McKinsey's interviewing page lists four practice cases, and two older ones are still live on the site. None of the six is the case most candidates drill first, where profits are down and you are asked why. Bain publishes two: a coffee-shop break-even and a five-year revenue decline that becomes a promotion choice by the second exhibit. The type you have rehearsed most is the one the firms publish least.
That gap tells you what a "type" is: not the industry, not the opening sentence. A case type is the number that will settle the decision, and the units that number comes in. Profitability closes on a bridge in dollars. Entry closes on a break-even volume against an accessible market. An acquisition closes on value created minus price paid. Name that number before you draw a tree and the tree mostly draws itself. By the end you will be able to do that in one sentence, and test it on a fact base run three ways.
What the firms actually publish
Firms do not publish frequency data on case types, and no office I know of runs a quota. What they publish is practice cases, the nearest thing to a primary source on what they consider representative. The tally as of 2026-09-24:
| Firm | Published case | The decision | The closing number |
|---|---|---|---|
| McKinsey | Beautify | Retrain in-store beauty consultants for virtual channels? | Payback on the training cost from incremental revenue |
| McKinsey | Electro-Light | Launch a low-sugar sports drink? | Break-even: $40 million fixed over $0.10 a bottle, and the 12.5% share that implies |
| McKinsey | Talbot Trucks | Build electric trucks? | The eTruck price at which total cost of ownership matches a €100,000 diesel truck |
| McKinsey | GlobaPharm | Buy a biologics start-up worth about $1 billion? | The Phase II success rate that breaks even on an extra $150 million trial |
| McKinsey | Diconsa | Deliver benefits and basic banking through 22,000 rural stores? | The saving to rural households from collecting benefits nearby |
| McKinsey | National Education | How does a country of 20 million fix its schools? | The outcome gap against comparison countries |
| Bain | Coffee Shop Co. | Open a coffee shop in Cambridge? | Break-even of 204,675 cups, about 3% of a 7,000,000-cup market |
| Bain | FashionCo | Why five years of falling revenue, and which fix? | Option B beats Option A by $50 million |
Six of the eight close on a threshold or a comparison: a break-even, a payback, a maximum price, a required success rate, a saving. Two open as a diagnosis, and neither is about profit: a school system, and a revenue decline that becomes a promotion choice within two exhibits. (McKinsey's last two are unlisted but were live on 2026-09-24.)
BCG's case interview preparation page describes the interview as structuring, questioning, analyzing data, quick calculation, and prioritizing; the firm also hosts three interactive case challenges: climate, digital strategy, IT architecture. Deloitte's case interview tips point to a practice tool whose cases, at last check, were five federal agencies and one commercial client. Again, no margin autopsy.
Two caveats. Published cases lean toward work a firm wants to be seen doing: pro bono, climate, inclusion. Live cases lean toward what the office sells. Neither is a random sample. But if your practice log is 60% profitability, no primary source supports that.
The type is the closing number
When a prompt lands, I am not asking "is this market entry?" I am asking what number would let the client say yes or no, and in what units. The tree is whatever moves that number, plus the two or three things that could veto it. Eight families, by closing number:
| Type | Closing number and units | Branches the interviewer is waiting for | Where the points leak |
|---|---|---|---|
| Profitability or turnaround | A profit bridge in currency: price, volume, mix, unit cost, fixed cost, each signed | Price and volume split before any cost talk; which costs should flex with volume | Naming a cost to cut before the bridge exists |
| Market entry or new product | Break-even volume or share against accessible demand, then payback in years | Accessible market, not total; channel; competitor response; capability gap | Quoting a total market where the accessible one belongs |
| M&A or build vs. buy | Standalone value plus synergies, minus premium and cost to achieve | Synergy timing; integration risk; the build or partner alternative | Booking 100% of synergies in year one |
| Pricing | Contribution at old and new price; break-even volume change in percent | Willingness to pay by segment; competitor reaction; discount leakage | Volume up, profit down |
| Growth | Incremental revenue by lever, capped by capacity | Core vs. adjacent vs. new; sales capacity; cannibalization | A list of ideas with no constraint |
| Operations | Units per hour at the bottleneck, or cost per unit | Demand vs. capacity vs. yield; where the minutes go | A customer framework on a factory problem |
| PE diligence | Multiple of money or IRR under a stated plan | EBITDA sustainability; exit multiple; leverage; what the plan assumes | A strategy case with no price and no clock |
| Public sector or nonprofit | Cost per outcome: per beneficiary, per point, per minute of waiting | The success metric the client named; feasibility; equity; who pays | A profit tree forced onto a service mandate |
The units are the test. "Market attractiveness" has none and settles nothing; "130 transactions a day against a lobby of 3,000 people" does.
Ned's rule. If you cannot tell me, before the first exhibit, what number would settle the case and what its units are, you have not identified the type. You have identified the industry.
Profitability is a subroutine, not a type
The profit bridge is the most useful 90 seconds of arithmetic in consulting, which is exactly why it rarely gets a case to itself. Beautify asks whether retraining consultants could be profitable: a payback inside a channel decision. Electro-Light's $0.10 a bottle is unit economics inside a launch. In diligence it sits inside a return calculation with a price and a clock.
In my own first-round scoring, perhaps one case in four opened as a margin problem, and about half of those had become a pricing, capacity, or portfolio decision by the second exhibit. That is an estimate, not a count. Either way: drill the bridge until it is fast and boring, then stop treating it as a destination.
One company, three prompts
Harlow & Finch is invented, numbers included; every figure reproduces with a pencil.
| Harlow & Finch, 40 bakery-cafés | Two years ago | Now |
|---|---|---|
| Revenue | $60.0M | $60.0M |
| Transactions | 6.25M | 6.00M |
| Average ticket | $9.60 | $10.00 |
| Ingredients (% of revenue) | 30% | 32% |
| Labor (% of revenue) | 30% | 32% |
| Rent and other (% of revenue) | 28% | 28% |
| EBITDA | $7.2M (12%) | $4.8M (8%) |
Prompt A: "Margin fell four points. Why?"
Closing number: a $2.4 million bridge from $7.2 million to $4.8 million.
| Bridge line | Calculation | Effect |
|---|---|---|
| Price | 6.00M transactions x ($10.00 - $9.60) | +$2.4M |
| Volume | (6.00M - 6.25M) x $9.60 | -$2.4M |
| Ingredients | 2 points x $60M | -$1.2M |
| Labor | 2 points x $60M | -$1.2M |
| Rent and other | No change | $0 |
| Net | -$2.4M |
What the interviewer wants to hear: revenue is flat only because a 4% price rise bought back what 250,000 lost transactions took away, and the two cost lines that rose are the two that should have flexed with volume. Ingredient cost per transaction rose from $2.88 to $3.20, up 11%: inflation or waste? Labor at flat dollars on fewer transactions is a scheduling problem. Nobody has said "cut marketing."
Prompt B: "Should Harlow & Finch put ten kiosks in hospital lobbies?"
Closing number: transactions per kiosk per day to break even, then payback.
Assume a $150,000 fit-out, $200,000 a year of fixed cost (seven-day staffing plus the hospital's fee), a $7 ticket, and 40% variable cost: $4.20 of contribution per transaction.
- Break-even: $200,000 / $4.20 = 47,619 a year, about 130 a day on a site that never closes.
- At 200 a day: 73,000 x $4.20 = $306,600, less $200,000, leaves $106,600 a year. Payback on $150,000 is about 17 months.
- At 150 a day: 54,750 x $4.20 = $229,950, less $200,000, leaves $29,950. Payback is five years.
The case now lives between 150 and 200 a day, so the branches are lobby footfall, capture rate (200 of 3,000 people passing through is under 7%), the hospital's own cafeteria and vending, and whether a company whose café labor already fails to flex can staff a 365-day kiosk. Prompt A's bridge matters here only as a judgment point about labor.
Prompt C: "A sponsor can buy Harlow & Finch for $38.4 million. Should it?"
Closing number: multiple of money and IRR under a stated plan.
$38.4 million is 8.0x current EBITDA. Assume half debt, half equity ($19.2 million each), a five-year hold, exit at 8.0x, and $8 million of debt repaid in either scenario (a simplification).
| Scenario | Exit EBITDA | Exit value | Equity out | Multiple | IRR |
|---|---|---|---|---|---|
| Margin stays at 8% | $4.8M | $38.4M | $27.2M | 1.4x | About 7% |
| Margin returns to 12% | $7.2M | $57.6M | $46.4M | 2.4x | About 19% |
The whole diligence is the probability of getting four margin points back: Prompt A again, now a subroutine inside a return calculation with a price and a clock attached. Same company, same numbers, three closing numbers, three trees.
What the scoring sheet actually rewards
Every scoring sheet I have used had four to six rows: structure, quantitative work, judgment, synthesis, sometimes communication. None had a row for "identified the case type." The type earns points only through the structure row, and only when it makes the structure more specific. Two openings to Prompt B, scored on a five-point scale that is mine, not any firm's:
| Row | "It's a market entry case, so I'll look at market, competition, capabilities, and financials." | "A kiosk needs roughly 130 transactions a day. I'd size footfall and capture rate first, then 365-day staffing given our cafés already struggle to flex labor, then lease terms and the hospital's alternatives." |
|---|---|---|
| Structure | 2 of 5: complete but interchangeable with any entry case | 4 of 5: ordered by what settles the decision, veto conditions named |
| Judgment | 1 of 5: nothing specific to this client yet | 4 of 5: connected the labor finding to the new format unprompted |
The second candidate did nothing clever: they named the closing number, then listed what moves it and what could veto it.
Practice this today
Take the eight published cases in the first table. Without opening them, write the closing number and its units for each, 20 seconds apiece. Then open them and check. Expect to miss two; the misses show which family you have been pattern-matching by industry.
Then do it live. Start a five-minute rep: three typed turns with instant scores, no account needed; the first score you see is structure. Follow with a structure drill, where the only task is turning a prompt into a tree that points at a number. For a full case, the live voice case with Ned runs about 18 minutes with a seven-score debrief; the case library has 64 original cases across every family above, including the one you have been avoiding.
Everything is open for seven days, no card; then $120 for a recruiting season or $49 a month. CoachNed is independent and not affiliated with McKinsey, BCG, Bain, or Deloitte; the published cases are theirs, the bakery is ours.
Frequently asked questions
What are the most common types of case interview?
Nobody outside a firm has frequency data, and firms do not publish it. Their own practice cases, tallied above, are dominated by "should we do this" decisions ending in a break-even, payback, or return, plus a steady share of public and social sector work. Pure profit diagnoses are rarer than prep folklore suggests, though the bridge appears inside many other cases.
Is a written case or a chatbot case like BCG's Casey a different type?
No, it is a different format. Written, interviewer-led, candidate-led, and chatbot cases change who drives and how data arrives; the closing number underneath is still one of the eight above. CoachNed's Casey-style UrbanBrew case lets you feel that format.
How do I identify the case type when the prompt is vague?
Ask what decision the client has to make, and by when. The answer almost always contains the closing number: "we want to understand the market" becomes "enter or not," which becomes break-even against accessible demand. If the interviewer will not narrow it, state your assumption aloud and build to that.
Sources
- Interviewing at McKinsey — the four listed practice cases. Checked 2026-09-24.
- McKinsey: Beautify — payback on virtual-channel training. Checked 2026-09-24.
- McKinsey: Electro-Light — $40 million fixed, $2 price, $1.90 cost, 12.5% share. Checked 2026-09-24.
- McKinsey: Talbot Trucks — TCO against a €100,000 diesel truck. Checked 2026-09-24.
- McKinsey: GlobaPharm — $1 billion target; $150 million Phase II break-even. Checked 2026-09-24.
- McKinsey: Diconsa — 22,000 stores; cost of collecting benefits. Checked 2026-09-24.
- McKinsey: Transforming a national education system — Loravia, population 20 million. Checked 2026-09-24.
- Bain: Coffee Shop Co. — 204,675 cups, about 3% of 7,000,000. Checked 2026-09-24.
- Bain: FashionCo — five years of decline; $50 million between options. Checked 2026-09-24.
- BCG case interview preparation — what the interview tests; interactive case challenges. Checked 2026-09-24.
- Deloitte case and scenario interview tips — five-step approach; link to the practice tool. Checked 2026-09-24.