A Full Sample Case Interview, Annotated Line by Line

A sample case interview walks through a full round from start to finish: the prompt, the candidate's opening restatement, clarifying questions, a structure built and presented out loud, a quantitative question worked live, a curveball, and a final recommendation. Below is a complete one, using a composite candidate I'll call Jordan, an invented stand-in built from patterns I saw across hundreds of real interviews at McKinsey. I annotate every move Jordan makes, the ones that earn points and the ones that would cost them.

One honesty note before we start. Jordan is not a real person and this is not a transcript of an actual interview. It's a teaching composite, built to show you the shape of a strong round without pretending any single candidate said all of this. If you want the theory behind why each move works before you read the applied version, start with what a case interview actually is.

The prompt

Here's what the interviewer opens with. "Our client is Peak Fitness, a regional gym chain with 40 locations across three states. Over the last year, membership counts have stayed flat, but profits are down 20 percent. The CEO wants to know why, and what to do about it."

That's it. No extra hints. The candidate has to do the rest.

The opening, annotated

Jordan: "Before I jump in, can I take a moment to make sure I've got this right and think through how I'd like to approach it?"

This line alone tells the interviewer something. Jordan isn't rushing to sound smart. Asking permission to pause is a small, deliberate signal that this candidate has done this before.

Jordan: "Our client is Peak Fitness, a 40-location regional gym chain. Membership has held steady over the past year, but profit is down 20 percent, and the CEO wants to understand the cause and a path to fix it."

This is a restatement, not a summary. Jordan didn't just parrot the numbers back in the same order the interviewer gave them. Notice the shift: instead of listing facts, Jordan connects them into cause and effect language, flat membership plus falling profit, which frames the real puzzle immediately. That's the difference between reciting and synthesizing, and it's the single biggest tell in the first 30 seconds of a case.

A weak version of this same moment sounds like: "So membership is flat and profits are down 20 percent, got it." Technically correct, but flat. No structure, no signal that the candidate is already thinking about why the two numbers might not match.

Clarifying questions, annotated

Jordan: "Two quick questions before I structure my thinking. First, is this 20 percent decline specific to Peak Fitness, or is the broader regional gym market seeing something similar? Second, has anything changed operationally in the last year, new locations, new pricing, new competitors nearby?"

Both questions are aimed directly at the objective. Jordan isn't asking about the founding year of the company or how many employees work at headquarters. Every question either rules something in or rules something out before the structure gets built, which is exactly what the questions are for.

Interviewer: "It's specific to Peak Fitness. Two new low-cost competitors opened near six of their locations this year, and Peak Fitness added a new premium membership tier."

That's a gift, and a good candidate treats it as one. Jordan now has two live threads: outside pressure from new competitors, and an internal pricing change that might be cannibalizing revenue instead of adding to it.

The structure, annotated

Jordan: "Can I take a minute to structure my approach?"

[60 seconds of silence while Jordan writes.]

This is the one place in the whole case where silence is expected and even rewarded. A candidate who fills that silence with nervous chatter is usually building a worse structure than one who sits quietly and writes. If you want the full mechanics of how to use that minute well, I go deeper on it in how to structure a case interview.

Jordan: "I'd like to look at this through three lenses. First, revenue, split into pricing and volume, since we know pricing changed this year. Second, cost, to rule out whether expenses simply grew faster than sales. Third, the competitive picture, since we know two new entrants showed up near some locations. I'd like to start with revenue, since the two facts we already have both point there."

This is a built structure, not a memorized one. Jordan starts from profit equals revenue minus cost, which is automatically clean, nothing overlaps and together the branches cover the whole problem, then bends it toward the two clues already on the table. That's the tell of a candidate who understands frameworks as a starting point rather than a script. For the classic templates this gets built from, see the case interview frameworks guide.

A rejected version of this moment reaches for a generic four-box profitability framework and presents it exactly the same way it would for any other case, without ever mentioning the competitors or the new pricing tier. The interviewer notices instantly. It reads as rehearsed instead of built for this problem.

The quantitative question, annotated

Interviewer: "Let's dig into pricing. The new premium tier is priced at 40 dollars a month higher than the standard tier. 15 percent of members switched to it. Standard membership is 60 dollars a month. What's happened to average revenue per member?"

Jordan: "So before the change, everyone paid 60 dollars. After the change, 85 percent still pay 60, and 15 percent pay 100. That's 0.85 times 60, which is 51, plus 0.15 times 100, which is 15. So average revenue per member moved from 60 to 66 dollars, an increase of 10 percent."

Jordan narrates the arithmetic instead of announcing a final number, which lets the interviewer follow the logic and catch it immediately if a step is wrong. The math checks out, and the framing at the end, stating it as a percentage change, shows Jordan translating a raw number into something a CEO would actually care about.

Jordan: "That's interesting, because revenue per member went up, but total profit went down. That points me back to volume or cost, since the pricing move alone should have helped, not hurt."

This is the moment that separates a strong candidate from an average one. Jordan doesn't just answer the question asked, they notice the tension between the math they just did and the problem they're solving, and say so out loud.

The curveball, annotated

Interviewer: "One more thing. Corporate just told us that member cancellations have doubled at the six locations near the new low-cost competitors, but stayed flat everywhere else."

Jordan: "That fits with what we found on revenue per member. The pricing change helped where there's no cheap alternative nearby, but at the six affected locations, members may be leaving for the lower-cost option instead of upgrading, which would explain flat membership overall hiding a shift underneath, some locations gaining, some losing."

Jordan doesn't treat this as a new case. The new fact slots into the tree that was already built, competition was branch three from the start, which means the structure did its job. A candidate who panics here usually starts a fresh mental framework from scratch, which wastes time and looks disorganized to the interviewer watching the clock.

The recommendation, annotated

Interviewer: "We're near time. What would you tell the CEO?"

Jordan: "I'd recommend two things. First, keep the premium tier company-wide, since it's lifting revenue per member where competition isn't an issue. Second, at the six locations facing new low-cost competitors, introduce a defensive lower-priced option or a loyalty discount, since those locations are losing members to price, not to quality. The risk is that a discount tier could cannibalize the standard tier elsewhere if it's not restricted to the affected locations, so I'd pilot it at those six first before rolling it out further."

This lands the answer first, then the reasoning, then a named risk with a mitigation attached. That order matters. A partner listening for thirty seconds gets the recommendation immediately, then can keep listening for the "why" if they want it. A weak recommendation buries the answer at the end of a long walk through the math, which is exactly backward from how a real client conversation works.

What actually separates this from a rejected round

Nothing in Jordan's performance required genius-level math or an unusual framework nobody's seen before. What it required was consistency: synthesizing instead of repeating, tailoring the structure to the two clues already on the table, narrating math instead of announcing it, and treating a curveball as new data for an existing tree rather than a reason to panic. Every one of those is a habit you build through reps, not a trait you're born with. If you want a bank of prompts to build those reps on, these case interview questions are a good place to start, and practicing case interviews properly covers how to structure the drilling itself so it actually transfers to the real room.

The bottom line

A sample case interview is only useful if you read it for the pattern, not the specific numbers. Jordan's gym chain case works because every response does one job: restate, ask, structure, calculate, adapt, recommend, in that order, with no step skipped and no step padded. Copy that rhythm into your own practice, not the specific line about premium memberships, and if you want the full roadmap for building toward it, the case interview prep roadmap is the place to start.

Go deeper

Reading one annotated case gets you the shape of it. Building the actual skill, structuring on the fly, narrating math cleanly, adapting to curveballs without losing your tree, is what the CaseMap method inside the course is built to train.

Get the free resume module →

Get the complete Cut to the Case course →

One payment, lifetime access, 30-day guarantee.