real-time feedback.
How many practice cases do I need before my interview?
There's no fixed number that works for everyone. Fifteen well-structured, well-reviewed cases across different case types beats forty rushed ones with no feedback, so focus on the mix and the review, not the count.
Do I need a partner to practice case interviews?
Not at first. Structure and math can both be drilled alone. A partner becomes valuable once your fundamentals hold up and you need real-time feedback and the pressure of talking through a case live, which usually means the second half of your prep window.
How do I find a case interview practice partner?
Consulting clubs, MBA case clubs, and online case-practice communities are the fastest routes. Look for someone at a similar prep stage as you, since pairing with someone much further along means you spend the session watching them think instead of building your own reps.
What's the biggest mistake people make when practicing cases?
Running full cases too early and too often without reviewing what went wrong. Volume without a specific target skill and without honest feedback just repeats whatever bad habit you already have, case after case.