25 Case Interview Questions to Practice Before Your First Round
Updated: Oct 6
The video call opens at 10:00 a.m. The interviewer, a consultant two years out of college, reads the prompt once: a regional gym chain, flat revenue, profits down by a third. Then she stops talking and waits. The next 30 seconds set up the whole case, and the only way to be calm in that silence is to have answered case interview questions like this one out loud, many times, before the real one.
I'm Stephen Turban. I went through McKinsey's on-cycle recruiting after Harvard, worked at McKinsey, and now run WallStreetGuide, where I help undergrads prepare for MBB, tier 2, and Big 4 strategy interviews. The students who walk into first rounds relaxed have already seen every major case type at least once.
Below are 25 practice cases grouped by type. Each has the prompt, the clarifying questions worth asking, a structure in plain words, the math worked step by step with the numbers an interviewer would hand you, and what a strong recommendation sounds like. Try each case on your own before reading the math, because reading a solved case feels like learning but doesn't build the skill. They're original prompts in the style firms use, so you can run them with a partner tonight.
Once you have worked through all 25, these free case consulting prep resources are the easiest way to keep practicing without paying for a course. If McKinsey is on your list, its interviewer-led format changes how you should run a case, as explained in this guide to the McKinsey case interview.
What kinds of case interview questions do consulting firms ask?
Most case interview questions fall into five core types: profitability, market entry, M&A, operations or cost reduction, and market sizing. Firms also mix in growth, pricing, new product, public sector, and exhibit-reading questions, which borrow the same building blocks. Interviewers want to see you break a vague problem into a few clear pieces, do the math without a calculator, and land on a recommendation. Consultants call that split an issue tree: smaller questions that don't overlap and together cover the whole problem. The prompt changes; the skill being graded doesn't.
How many practice cases should you do before your first round?
Quality of practice matters more than the number of cases, because a case you review carefully teaches you more than three you rush. Aim for enough live cases with a partner that you've seen every type below once and your weakest type two or three times. For MBB internships in Summer 2028, applications open in early 2027 (January through March for McKinsey and Bain; late spring for BCG), with first rounds following throughout the spring and early summer.. A sophomore who starts practicing this fall has time to see every case type twice before applications open.
Profitability case interview questions
Every profitability case starts with the same split: profit equals revenue minus costs, and revenue equals price times volume. The skill is finding which piece moved.
1. Profits are down at a regional gym chain
The gym chain case tests whether you split revenue into price and volume first.
The prompt: Our client runs 120 gyms across the Midwest. Revenue has been flat at $180 million for three years, but operating profit fell from $27 million to $18 million. What's going on, and what should they do?
Clarifying questions:
Is operating profit measured after rent and staff but before interest and taxes?
Has the number of gyms changed?
Has a low-price competitor entered the market?
Structure: Profit is revenue minus costs. Revenue is members times monthly dues, plus extras like personal training. Costs are fixed (rent, equipment, managers) or variable, rising with each member (cleaning, towels, payment fees). Flat revenue is the clue here, because it usually means something inside revenue shifted while the total held still.
The math:
Three years ago: 250,000 members paying $60 a month. 250,000 × $60 × 12 = $180 million.
Today, after a budget chain opened and the client cut dues: 300,000 members paying $50. 300,000 × $50 × 12 = $180 million. Same total, 20 percent more members at a 17 percent lower price.
Costs rose $9 million, from $153 million to $162 million. Each member costs $10 a month to serve, so 50,000 extra members × $10 × 12 = $6 million. The other $3 million is higher rent on renewed leases.
Margin fell from 15 percent ($27M ÷ $180M) to 10 percent ($18M ÷ $180M), and the whole $9 million drop traces to the $10 price cut.
A strong recommendation: "Profit fell $9 million: a $10 dues cut brought 50,000 members who added no revenue but cost $6 million a year, and lease renewals added $3 million. Keep $50 as the base and add a $65 premium tier. If 30 percent upgrade, 90,000 × $15 × 12 = $16.2 million a year covers the gap."
When revenue is flat, check whether price and volume moved in opposite directions before you look at costs.
2. A craft brewery grows volume but loses profit
The brewery case tests whether you spot a product mix shift behind volume growth.
The prompt: Our client is a regional craft brewery that sells to grocery stores. Case volume grew 20 percent last year and revenue grew too, but gross profit fell. Why?
Clarifying questions:
How many product lines are there, and did prices change?
Did the cost to brew a case change?
Did any large grocery customer change its orders?
Structure: Gross profit means revenue minus the cost of making the beer. Split it by product, cases sold times profit per case, last year against this year. Total volume can grow while gross profit falls if the extra cases come from the beer that earns the least.
The math:
IPA sells for $36 a case and costs $18 to make, so it earns $18 a case. Light lager sells for $24 and costs $19, so it earns $5 a case.
Last year: 100,000 cases of IPA and 100,000 of lager. Gross profit = (100,000 × $18) + (100,000 × $5) = $1.8M + $0.5M = $2.3 million. Revenue = $3.6M + $2.4M = $6.0 million.
This year: 80,000 IPA and 160,000 lager, 240,000 cases total, which is the 20 percent growth. Gross profit = (80,000 × $18) + (160,000 × $5) = $1.44M + $0.8M = $2.24 million. Revenue = $2.88M + $3.84M = $6.72 million, up 12 percent.
Gross profit fell $60,000. To close it with a lager price increase: $60,000 ÷ 160,000 cases = $0.375 (about 38 cents a case).
A 38 cent increase per case of lager, under 2 percent of its price, closes the whole $60,000 gap.
A strong recommendation: "Gross profit fell $60,000 because the mix moved toward light lager, at $5 a case against $18 for IPA. A 38 cent increase on lager closes the gap, but the real question is why IPA fell 20 percent. If a grocery chain cut its shelf space, winning that back beats any price move."
Volume growth can hide a profit problem when the growth comes from your lowest-margin product.
3. A pizza chain's delivery apps eat its growth
The pizza chain case tests whether you reconcile every change to the exact drop.
The prompt: Our client owns 40 pizza restaurants in the Southeast. Revenue grew from $50 million to $54 million over two years, but profit dropped from $5 million to $3 million. What happened?
Clarifying questions:
Did any restaurants open or close?
Did menu prices change?
How do the delivery apps charge the client?
Structure: Start from last year's $5 million and add up everything that changed: revenue growth adds profit, any cost that rose takes some away. The pieces have to sum to exactly $3 million, and if they don't, you've missed a cost.
The math:
Before app fees, each extra dollar of sales leaves 40 cents after ingredients, hourly labor, and packaging. This is the contribution margin: what a sale contributes after the costs that come with it. $4 million of new revenue × 40 percent = $1.6 million more profit.
Delivery apps charge a 25 percent commission. They went from 10 percent of sales to 30 percent. Before: $50M × 10% = $5M of app sales × 25% = $1.25M in commissions. After: $54M × 30% = $16.2M × 25% = $4.05M. Increase: $2.8 million.
Cheese prices added $0.8 million in costs.
Reconcile: $5.0M + $1.6M minus $2.8M minus $0.8M = $3.0 million.
The four pieces tie out to exactly $3 million, and the $2.8 million jump in app commissions is the whole story.
A strong recommendation: "App commissions grew $2.8 million and wiped out the $1.6 million of profit from new sales. Push customers to order direct with a 10 percent discount. If half of app sales, $8.1 million, move direct, the client saves $2.03 million in commission, gives up $0.81 million in discounts, and nets about $1.2 million."
A cost that scales with sales, like an app commission, can eat all of your revenue growth and more.
Market entry case interview questions
A market entry case asks two questions in order: is the market worth entering, and can this client win in it. The third, which most candidates forget, is how exactly the client gets in.
4. A Texas car wash chain looks at Arizona
The car wash case tests single-site economics and the choice of how to enter.
The prompt: Our client operates 60 express car washes in Texas and is considering expanding into Arizona. Should they?
Clarifying questions:
What payback period would make the client say yes?
Would they build new sites, buy existing ones, or partner?
Does the Texas business rely on monthly memberships?
Structure: Is Arizona attractive, can the client win there, what does one site earn, and how should they enter? The economics of a single site decide most market entry cases, so build that branch first.
The math:
A new site costs $4 million to build. It does 250 washes a day at $12 each: 250 × $12 = $3,000 a day. Over 360 open days, $3,000 × 360 = $1.08 million a year
Operating costs are $600,000, so each site earns $480,000 a year.
Payback, the years it takes profit to repay the investment: $4,000,000 ÷ $480,000 = about 8.3 years.
Buying instead: independent washes earning $500,000 a year sell for about 6 times annual profit, or $3 million.
Buying pays back in 6 years ($3M ÷ $0.5M) against about 8.3 for building, so acquisition wins on the numbers.
A strong recommendation: "Buy existing washes in Arizona rather than building. A new site pays back in about 8 years; an acquired one in about 6, with customers on day one and a quick switch to our membership model. The risk is overpaying if other chains bid, so I'd set a walk-away price of 6 times profit."
A strong market entry answer names the way in, whether that's building, buying, or partnering, along with the yes or no.
5. A European sparkling water brand considers the US
The sparkling water case tests compounding math and competitive realism in a market entry.
The prompt: Our client is a well-known European sparkling water brand with no US presence. Should they launch in the United States?
Clarifying questions:
Why the US, and why now?
Grocery stores, restaurants, or both?
What market share would count as success, and by when?
Structure: Market attractiveness (size, growth, competition), client fit (brand, positioning, capacity), financials (investment, revenue, payback), and go-to-market (cities, channel, partner). The financial branch decides this case, because the target share is small and the launch cost is large.
The math:
The interviewer says the US market is $4 billion today and grows 8 percent a year. The top three brands hold 70 percent of it. The client targets 3 percent share in year five.
Five years of 8 percent growth: 1.08 × 1.08 = 1.1664. Squared again: 1.1664 × 1.1664 = about 1.3605. One more year: 1.3605 × 1.08 = about 1.469.
Year five market: $4B × 1.469 = about $5.9 billion. 3 percent of that is about $176 million in revenue.
At a 15 percent operating margin, that's about $26 million a year of profit.
Against $150 million of launch spending, 26 million a year takes about 5.8 years to pay back (150M ÷ $26M), which is slow"
A strong recommendation: "Launch, but narrowly. A 3 percent share in year five is worth about $26 million a year, and paying back $150 million takes close to six years after that. I'd start in three coastal metros through an existing distributor, with a checkpoint at year two: if share isn't above 1 percent there, stop."
Compounding shows up in market entry cases, so practice multiplying 1.08 by itself until it feels routine.
6. Dental scheduling software eyes veterinary clinics
The vet software case tests whether you stress-test the assumption everything rests on.
The prompt: Our client sells appointment scheduling software to 20,000 US dental offices. Should they build a version for veterinary clinics?
Clarifying questions:
How different is a vet clinic's day from a dental office's?
Who already sells scheduling software to vets?
Is the goal revenue growth or using spare engineering capacity?
Structure: Market (clinics and what they use today), fit (how much of the product works as is), economics (price, adoption, cost to build and sell), and risks. Adoption is the assumption everything else rests on, so the structure needs a way to test it.
The math:
The interviewer gives you: 30,000 vet clinics, a price of $300 a month, expected adoption of 10 percent in three years, $5 million to adapt the product, $2,000 in sales cost per clinic won, and $50 a month to serve each clinic.
Clinics won: 30,000 × 10% = 3,000.
Yearly revenue: 3,000 × $300 × 12 = $10.8 million. Yearly serving cost: 3,000 × $50 × 12 = $1.8 million. Yearly contribution: $9.0 million.
One-time costs: $5M + (3,000 × $2,000) = $5M + $6M = $11 million. Payback: $11M ÷ $9M = about 1.2 years once the client has 3,000 clinics.
Stress test at half the adoption, 1,500 clinics: contribution $4.5 million, one-time cost $5M + $3M = $8 million, payback about 1.8 years.
Payback is about 1.2 years at 10 percent adoption and about 1.8 years at 5 percent, so the case survives its own stress test.
A strong recommendation: "Build it. Even at 5 percent adoption, payback is under two years. I'd pilot with a few hundred clinics first to confirm the adoption rate, because every number here rests on it."
Test your recommendation against the one assumption that matters most, then say whether it survives.
Growth strategy case interview questions
Growth cases reward candidates who size each growth lever in dollars instead of listing ideas. Sell to more people, sell more to the same people, or buy growth.
7. A salon chain wants to double revenue
The salon case tests whether you turn a vague goal into a growth rate.
The prompt: Our client runs 300 hair salons across the US with $300 million in revenue. The CEO wants to double revenue in five years. How?
Clarifying questions:
Does the growth need to be profitable?
How many salons has the client opened per year historically?
Is franchising or buying other chains on the table?
Structure: Organic growth (more customers, more visits, higher spend per visit, new salons) and inorganic growth (buying other chains). Turning "double in five years" into a yearly growth rate is the first move, because it tells you how much each lever has to carry.
The math:
Doubling in five years needs about 15 percent growth a year. Quick check with the rule of 72, a shortcut where 72 divided by the growth rate gives the years to double: 72 ÷ 15 = 4.8 years. Exact check: 1.15 × 1.15 = 1.3225, squared is about 1.749, times 1.15 is about 2.01.
If existing salons grow 5 percent a year: 1.05 to the fifth power is about 1.276, so revenue reaches $300M × 1.276 = about $383 million.
Gap to $600 million: $600M minus $383M = $217 million. Each mature salon does $1 million, so the client needs about 217 more salons, or roughly 43 a year.
The interviewer says the client has never opened more than 15 salons in a year.
Organic growth leaves a $217 million gap, and closing it needs 43 new salons a year against a best-ever pace of 15.
A strong recommendation: "Existing salons get the client to about $383 million. The remaining $217 million needs roughly 200 more salons, three times the client's best opening pace. I'd open 20 a year for 100 total, buy one regional chain of about 100, and push color services and product sales to lift spend per visit."
Size each growth lever in dollars so you can show which ones actually close the gap.
8. A kids' reading app wants more revenue
The reading app case tests funnel math and spotting the biggest lever.
The prompt: Our client makes a reading app for children. It has 2 million free users and 50,000 paying subscribers at $8 a month. How can it grow revenue?
Clarifying questions:
Who pays: parents, schools, or both?
How long does the average subscriber stay?
How do competitors' conversion rates compare?
Structure: Revenue equals free users times conversion rate (the share of free users who pay) times price times months retained: four levers. Sizing each lever in dollars shows which one has the most room, and that's the one to pull first.
The math:
Today: 50,000 × $8 × 12 = $4.8 million a year. Conversion: 50,000 ÷ 2,000,000 = 2.5 percent. The interviewer says competitors convert 4 to 5 percent.
Raise price to $10, lose 10 percent of subscribers: 45,000 × $10 × 12 = $5.4 million. Gain: $0.6 million.
Raise conversion to 4 percent: 2,000,000 × 4% = 80,000 subscribers. 80,000 × $8 × 12 = $7.68 million. Gain: $2.88 million.
Grow free users 50 percent to 3 million at 2.5 percent conversion: 75,000 × $8 × 12 = $7.2 million. Gain: $2.4 million, but it costs marketing money.
Lifting conversion to 4 percent adds $2.88 million a year, more than either a price increase or a bigger user base.
A strong recommendation: "Focus on conversion first. Moving from 2.5 to 4 percent, still below competitors, adds about $2.9 million a year without spending on new users. I'd test a seven-day free trial and a cheaper annual plan, and raise prices once parents see more value."
The best growth lever is usually the one with the biggest gap to competitors, not the most obvious one like price.
M&A case interview questions
Every M&A case comes down to one test: is the target worth more to the client than the price, including synergies, and is there a better use of the same money? Synergies are the extra profit the two companies make together that neither makes alone.
9. A private equity firm eyes a vet clinic chain
The vet clinic buyout tests multiples and working backward from a return target.
The prompt: Our client is a private equity firm looking at buying a chain of 50 veterinary clinics for $300 million. Should they do it?
Clarifying questions:
What return does the firm need, and over how many years?
Should we ignore debt financing for now?
How proven is the plan to open new clinics?
Structure: What is the chain earning now, what will it earn at exit, what will it sell for, and does that clear the return target? A financial buyer's question is simply whether exit value divided by price meets its return target.
The math:
The chain's EBITDA is $25 million. EBITDA is earnings before interest, taxes, depreciation, and amortization, a rough measure of the cash the business generates from operations. Price ÷ EBITDA = $300M ÷ $25M = 12 times. That "12 times" is the multiple.
The plan grows EBITDA to $40 million in five years. Selling at the same 12 times: $40M × 12 = $480 million.
Money back: $480M ÷ $300M = 1.6 times in five years, about 10 percent a year (1.1 to the fifth power is about 1.61).
The firm wants 2.5 times, about 20 percent a year. That needs an exit value of $300M × 2.5 = $750 million, or EBITDA of $750M ÷ 12 = $62.5 million.
The most the firm can pay and still earn 2.5 times on the current plan is $192 million ($480M ÷ 2.5), far below the $300 million ask.
A strong recommendation: "Don't pay $300 million. The current plan returns 1.6 times, well short of 2.5. I'd bid up to about $190 million, and go higher only with a credible path to EBITDA past $60 million, such as a faster clinic rollout. Real deals use debt, which lifts returns, so I'd rerun this with financing."
In an M&A case, work backward from the buyer's return target to the most it can pay.
10. A snack giant wants a protein bar brand
The protein bar deal tests whether you can value synergies and find the break-even.
The prompt: Our client, a large US snack food company, wants to buy a fast-growing protein bar brand for $1 billion. The brand has $200 million in revenue, a 5 percent operating margin, and sells in 15,000 stores. Is it worth it?
Clarifying questions:
How many stores does the client reach?
What margin could the brand earn using the client's factories and trucks?
Could the client build its own protein bar for less?
Structure: Standalone value, synergies (more stores, lower costs), price against combined value, and alternatives like building a brand. The break-even assumption matters more than the point estimate, because it tells the buyer how much room there is to be wrong.
The math:
The interviewer gives you: the client sells in 60,000 stores. In new stores, the bars would sell 60 percent as well per store as today. Margin rises to 20 percent after integration. Snack businesses are valued at about 15 times operating profit.
Standalone: $200M × 5% = $10 million profit × 15 = $150 million. The $1 billion price is mostly a bet on synergies.
Sales per current store: $200M ÷ 15,000 = about $13,333.
New stores: 45,000 × $13,333 = $600 million at the full rate; at 60 percent, $360 million. Total revenue: $560 million.
Profit: $560M × 20% = $112 million. Value: $112M × 15 = $1.68 billion. That's above the $1 billion price.
Break-even: $1B ÷ 15 = about $66.7 million of profit needed, which is $66.7M ÷ 20% = $333 million of revenue.New stores must add $133 million, or $133M ÷ $600M = about 22.2 percent of the current per-store rate.
The deal clears $1 billion as long as new stores sell at least about 22 percent of the current per-store rate.
A strong recommendation: "Buy it. The deal is worth $1 billion as long as the bars sell at least a fifth as well in our stores as in theirs, and the 60 percent estimate gives about $1.7 billion of value. Before signing, I'd test the brand in a few hundred of our stores."
When the price is far above standalone value, the case is really asking whether you believe the synergies.
Pricing case interview questions
There are three ways to set a price: from your cost, from competitors' prices, or from the value the product creates for the buyer. Strong answers show all three and say which one should drive the decision.
11. Pricing a new knee treatment
The knee treatment case tests value-based pricing and thinking about who actually pays.
The prompt: Our client, a pharmaceutical company, has a new pill that treats a knee condition that today often needs surgery. How should they price it?
Clarifying questions:
Who pays: patients, insurers, or hospitals?
Are there competing treatments besides surgery?
What does one course of the pill cost to make?
Structure: A floor (the cost to produce), a ceiling (the value to the buyer), and a point between them that gives the buyer a clear reason to say yes. The ceiling comes from what the pill saves whoever pays the bill, and here that's the insurer, not the patient.
The math:
The interviewer says surgery costs $40,000. The pill avoids surgery for 70 percent of patients. One course costs $500 to make.
Expected savings per patient: 70% × $40,000 = $28,000. That's the ceiling.
At a price of $18,000, an insurer's expected cost per patient is $18,000 for the pill plus 30% × $40,000 = $12,000 for the patients who still need surgery, $30,000 in total.
At $18,000 a course, the insurer pays $30,000 per patient instead of $40,000 for surgery alone, a $10,000 saving.
A strong recommendation: "Price between $15,000 and $20,000 a course. The pill saves insurers about $28,000 per patient in avoided surgery, and $18,000 leaves them $10,000 of that, which makes coverage an easy yes. I'd add a partial refund when a patient still needs surgery, so the value argument is harder to turn down."
Value-based pricing starts with what the product saves the buyer, then leaves the buyer a reason to say yes.
12. A streaming service wants a $2 price increase
The streaming case tests how many customers the client can lose and still gain.
The prompt: Our client is a streaming service with 10 million subscribers paying $12 a month. They want to raise the price to $14. Should they?
Clarifying questions:
What does each subscriber cost to serve each month?
What happened to cancellations the last time prices rose?
Is the goal profit, or subscriber growth ahead of a funding round?
Structure: Profit today, profit at the new price with some cancellations, and the break-even number of subscribers. Break-even has to use profit per subscriber rather than revenue, because a lost subscriber also stops costing money to serve.
The math:
The interviewer says each subscriber costs $2 a month to serve (streaming and licensing fees).
Today: each subscriber contributes $12 minus $2 = $10. 10 million × $10 = $100 million a month.
At $14: each subscriber contributes $12. Break-even: $100M ÷ $12 = about 8.33 million subscribers. The client can lose about 1.67 million, or 16.7 percent, and still earn the same.
Using revenue alone gives the wrong answer: $120M ÷ $14 = 8.57 million, a 14.3 percent limit.
If 8 percent cancel: 9.2 million × $12 = $110.4 million a month, $10.4 million more than today, or about $125 million a year.
The client can lose 16.7 percent of subscribers before profit falls, and at 8 percent cancellations it gains about $125 million a year.
A strong recommendation: "Raise the price. The client can lose about 17 percent of subscribers before profit falls, and a $2 increase is unlikely to drive losses that high. At 8 percent cancellations, profit rises about $125 million a year. I'd pair it with a cheaper ad-supported plan to catch customers who would otherwise leave."
Break-even analysis turns a pricing debate into one question: how many customers can we lose and still come out ahead?
Market sizing case interview questions
Market sizing questions grade your assumptions and your arithmetic, not whether you hit the exact answer. Say each assumption out loud, round to friendly numbers, and sanity-check the result.
13. The US dog food market
The dog food question tests top-down sizing: clean segments and multiplying large numbers.
The prompt: Estimate the annual US market for dog food in dollars.
Clarifying questions:
Meals only, or treats too?
Retail dollars (what shoppers pay) or manufacturer revenue?
Today's market, or a forecast?
Structure: Top-down means starting from a big known number and narrowing it: households, times the share with a dog, times dogs per household, times yearly spending per dog. Splitting dogs by size matters because a Great Dane eats far more than a Chihuahua, and one average hides that.
The math:
About 340 million people ÷ about 2.6 people per household = about 130.8 million (call it 130 million) households.
Assume 40 percent own a dog: 52 million households. At 1.5 dogs each: 78 million dogs.
Split dogs into thirds by size: small at $250 a year, medium at $450, large at $700. Average: ($250 + $450 + $700) ÷ 3 = $1,400 ÷ 3 = about $467.
78 million × $467: 80 × 467 = 37,360, minus 2 × 467 = 934, gives 36,426.
Americans spend about $36 billion a year on dog food, counting meals only.
A strong recommendation: "About $36 billion a year, meals only. That fits a product that fills a full aisle in almost every grocery store. At 45 percent ownership instead of 40, the market is closer to $41 billion."
Segmenting by one driver, like dog size, makes a sizing answer more accurate and easier for the interviewer to follow.
14. Revenue at a busy downtown burrito shop
The burrito shop question tests bottom-up sizing by time of day.
The prompt: How much revenue does a busy fast-casual burrito shop in a downtown office district make in a year?
Clarifying questions:
Is it open on weekends?
Does it do delivery or catering?
Is "busy" a line out the door at lunch, or steady all day?
Structure: Bottom-up means building from how one unit operates: orders per hour in each part of the day, times average ticket, weekdays and weekends separately. The lunch rush is where the money is, so estimate its throughput carefully and be rough everywhere else.
The math:
Open 10:30 a.m. to 9 p.m., 10.5 hours.
Lunch rush, 11:30 to 1:30: the line moves one order every 40 seconds, or 90 an hour. 2 hours × 90 = 180 orders.
Dinner, 5:30 to 7:30: 40 an hour × 2 = 80 orders.
The other 6.5 hours: about 15 an hour, roughly 100 orders.
Weekday total: about 360 orders × $14 average ticket = $5,040.
Weekends are slower downtown: 150 orders × $14 = $2,100 a day.
Week: (5 × $5,040) + (2 × $2,100) = $25,200 + $4,200 = $29,400. Year: $29,400 × 52 = about $1.53 million.
A busy downtown burrito shop does about $1.53 million a year, and lunch alone drives about half of weekday orders.
A strong recommendation: "About $1.5 million a year, with lunch driving about half of weekday orders. The number moves most with lunch throughput, so I'd ask how many registers are open during the rush."
For a single location, build from capacity at the busiest hours, because that's where most of the revenue sits.
15. How many wedding photographers work in the US?
The photographer question tests supply-side sizing: estimate demand, then divide by one provider's capacity.
The prompt: How many full-time wedding photographers are there in the US?
Clarifying questions:
Full-time equivalents, or anyone who shoots a wedding?
Do second shooters count?
Only weddings that hire a professional?
Structure: Weddings per year, times the share that hire a professional, divided by the weddings one full-time photographer can shoot in a year. The supply side comes last, because you can't count providers until you know how much work there is.
The math:
Weddings per year: 340 million people ÷ an 80-year life expectancy = about 4.25 million people per single-year age cohort. If about 70 percent marry, that's about 3 million people, or 1.5 million weddings, plus remarriages. Call it 2 million. TThe CDC's official count of 2.04 million marriages a year (provisional FastStats data) validates this estimate.
70 percent hire a professional photographer: 1.4 million weddings.
A full-time photographer shoots most weddings on Saturdays: about 25 peak-season weekends plus some Fridays, Sundays, and off-season dates. Call it 35 weddings a year.
Dividing 1.4 million weddings by 35 a year gives about 40,000 full-time photographers.
A strong recommendation: "About 40,000 full-time equivalents. The real headcount is higher, since many wedding photographers work part time or as second shooters."
Supply-side sizing means estimating demand first, then dividing by what one provider can handle.
Operations and cost reduction case interview questions
Operations cases are about flow: where work piles up, why, and what the cheapest fix is. Don't jump to "hire more people" before you've found the stuck step.
16. An emergency room's wait time doubles
The emergency room case tests whether you compare demand to capacity at each step.
The prompt: A city hospital's emergency room wait time doubled in a year, from 45 minutes to 90 minutes on average. How would you bring it back down?
Clarifying questions:
Did patient numbers change, or staffing?
Is the wait for a doctor, for test results, or for a bed after admission?
Are there budget limits on new hires?
Structure: Arrivals (demand), capacity at each step (triage, doctor, tests, admission), and the mix of patients by how sick they are. Waits explode when demand reaches capacity, because no slack is left to absorb a single slow case.
The math:
The interviewer says peak arrivals rose from 8 to 10 patients an hour. Four doctors each see 2.5 patients an hour, so capacity is 4 × 2.5 = 10 an hour.
Utilization, how busy the doctors are: 8 ÷ 10 = 80 percent last year, 10 ÷ 10 = 100 percent now.
Option A: add a fifth doctor at $400,000 a year. Capacity becomes 12.5 an hour; utilization 10 ÷ 12.5 = 80 percent.
Option B: a fast track. 40 percent of arrivals, 4 an hour, are minor cases. A nurse practitioner at $150,000 a year sees 4 minor cases an hour. Doctors now handle 6 an hour against capacity of 10, so utilization drops to 60 percent.
A fast track costs $150,000 a year and cuts doctor utilization to 60 percent, while a fifth doctor costs $400,000 and only gets to 80.
A strong recommendation: "Add a nurse practitioner fast track during peak hours. It costs about $150,000 a year, less than half a doctor, and drops doctor utilization to 60 percent instead of 80. I'd also confirm admitted patients aren't stuck in the ER waiting for beds, which needs a different fix."
In an operations case, compare demand to capacity at each step before suggesting a fix.
17. An online retailer's shipping costs jump
The shipping cost case tests whether you split a blended average into its parts.
The prompt: An online home goods retailer's shipping cost per order rose from $8 to $11 in two years. It ships 10 million orders a year. Where's the money going, and what should it do?
Clarifying questions:
Did carrier rates change?
Did the mix of shipping speeds, package sizes, or distances change?
Did the company change any shipping promises to customers?
Structure: Cost per order equals the share of orders at each shipping speed times the cost of that speed, plus distance and package size. A blended average can rise with no price change at all if the mix underneath it shifts.
The math:
Total impact: $3 × 10 million = $30 million a year.
The interviewer says rates didn't change. Two-day shipping costs $16 an order and ground costs $6. Two-day went from 20 percent of orders to 50 percent after the company launched free two-day shipping on orders over $50.
Before: (20% × $16) + (80% × $6) = $3.20 + $4.80 = $8.00..
After: (50% × $16) + (50% × $6) = $8.00 + $3.00 = $11.00.
If two-day falls back to 30 percent: (30% × $16) + (70% × $6) = $4.80 + $4.20 = $9.00, saving $2 an order.
The whole $3 increase is the shift to two-day shipping, and getting back to 30 percent two-day saves $20 million a year.
A strong recommendation: "All $30 million of the increase comes from more orders going two-day. I'd raise the free two-day threshold to $100 and show a delivery date on ground orders, since many customers only want to know when it arrives. Getting back to 30 percent two-day saves about $20 million a year, as long as conversion holds."
A blended average that rises usually means the mix changed, so split it before blaming prices.
18. A factory can't keep up with demand
The factory case tests whether you find the bottleneck before spending on capacity.
The prompt: Our client's factory makes 1,000 kitchen appliances a day, but demand is now 1,300 a day. Management wants to build a second factory for $50 million. Is that the right call?
Clarifying questions:
What are the production steps, and what can each handle per day?
How many shifts run now?
Will the higher demand last?
Structure: Capacity at each step, the bottleneck, the cheapest way to lift it, and the next bottleneck after that. A second factory adds capacity at every step, but the client only needs it at one.
The math:
The interviewer says: stamping can do 1,500 a day, assembly 1,000, painting 1,400. Assembly is the bottleneck.
Assembly runs two 8-hour shifts, 16 hours, so it makes 1,000 ÷ 16 = 62.5 units an hour.
To add 300 units: 300 ÷ 62.5 = 4.8 hours. A five-hour third shift covers it.
Painting at 1,400 can handle 1,300, so it isn't the new limit.
Cost of the extra shift: 40 workers × $35 an hour × 5 hours = $7,000 a day, or $1.75 million over 250 working days
A five-hour third shift on assembly costs about $1.75 million a year, against $50 million for a second factory.
A strong recommendation: "Don't build a new factory yet. Assembly is the only step below 1,300, and a five-hour third shift fixes it for about $1.75 million a year. If demand passes 1,400 a day, painting becomes the limit, and that's when to revisit a second plant."
Find the bottleneck step first, because the whole line can only run as fast as its slowest step.
New product case interview questions
New product cases test whether you check what the new product takes away from products the client already sells. That effect is called cannibalization.
19. A kids' sneaker subscription
The sneaker subscription case tests subscription unit economics and cannibalization.
The prompt: Our client, a large athletic shoe company, wants to launch a subscription for kids' sneakers. Parents pay a yearly fee and get a new pair every time the child outgrows the old one. Should they launch it?
Clarifying questions:
How often do kids outgrow shoes?
Do subscribing families already buy the client's shoes in stores?
What happens to the returned shoes?
Structure: Customer value, unit economics (profit per subscriber per year and over their lifetime), cannibalization, and risks. Cannibalization here means asking whether subscribers would have bought the client's shoes in a store anyway.
The math:
The interviewer says kids outgrow shoes about every six months, so 2 pairs a year. A pair costs $70 in stores. The subscription is $120 a year. Each pair costs $30 to make, $10 to ship, and $5 for a prepaid return label, so $45 a pair. Winning a subscriber costs $60 in marketing. Families stay three years on average.
Parents: $120 versus 2 × $70 = $140 in stores. They save $20.
Client profit per year: $120 minus (2 × $45) = $120 minus $90 = $30.
Over three years: 3 × $30 = $90, minus $60 to acquire = $30 per subscriber.
Cannibalization: the client earns about $15 a pair selling through stores, or $30 a year for a family that already buys its shoes.
For a family that already buys the client's shoes, the subscription earns the same $30 a year and adds shipping risk.
A strong recommendation: "Launch a pilot, not a national rollout. For families who already buy our shoes, the subscription earns about what stores earn, so it only pays with new families. I'd aim the pilot at families who buy other brands and test reselling returned shoes to lift the $30 margin."
A new product only creates value if its profit beats what it takes from products you already sell.
20. A grocery chain's store-brand organic line
The store-brand case tests whether cannibalization helps or hurts when margins differ.
The prompt: A regional grocery chain wants to launch its own store-brand line of organic packaged foods. Should it?
Clarifying questions:
How much does the chain earn on the national brands today?
What will it cost to design, source, and package the line?
Will national brands pull promotional money in response?
Structure: Profit per unit for the national brand against the store brand, how many shoppers switch or join, launch cost, and payback. Cannibalization can help here, because the sale the store brand replaces may earn less than the one it creates.
The math:
The interviewer uses organic pasta: the national brand sells for $4 and the chain earns $1 (25 percent). The store brand sells for $3 and earns $1.20 (40 percent). The interviewer expects 30 percent of buyers to switch.
Per 100 units before: 100 × $1 = $100 profit.
After: (70 × $1) + (30 × $1.20) = $70 + $36 = $106. That's 6 percent more.
If the lower price also brings 10 new units per 100: 10 × $1.20 = $12 more, so $118.
The category sells 10 million units a year, earning $10 million. A 6 percent lift is $0.6 million a year; with new buyers, 18 percent is $1.8 million.
Against $2 million to launch, payback runs from about 1.1 years with new buyers to 3.3 years on switching alone.
A strong recommendation: "Launch it. Every shopper who switches earns the chain 20 cents more per unit, the lower price may pull in new buyers, and payback is one to three years on a $2 million launch. The risk is national brands cutting promotional money, so I'd start in three categories and watch."
Cannibalization is fine when the product that replaces the old sale earns more per unit.
Public sector and nonprofit case interview questions
Public sector and nonprofit cases have goals beyond profit, so name those goals before you touch the math.
21. A transit agency's budget gap
The transit case tests whether you can balance conflicting goals against a fare increase.
The prompt: A big-city public transit agency faces a $200 million budget gap next year, and ridership is still below where it was before the pandemic. What should it do?
Clarifying questions:
Must the budget balance by law?
What else does the board care about: ridership, service levels, access for low-income riders?
Is new state or city funding possible?
Structure: Fare revenue (price per ride times rides), costs (operations, overtime, underused routes), outside funding, and the effect of each move on ridership. Every lever in a transit case has a ridership cost attached, so size that cost alongside the dollars.
The math:
The interviewer says: 300 million rides a year at an average fare of $1.50, or $450 million. Price elasticity is 0.3. Elasticity measures how much demand changes when price changes: here, a 10 percent fare increase means about 3 percent fewer rides.
A 25 cent increase: $0.25 ÷ $1.50 = about 16.7 percent. Rides fall 16.7% × 0.3 = 5 percent, to 285 million.
New fare revenue: 285 million × $1.75 = $498.75 million. Gain: about $49 million, a quarter of the gap.
Fare evasion: if 5 percent of rides go unpaid, that's 15 million × $1.50 = $22.5 million.
A 25 cent fare increase raises about $49 million and fare enforcement up to $22.5 million, together roughly a third of the $200 million gap.
A strong recommendation: "No single lever closes $200 million. I'd combine about $20 million from fare enforcement, $49 million from a 25 cent fare increase, and $60 million from cutting the least-used routes, and ask the city and state for the remaining $70 million. I'd protect frequency on the busiest lines, because cutting there loses riders."
Public sector cases usually need several smaller moves, because the single big lever almost always hurts one of the agency's other goals.
22. A food bank serves more families on the same budget
The food bank case tests whether you turn a goal into a cost-per-unit target.
The prompt: A food bank network wants to serve 20 percent more families next year without increasing its $10 million budget. How can it do that?
Clarifying questions:
How is the budget split between food, logistics, and staff?
How much food is donated today?
Can volunteers take on any paid roles?
Structure: Cost per family, split into food purchases, trucks and warehouses, and staff and overhead, then savings in each that add up to the target. A fixed budget and a bigger goal mean the real question is how much cheaper each family has to get.
The math:
Today: $10 million ÷ 50,000 families = $200 per family.
Goal: 60,000 families, so $10 million ÷ 60,000 = about $166.67 per family. The network must cut about $33 per family, or one sixth.
The interviewer splits the $200: $100 food purchases, $60 logistics, $40 staff and overhead.
More donated food from grocery partners cuts purchases 20 percent: $20. Combining delivery routes cuts logistics 10 percent: $6. Volunteers in some roles cut overhead 10 percent: $4. Total: $30, so cost per family becomes $170.
$10 million ÷ $170 = about 58,824 (roughly 58,800) families, 17.6 percent more
Cutting cost per family to $170 serves about 58,800 families, which leaves the network about 1,200 short of the goal.
A strong recommendation: "Three changes get the network to about 58,800 families: more donated food, shared delivery routes, and volunteers in some roles. To close the last 1,200, negotiate bulk purchasing with other food banks: a 4 percent discount on the remaining $80 of food per family saves another $3.20, within a few dozen families of the goal."
When the budget is fixed, turn the goal into a cost-per-unit target and find savings that add up to it.
Unconventional case interview questions
An unusual prompt is still a structure question, so break it into quantities, time, and cost like any other case.
23. What to charge for washing a skyscraper's windows
The window washing quote tests whether you stay structured with no familiar template.
The prompt: A building owner asks what you'd charge to wash the outside windows of a 40-story office tower once. What price would you quote?
Clarifying questions:
How big is each floor?
Is there a platform on the roof, or do we bring equipment?
One-time job or recurring contract?
Structure: Count the windows, estimate the labor time, cost the labor and equipment, then add a margin. A price quote is an estimate with a markup on top, so this is a sizing case plus one step.
The math:
A floor measures 150 feet by 100 feet. Perimeter: 2 × (150 + 100) = 500 feet.
Windows are 5 feet wide: 500 ÷ 5 = 100 windows per floor. Times 40 floors: 4,000 windows.
A two-person crew on a platform cleans one window in about 3 minutes, so 20 an hour. 4,000 ÷ 20 = 200 crew-hours. Add 25 percent for moving the platform and setup: 250 crew-hours.
A crew costs $150 an hour including wages, insurance, and equipment. 250 × $150 = $37,500.
A 30 percent markup on $37,500 of cost ($37,500 × 1.3) gives a quote of $48,750 for 250 crew-hours of work.
A strong recommendation: "I'd quote about $49,000, roughly $12 a window. If the owner wants two washes a year, I'd offer a contract at a small discount, since the second job reuses the setup."
End an odd estimate with a price and the reason behind it, because a bare number doesn't answer a business question.
Chart and exhibit case interview questions
Exhibit questions test whether you can find the one thing that matters in a table or chart in under a minute. Read the title and units first, then look for the biggest change.
24. Revenue by region at an electronics retailer
The revenue exhibit tests finding the "so what" and doing percentages fast.
The prompt: Here's revenue by region for a consumer electronics retailer. What stands out?
The exhibit: revenue in millions of dollars, 2023 to 2025.
Northeast: $400 to $420.
South: $300 to $390.
Midwest: $250 to $240.
West: $350 to $300.
Total: $1,300 to $1,350.
Clarifying questions:
Did the store count in any region change?
Is this all channels, or stores only?
Are the figures adjusted for inflation?
Structure: Total change first, then each region's change in dollars and percent, then the question the data raises. The total hides the story whenever two regions move in opposite directions.
The math:
Total: $50 million more. $50 ÷ $1,300 = about 3.8 percent.
Northeast: $20 ÷ $400 = 5 percent up.
South: $90 ÷ $300 = 30 percent up.
Midwest: $10 ÷ $250 = 4 percent down.
West: $50 ÷ $350 = about 14 percent down.
The South's $90 million gain covers the West's $50 million drop, with $40 million left over.
Total revenue grew about 3.8 percent, but the South rose 30 percent while the West fell 14 percent.
A strong recommendation: "Total revenue grew about 4 percent, but that hides two stories: the South grew 30 percent and the West fell 14 percent. I'd want to know why the West dropped: store closures, a new competitor, or fewer visits per store."
Say the one-sentence takeaway from an exhibit first, then the numbers that back it up.
25. Which plant should close?
The plant closure exhibit tests whether you know average cost from marginal cost.
The prompt: Here's cost and volume data for our client's three plants. Which, if any, should it close?
The exhibit: units, average cost, and capacity by plant.
Plant A: makes 2.0 million units a year at an average cost of $5.00. Capacity: 2.6 million.
Plant B: 1.5 million units at $5.40. Capacity: 1.6 million.
Plant C: 0.5 million units at $7.20. Capacity: 1.0 million.
Clarifying questions:
What does it cost Plant A to make one more unit?
What does closing Plant C cost one time?
Do Plant C's fixed costs, like rent and managers, disappear if it closes?
Structure: Can the other plants absorb Plant C's volume, what's the saving per unit moved, and how fast does that repay the one-time closing cost? Average cost is total cost divided by units, marginal cost is what one more unit costs, and only marginal cost belongs in this decision.
The math:
Plant A has 2.6M minus 2.0M = 0.6 million units of spare capacity, enough to take Plant C's 0.5 million.
The interviewer says each extra unit at Plant A costs $4.50. That's below its $5.00 average, because A's fixed costs are already paid.
The interviewer confirms C's fixed costs go away with the closure, so moving a unit saves C's full $7.20 average cost. Saving per unit: $7.20 minus $4.50 = $2.70. Total: 0.5 million × $2.70 = $1.35 million a year.
Closing C costs $4 million one time. Payback: $4M ÷ $1.35M = about 3 years.
The trap: using A's $5.00 average cost gives a saving of only $1.1 million a year, which understates the case.
Closing Plant C saves $1.35 million a year and repays the $4 million closing cost in about three years.
A strong recommendation: "Close Plant C and move its volume to Plant A. A has the room, each moved unit saves $2.70, and the $4 million closing cost pays back in about three years. Before deciding, I'd check shipping costs from A to C's customers and whether A at 96 percent capacity can absorb demand spikes."
Exhibit questions often hide a trap, so check whether the decision needs the average cost shown or the extra cost of one more unit.
How to practice these case interview questions
Practicing out loud with a partner is the closest thing to a real case interview, and nothing else builds the same skill. Reading solved cases teaches you structures. Only speaking them builds the calm you need in the room.
One person plays interviewer and holds the data, sharing numbers only when asked. Run the case for about 25 minutes, then spend 10 on feedback: structure, math, the recommendation, and how clearly you spoke. Swap roles next time.
Practice both interviewer-led and candidate-led styles, because McKinsey and the other firms run cases differently. When I went through McKinsey's process, the interviewer led step by step: a structure, then an exhibit, then a specific calculation. BCG and Bain cases give the candidate more room to choose where to go next. The McKinsey case interview guide walks through that format.
Between partner sessions, drill math with no calculator (percentages, dividing by 12, 1.08 to the fifth power) and record yourself once a week to catch filler words and long silences. Plan about two minutes for restating and clarifying, a minute or two for structure, 15 minutes of analysis, and one to two minutes for the recommendation.
Spread practice over weeks rather than cramming, because the skill you're building is staying calm, and calm comes from repetition over time. For a full prep plan, see how to prepare for a consulting interview and these 10 tips to beat the case interview. And don't skip the behavioral half: at McKinsey, the Personal Experience Interview is graded as seriously as the case.
What to say in the first two minutes of a case: say this, don't say that
The first two minutes of a case decide how the interviewer reads the next 20, so script your opening moves. Using case 1:
Restating the prompt
Don't say: "Okay, so profits are down. Got it."
Say: "Let me make sure I have it: 120 gyms, revenue flat at $180 million, operating profit down from $27 million to $18 million over three years. We want to find out why and fix it. Is that right?"
Asking clarifying questions
Don't say: "Can you tell me more about the company?"
Say: "Two questions before I structure: has the number of gyms changed, and does the client have a profit target in mind?"
Asking for time
Don't say: "Um, so I guess first I'd look at, like, revenue..."
Say: "Could I take a minute to organize my thoughts?"
Presenting the structure
Don't say: "I'll use the profitability framework."
Say: "I'd look at this in two parts. Revenue, split into members and dues. Costs, split into ones that change with membership and ones that don't. Since revenue is flat, I'd start by checking whether members and dues moved in opposite directions."
Starting the math
Don't say: nothing, while you compute in silence for a full minute.
Say: "I'll multiply 300,000 members by $50 by 12 months. That's $15 million a month, so $180 million a year."
Naming a framework tells the interviewer nothing, but naming the first thing you'd check tells them how you think.
What about the firms' own practice cases not on this list?
The firms publish free practice cases, and they're the best way to learn each firm's style, so use them alongside the 25 above.
McKinsey posts Electro-Light, Talbot Trucks, Parks Inc., and Conservation International on its candidate preparation portal, alongside interactive guidance on the Personal Experience Interview.
BCG has a case interview preparation page with dos and don'ts and a link to OneDay@BCG, a free job simulation.
Bain posts Coffee Shop Co. and FashionCo. on its case interview page.
Case in Point by Marc Cosentino is the standard starting book, and Case Interview Secrets by Victor Cheng is strong on building your own structure. One book is enough, and after that every hour is better spent on live cases with a partner. The PrepLounge case library is useful for finding those partners, and 10 free resources for case consulting prep ranks what's free and how to use each one.
Where to start this week
Pick the three case types you'd least like to get in a first round and run one of each this week with a partner. For most students, that's market sizing, an exhibit, and one M&A case. Keep a list of every mistake and practice the math behind each one until it's quick.
Case interviews test a skill you can build, and the students who get offers are the ones who practiced it out loud until the first 30 seconds stopped being scary. If MBB is your target, the guide to breaking into MBB covers the rest of the path.
Comments