Case Interview Guide: How to Structure, Solve, and Talk Through a Consulting Case

A case interview is the only common interview format where saying "I don't know — but here's how I'd figure it out" is exactly the right answer. Consulting firms, and increasingly private equity funds and corporate strategy teams, use cases because the job itself is being handed ambiguous problems with incomplete data and a client watching you think. The interview is a 30-minute work sample, not a quiz.

That's good news, because work samples can be trained. This guide covers what interviewers actually score, why memorized frameworks backfire, a worked mini-case with annotated dialogue, a market-sizing example with real numbers, and a practice plan that gets you to interview-ready.

Table of Contents

Why Firms Use Case Interviews

Interviewers score four things, and it helps to know the rubric:

  • Structure. Can you break a messy problem into parts that together cover it? This is weighted heaviest.
  • Quantitative comfort. Percentages, break-evens, and market math, done aloud, without a calculator.
  • Business judgment. Do your hypotheses sound like someone who understands how companies make money?
  • Communication. Do you think out loud in organized sentences, and can you give a recommendation a CEO would sit still for?

Notice what's not on the list: getting "the right answer." Most cases have several defensible answers. Two candidates can reach opposite recommendations and both get offers, because the recommendation matters less than the reasoning that produced it.

What a Case Interview Looks Like

A typical case runs 25–40 minutes in five beats:

  1. The prompt. "Your client is a regional coffee chain whose profits have fallen 20% in two years. What's going on, and what should they do?"
  2. Clarifying questions. Two or three, to pin down the objective and scope — not a fishing expedition.
  3. Structure. You ask for a minute, sketch your breakdown on paper, and present it.
  4. Analysis. The interviewer feeds you data — charts, tables, facts — as you work down your structure. This is the bulk of the time.
  5. Recommendation. A crisp, answer-first close: what the client should do, why, the risks, and the next steps.

Interviewer-led cases (McKinsey's style — see our McKinsey interview guide) steer you through set questions; candidate-led cases (BCG, Bain, most boutiques) leave the driving to you. The skills are identical; the difference is who holds the map.

Many firms also screen with cognitive tests before or alongside cases — if you've been invited to one, our CCAT guide covers the most common format.

Frameworks vs. Structured Thinking

Every case prep book teaches frameworks: profitability trees, the 4Ps, Porter's Five Forces, market entry checklists. Learn them — as vocabulary. Then understand why reciting them fails.

A framework is someone else's structure for a generic problem. Interviewers hear "I'd look at Product, Price, Place, and Promotion" delivered identically to a profitability case, a market entry case, and a merger case, and they conclude — correctly — that the candidate is pattern-matching, not thinking. Worse, generic frameworks routinely miss the actual driver: the 4Ps have nothing to say about a client whose costs exploded.

Structured thinking means building a breakdown for this specific question that is MECE — mutually exclusive, collectively exhaustive. No overlaps, no gaps. The reliable method:

  1. Anchor on the objective. Profit problem? Start from Profit = Revenue − Costs, because that equation must contain the answer.
  2. Break each branch into drivers. Revenue = customers × visits × spend per visit. Costs = fixed + variable.
  3. Tailor to the client. A coffee chain's revenue splits naturally by daypart or by store cohort; a software firm's by new vs. renewal. Using the client's real levers is what separates a structure from a template.
  4. State a hypothesis. "Given that the market is growing, I'd guess this is competitive share loss rather than category decline — I'd like to test revenue per store first."

The math-tree anchor guarantees exhaustiveness; the tailoring shows judgment. That combination is what "good structure" means on the scoring sheet.

Worked Mini-Case With Annotated Dialogue

Prompt: "Your client is FreshBrew, a 40-store coffee chain in the Midwest. Profits are down 20% over two years. The CEO wants to know why, and what to do."

Candidate: "Before I structure this — has the profit decline been even across stores, and has the CEO ruled anything out?"

Interviewer: "Good question. The decline is concentrated in urban stores. Nothing's been ruled out."

Annotation: One sharp clarifying question just narrowed the problem by half. Note the candidate asked about the shape of the problem, not for the answer.

Candidate: "Can I take a minute to lay out an approach?" (sketches) "Profit is revenue minus cost, so I'd examine both, focused on urban stores. On revenue: number of customers, visit frequency, and average ticket. On costs: fixed — mainly rent — and variable — labor and ingredients. Since the decline is urban-specific, my hypothesis is it's driven by something that differs by geography: urban rent escalation or new urban competition. I'd like to start with revenue per store."

Annotation: MECE tree, anchored on the profit equation, tailored with a geographic hypothesis. Under 45 seconds.

Interviewer: "Here's data: urban revenue per store fell from $900K to $720K. Costs per store are flat. Two national chains opened 15 urban locations nearby in the period."

Candidate: "So this is a revenue problem — a $180K per-store decline, or 20%, matching the profit drop since costs held. New competition is the likely driver. I'd want to know whether we're losing customers entirely or losing visits — do we have traffic data?"

Annotation: The candidate synthesizes each data point back into the structure before asking for the next one. That rhythm — take data, interpret, connect, proceed — is what interviewers mean by "driving the case."

Interviewer: "Transactions per store are down 25%; average ticket is actually up slightly. We have 90 seconds — what do you tell the CEO?"

Candidate: "FreshBrew's profit decline is a customer-traffic problem in urban stores, driven by new national competitors — we're losing visits, not price. I'd recommend a retention-first response: launch a loyalty program to protect the remaining base, differentiate on what nationals can't match — speed and local identity — and reassess the three worst-performing urban leases rather than defending every location. Main risk: a price war, which we should avoid because our ticket is holding. First step: 60-day loyalty pilot in the five most affected stores."

Annotation: Answer first, reasoning second, risk and next step last. No "it depends."

Market Sizing: A Worked Guesstimate

Sizing questions ("How many cups of coffee are sold in Chicago per day?") test whether you can build an estimate from defensible assumptions. The method: segment, assume round numbers, multiply, sanity-check.

How many cups of coffee are sold in Chicago daily?

  1. Chicago proper: ~2.7 million people → call it 2.5M to keep math clean, plus commuters ≈ 3M daytime population.
  2. Segment by drinking behavior: ~60% drink coffee (1.8M people). Of those, ~two-thirds average 1 cup/day and one-third average 2 → 1.2M + 1.2M = 2.4M cups.
  3. Adjust for sold vs. home-brewed: assume 40% of cups are purchased → ~1 million cups sold per day.
  4. Sanity-check: 1M cups across maybe 2,500 outlets ≈ 400 cups per outlet per day — about one every 90 seconds over a 10-hour day. Plausible for a mix of busy and quiet locations. Done.

Say every assumption out loud and invite challenge ("I'm assuming 60% coffee drinkers — happy to flex that"). The interviewer is scoring the scaffolding, not the final number; anything within a factor of two or three with clean logic passes.

Practice Plan

  • Weeks 1–2 — Foundations. Learn the core structures (profitability, market entry, pricing, M&A) as starting materials. Drill mental math daily: percentages, break-evens, big-number multiplication. Ten minutes a day beats one weekend binge. Timed numerical reasoning practice builds exactly this speed-under-pressure muscle.
  • Weeks 3–5 — Live cases. Case with a partner, alternating interviewer and candidate — interviewing someone else teaches you the rubric faster than anything. Target 15–20 cases, with written feedback on structure, math, and communication after each.
  • Week 6 — Polish. Record yourself. Fix filler words, un-hedge your recommendations, rehearse 60-second closes. Do two full mock interviews with the fit portion attached — firms score both, and the common interview questions deserve real reps too.

Twenty to forty total live cases is the norm among successful candidates. The 41st case teaches you less than honest feedback on the 15th.

Common Mistakes

  • Framework recitation. Announcing "I'll use Porter's Five Forces" is a rejection speed-run. Build the structure from the question.
  • Silent math. Long pauses while you calculate privately read as being lost. Narrate: "720 from 900 is 180, so a 20% decline."
  • Data hoarding. Receiving a chart and moving on without interpreting it. Every exhibit exists to change your thinking — say how.
  • The "it depends" close. You will never have complete information. Commit to a recommendation and own the risks.
  • Ignoring the fit interview. Half the session is behavioral. A brilliant case with a mumbled "why consulting?" still ends in rejection.
  • Practicing solo only. Reading casebooks feels productive and transfers poorly. The skill is conversational; train it in conversation.

FAQ

How long does a case interview take?

Usually 25–40 minutes, paired with 10–15 minutes of fit questions in the same session.

Do I need to memorize frameworks like the 4Ps or Porter's Five Forces?

Learn them as vocabulary, never as scripts. Interviewers reject candidates who force generic frameworks onto specific problems; build a MECE structure tailored to the actual question.

How much math is in a case interview?

Constant but not advanced: percentages, ratios, break-evens, market sizing, done by hand while talking. Speed and composure matter more than mathematical sophistication.

How many practice cases should I do?

Most successful candidates do 20–40 live cases with partners. Feedback quality beats case count.

What if I get stuck during a case?

Say so and return to your structure aloud. Recovering in an organized way is itself scored — often more favorably than never stumbling.

Conclusion

Case interviews reward a learnable craft: anchor on the objective, build a structure that covers the problem, narrate your math, and close with a recommendation you're willing to defend. Do the live reps and the format loses its mystery. Since most consulting and finance processes also include aptitude screening, sharpen the underlying speed with numerical, abstract, and verbal reasoning practice — or cover every test you might face with the All-Test Package.

About the Author

Ingmar van Maurik is a career and assessment preparation expert who helps candidates improve their interview and test performance.

References

  • Assessment-Training.com
  • McKinsey & Company careers
  • Glassdoor