Crack any case in five moves
The one method this site teaches: build the structure from the goal and the business, not from a memorized list. Two full examples: a hospital in India and a food delivery launch in Riyadh.
Key takeaways
- We do not teach frameworks to memorize. The structure comes from the maths of the goal in this exact question, and what you know about the business tells you where to look first.
- You will still hear the names, so know what they mean. If an interviewer says "think about the customers, the competitors and the company", you know the idea: three places to look.
- Both start from profit, but only the stronger answer knows what this business sells and where its profit moves.
Here is the whole method on one page. Every case on this site, in any industry, is cracked with the same five moves. You do not need a list of frameworks to memorize. You need to understand the business in front of you and turn the question into numbers. This lesson teaches the five moves, then runs two very different cases through them: a hospital in India and a food delivery app in Saudi Arabia.
Key idea
We do not teach frameworks to memorize. The structure comes from the maths of the goal in this exact question, and what you know about the business tells you where to look first.
Crack any case in five moves
- 1Pin the question. What decision does the client face, what does success look like, and by when? Say it back in one sentence. Ask the one or two clarifying questions that would change the answer.
- 2Write the maths of the goal. Every business question is a number moving. Profit is revenue minus cost. Revenue is customers x purchases x price. A market is buyers x units x price. Break the goal into its drivers until each branch is something you can measure. That is your structure, and it fits because it comes from this question, not from a list.
- 3Use what you know about the business. How does this business make money? Where are its biggest costs? Which measures does it watch? Which business model pattern does it follow? This tells you which branches matter here. Pick the branch where the answer probably sits and say why: that is your hypothesis.
- 4Find the facts that decide it. Ask for, or estimate, the numbers that test your hypothesis. Do the maths cleanly, find the one number in each chart that matters, and change your hypothesis when the data disagrees.
- 5Say so what. Answer first: the decision, two or three reasons with numbers, the main risk, and the next step.
Later in this module, "The case, start to finish" walks through the moments of a live case: summarizing, asking for time, presenting your structure, the mid-case summaries and the close. The five moves are the thinking inside those moments. Every case type, industry and worked case on this site uses them.
Example 1: a hospital in India whose profit fell
Interviewer: "Our client runs a 250-bed private hospital in Pune, India. Two years ago it made an operating profit of INR 50 crore. This year it made INR 20 crore. Why, and what should it do?" (One crore is 10 million rupees. The hospital is fictional and the figures are rounded and illustrative.)
Move 1. Pin the question
- You: "So the hospital's operating profit fell from INR 50 crore to INR 20 crore in two years, and the client wants the cause and a plan. Is the goal to get back to INR 50 crore, and by when?" Interviewer: "Yes, within two years."
- You: "Has the number of beds changed, or has the hospital opened or closed a service?" Interviewer: "No. Still 250 beds and the same services." This one question matters: if beds had changed, the whole analysis would start somewhere else.
Move 2. Write the maths of the goal. A hospital sells bed days: one patient in one bed for one day. So profit = revenue minus cost. Revenue = beds x 360 days x occupancy (the share of beds that are filled) x revenue per occupied bed per day, which Indian hospital chains call ARPOB. Cost = fixed costs that do not change with the number of patients (staff, the building, equipment) plus variable costs for each bed day (drugs, implants and other supplies). We use 360 days a year to keep the maths simple.
- Operating profit: INR 50 crore to INR 20 crore
- Revenue = beds x days x occupancy x ARPOB
- Beds: 250, unchanged
- Key: Occupancy: how full the beds are
- Key: ARPOB: what each filled bed earns a day
- Payer mix: who pays, and at what rate
- Case mix and length of stay
- Cost
- Fixed: staff, building, equipment
- Variable for each bed day: drugs, implants, supplies
This comes from the question: profit is the number that moved, and a hospital earns it one bed day at a time.
Move 3. Use what you know about the business. A hospital is a utilization business: most of its cost is fixed for the day, whether a bed is full or empty. So profit moves mostly with how full the beds are and with how much each filled bed earns. In India, who pays matters a lot: patients who pay themselves or through private insurance pay full rates, while government insurance schemes pay lower, fixed package rates for each treatment. Hypothesis: occupancy fell, because emptier beds are the most common reason a hospital's profit drops. If occupancy is flat, I will test revenue per bed day and the payer mix next.
The industry brief "Healthcare providers and payers" explains how hospitals earn money, why payer mix matters, and the measures to ask for first, such as occupancy and ARPOB. Hospitals follow the utilization pattern (on the Business model patterns page), the same idea as airlines, hotels and telecom, so once you know it, those cases feel familiar too.
Read the healthcare brief| Measure | Two years ago | This year |
|---|---|---|
| Beds | 250 | 250 |
| Occupancy | 80% | 80% |
| Share of bed days paid by a government scheme | 20% | 35% |
| Revenue per bed day, private and self-paying patients (INR) | 55,000 | 55,000 |
| Revenue per bed day, government scheme patients (INR) | 30,000 | 30,000 |
| Fixed costs (INR crore) | 220 | 223 |
| Variable costs (INR crore) | 90 | 90 |
| Operating profit (INR crore) | 50 | 20 |
So-what
Occupancy did not move. What changed is who fills the beds: more scheme patients, who pay much less per day.
Move 4. Find the facts that decide it
- 1Test the hypothesis first. Occupancy is 80 percent in both years, so the hypothesis is wrong. Say so calmly: "Occupancy is flat, so the beds are as full as before. The fall must be in what each bed earns, or in cost."
- 2Cost: fixed costs rose by only INR 3 crore (staff pay) and variable costs did not change. Cost explains INR 3 crore of the INR 30 crore fall.
- 3Revenue per bed day: two years ago, 80 percent of bed days paid INR 55,000 and 20 percent paid INR 30,000, an average of INR 50,000. This year, 65 percent pay INR 55,000 and 35 percent pay INR 30,000, an average of INR 46,250. That is 7.5 percent less for every filled bed.
- 4Revenue: 250 beds x 360 days x 80 percent is 72,000 bed days in both years. At INR 50,000 that was INR 360 crore; at INR 46,250 it is INR 333 crore. Revenue fell by INR 27 crore, which is 90 percent of the INR 30 crore fall in profit.
Waterfall chart: From INR 50 crore to INR 20 crore: where the profit went (INR crore). Values in INR crore. Profit two years ago, total: 50; Lower revenue per bed day (payer mix), change: -27; Higher staff pay, change: -3; Profit this year, total: 20.
So-what
Nine tenths of the fall comes from the payer mix, not from emptier beds or higher costs.
This year 65 percent of bed days pay INR 55,000 and 35 percent pay INR 30,000. What is the average revenue per occupied bed day, in INR?
The hospital has 250 beds, 360 days a year, 80 percent occupancy and INR 46,250 per occupied bed day. What is its yearly revenue, in INR crore? (One crore is 10 million.)
"The profit fall is a payer mix problem, not a demand problem. The beds are as full as before, but government scheme patients now fill 35 percent of bed days instead of 20 percent, and they pay about INR 30,000 a day against INR 55,000. That cut revenue by INR 27 crore, 90 percent of the fall; higher staff pay explains the other INR 3 crore. I recommend three things: grow private and insured patients in the specialties where the hospital is strongest; shorten the stay of scheme patients where it is safe, because the scheme pays a fixed package per treatment, so each day saved frees a bed; and find which scheme treatments lose money and raise them with the scheme. The main risk is trust: the hospital must keep its scheme commitments. Next, I would size how many full-rate patients the strongest specialties could add."
Example 2: should a food delivery app launch in Riyadh?
Interviewer: "Our client is a food delivery app that makes a profit in the UAE. It wants to launch in Riyadh, Saudi Arabia. Should it?" (A fictional company. The figures are rounded and illustrative, in Saudi riyals, SAR.)
Move 1. Pin the question
- You: "So the decision is whether to launch in Riyadh. What would success look like, and by when?" Interviewer: "The board wants the city to cover its own costs within two years."
- You: "Would we launch alone, or with a partner who already has restaurants and riders?" Interviewer: "Alone, from scratch." This changes the answer: alone, every customer, restaurant and rider has to be won.
Move 2. Write the maths of the goal. The goal is a city that covers its costs, so: city profit = orders x contribution per order, minus fixed city costs (the local team, offices and brand marketing). Contribution per order = what the app earns on one order (commission from the restaurant, the delivery fee, restaurant ads) minus what that order costs (rider pay, discounts, payment fees, support and refunds). Orders = all delivery orders in the city x our share.
- Can Riyadh cover its own costs within two years?
- Orders a day = city orders x our share
- Delivery orders in the city, all apps
- Share we can win from apps already there
- Key: Contribution per order
- Earned: commission, delivery fee, ads
- Key: Paid out: rider, discounts, payment, support
- Fixed city costs: team, offices, brand marketing
This comes from the question: "should we launch" means "will the city cover its costs", and a delivery app earns its money one order at a time.
Move 3. Use what you know about the business. A delivery app is a marketplace: it matches customers, restaurants and riders and keeps a cut of each order (its take rate). Three facts about this kind of business tell you where to look. The rider, the last mile, is usually the biggest cost of each order. Many customers keep two or three apps and switch for a discount, so a newcomer pays for its first orders with discounts. And the network is built city by city: being big in Dubai helps little in Riyadh. Hypothesis: the launch lives or dies on contribution per order, because a newcomer pays more in discounts and has fewer orders per rider trip than the apps already there. So I will test the economics of one order first, then the share needed to cover fixed costs.
The industry brief "Internet platforms, marketplaces and digital ads" works through the yearly economics of one food delivery customer, and the "Restaurants and food service" brief shows the restaurant's side of the commission. The marketplace pattern (on the Business model patterns page) gives you the questions to ask in any business that matches two sides.
Read the internet platforms brief| Item | Value |
|---|---|
| Average order value (SAR) | 80 |
| Commission from the restaurant, 20 percent (SAR) | 16 |
| Delivery fee paid by the customer (SAR) | 6 |
| Restaurant ads, per order (SAR) | 2 |
| Rider pay (SAR) | 14 |
| Discounts to win customers (SAR) | 6 |
| Payment fees, 2 percent of the order (SAR) | 1.6 |
| Support and refunds (SAR) | 0.4 |
| Fixed city costs (SAR a month) | 3 million |
| Delivery orders in the city, all apps (a day) | 500,000 |
So-what
The app earns SAR 24 on an order and pays out SAR 22, so each order leaves only SAR 2 to cover the fixed costs.
Move 4. Find the facts that decide it
- 1Earned on one order: SAR 16 commission plus SAR 6 delivery fee plus SAR 2 ads is SAR 24.
- 2Paid out on one order: SAR 14 rider plus SAR 6 discounts plus SAR 1.6 payment plus SAR 0.4 support is SAR 22. Contribution is SAR 2 an order. The hypothesis holds: each order barely pays for itself.
- 3Orders needed: SAR 3 million a month of fixed costs divided by SAR 2 is 1.5 million orders a month, or 50,000 a day. That is 10 percent of the 500,000 orders the city places each day, all won from apps that are already there.
- 4Test the risk: if rivals answer with bigger discounts and ours rise to SAR 10 an order, the order costs SAR 26 and earns SAR 24. Every extra order then loses SAR 2.
- 5Find what would change it: in dense districts a rider can carry two orders a trip, so rider pay falls to SAR 11 an order. Contribution rises to SAR 5, and the orders needed fall to 20,000 a day, a 4 percent share.
Fixed city costs are SAR 3 million a month (use 30 days). Each order leaves SAR 2 after its own costs. How many orders a day does the city need to cover its fixed costs?
With two orders per rider trip, rider pay falls from SAR 14 to SAR 11 an order. The app still earns SAR 24 an order and pays SAR 6 discounts, SAR 1.6 payment and SAR 0.4 support. What is the new contribution per order, in SAR?
"Do not launch across the whole of Riyadh now. At launch each order leaves only SAR 2, so the city needs 50,000 orders a day, a 10 percent share won from apps already there, just to cover its costs, and one round of bigger discounts from rivals makes every order lose money. A focused launch can work: start in a few dense districts where riders carry two orders a trip, which lifts contribution to SAR 5 and cuts the share needed to about 4 percent. The main risk is a discount war. Next, I would run a three-month test in two districts to measure repeat orders and orders per rider trip before deciding on the whole city."
Same hospital prompt, two first answers
Weaker answer
"I will use the profitability framework. Revenue is price times volume, and costs are fixed and variable. Could I have all the data?" It is true for every company, so it shows nothing about this one.
Stronger answer
"A hospital sells bed days, and most of its cost is fixed for the day. So profit moves with how full the beds are, what each filled bed earns, and who pays. I suspect occupancy fell, so I will check it first, then revenue per bed day and the payer mix."
Why the stronger answer wins: Both start from profit, but only the stronger answer knows what this business sells and where its profit moves. That is the difference between a structure taken from a list and one built from the goal and the business.
Why we do not teach frameworks
Many guides teach a set of named frameworks to memorize: the profitability tree, the three Cs, the four Ps, Porter's five forces, SWOT and others. We do not, for three reasons.
- They sound generic. A memorized list can be pasted onto any prompt, so it shows the interviewer nothing about how you think, and interviewers hear the same lists again and again.
- They break on unusual cases. A temple that wants more visitors, a city deciding on a second airport, or a charity choosing between two programmes fits no list. The maths of the goal always exists: visitors, passengers and people helped are numbers you can break into drivers.
- They skip the business. "Revenue and cost" is true for every company, but it does not tell you that a hospital's cost is fixed for the day, or that a delivery app lives or dies on the rider cost of each order. Knowing the business is what finds the answer fast.
BCG's case interview preparation page says there is not always a single right answer, and that what matters most is how you approach the problem and the quality of your reasoning (careers.bcg.com, checked on 2026-10-01). That is a test of thinking, not of recall.
You will still hear the names, so know what they mean. If an interviewer says "think about the customers, the competitors and the company", you know the idea: three places to look. Those reminders fall out of moves 2 and 3 on their own, as questions about this business, not as the structure itself. The frameworks section explains each name, the idea worth keeping, and what to do instead.
Frameworks, and why we do not teach themCheck yourself
Which of the five moves turns the question into a structure?
In the hospital case, why look at occupancy and revenue per bed day before cost?
Your hypothesis was that occupancy fell. The data shows occupancy flat at 80 percent. What do you say?
An interviewer asks: "What about the customers, the competitors and the company?" What is the best response?
A cinema chain in Singapore wants to raise its profit by 20 percent in two years. Write the maths of the goal as a tree, say which branch you would test first and why, then reveal a model answer.
Sources for this lesson (2)
- Recognized public explanations of case-interview concepts and terms
- BCG Careers, case interview preparation (checked 2026-10-01)
My notes on this lesson
0 of 5,000 characters. Saves automatically.
Try the 9 remaining checks and drills above to complete this lesson (0 of 9 done).
Spotted something wrong or out of date? Report a mistake. We check every report and correct the page.