Profitability is the most common case type, and the profit tree is the first framework most candidates learn. That is exactly the problem with it. Interviewers have heard “revenue and costs, price and volume, fixed and variable” thousands of times, and a candidate who recites it has told them nothing about how they think.
The tree is still the right backbone. What separates a strong answer is what you do with it: which branches you name for this particular client, the question you ask first, the hypothesis you commit to, and whether you notice that the problem is not the client’s at all. This post covers the tree, the five questions that turn it into a diagnosis, one profitability case worked end to end with the math, and the habits that mark a framework as recited.
The profit tree
Everything in a profitability case hangs off one identity. Profit is revenue minus costs. Revenue is price times volume, summed over every product, customer segment and channel. Costs are fixed costs, which do not move with volume (rent, salaried staff, equipment), plus variable costs, which do: cost per unit times volume (ingredients, materials, shipping, commissions).
The tree alone is not an answer. McKinsey’s interviewing page says it plainly: interviewers want candidates to listen carefully and give tailored solutions rather than standard frameworks. So the tree is where you start thinking, and the version you say out loud should already have this client’s names on it: “revenue from in-store sales and from wholesale to cafes” rather than “revenue.”
Five questions that turn the tree into a diagnosis
The tree tells you where a problem could be. The questions tell you where it is. Ask them in this order, because each one cuts the tree roughly in half, and the order keeps you from spending ten minutes on cost lines when the problem was price.
- 1Revenue or costs?Ask for both over the last few years. Which one moved?
- 2The client or the market?If competitors are hurting too, the fix is different.
- 3Which product, segment or channel?Totals hide the one line that fell.
- 4If revenue: price or volume?Discounting and lost customers need different fixes.
- 5If costs: which line, fixed or variable?A variable cost per unit rising is a supplier or process problem.
Revenue or costs? Ask for both over the last two or three years. Most cases are written so that one of them moved and the other did not.
The client or the market? This is the question that most often separates strong candidates. If the whole industry’s margins fell because an input got more expensive, the client’s problem is not a bad manager, and the fix is about pricing power and contracts. If competitors are fine, something specific to the client broke.
Which product, segment or channel? Totals hide the line that fell. A flat revenue figure can be one product growing and another collapsing.
Price or volume? Revenue falling because of discounting needs a different fix than revenue falling because customers left.
Which cost line, and fixed or variable? A fixed cost that grew is often a one-time decision (a new lease, a new hire). A variable cost per unit that grew is a supplier, input-price or process problem, and it gets worse as volume grows.
A worked profitability case
Here is a case written for this post, with round numbers, so you can follow every step.
The prompt. Your client is a regional bakery chain with 20 stores. Profit has fallen from $4.0 million to $2.5 million over two years. The owner wants to know why, and what to do about it.
A tailored structure. Rather than “revenue and costs,” say: “I’d look at revenue by store and by product, since breads, pastries and catering likely have different margins. On costs, ingredients are the big variable cost for a bakery, then labor, rent and everything else. And I want to know whether other bakeries in the region saw the same thing, because flour and butter prices move for everyone.” That last sentence is question two, said before you have any data.
The data, as the interviewer would give it.
| Two years ago | This year | |
|---|---|---|
| Revenue | $40.0M | $40.0M |
| Ingredients | $12.0M (30% of revenue) | $13.2M (33% of revenue) |
| Labor | $14.0M | $14.0M |
| Rent | $6.0M | $6.3M |
| Other | $4.0M | $4.0M |
| Profit | $4.0M | $2.5M |
The interviewer adds, when you ask: prices and the number of items sold are both flat, and the two biggest competitors raised prices about 5 percent last year.
Reading it. Revenue is flat, so the whole $1.5 million drop is in costs. Ingredients rose $1.2 million and rent $0.3 million; labor and other costs did not move. With prices and volume flat, the ingredient increase is a rise in cost per item of about 10 percent ($13.2 million divided by $12.0 million). That is the market, not the client: flour and butter got more expensive for everyone. The useful insight is the second fact. Competitors passed the increase on to customers, and the client did not.
The math on the fix. Test a price increase smaller than the competitors’, say 4 percent, and assume it costs 2 percent of volume. Revenue becomes $40.0 million times 1.04 times 0.98, about $40.8 million, an extra $0.8 million. Ingredients scale with volume, not price, so they fall 2 percent, from $13.2 million to about $12.9 million, saving another $0.3 million. Labor and rent stay where they are. Profit goes from $2.5 million to about $3.5 million.
The recommendation. “Raise prices about 4 percent, which keeps you below the two biggest competitors and recovers about $1 million of the $1.5 million. Then go after the rest on the cost side: renegotiate the butter and flour contracts or lock in prices for a year, and look at shifting the product mix toward pastries and catering if their margins are higher. The risk is the volume assumption. If you lose more than 2 percent of items sold, the gain shrinks, so test the increase in a few stores first.”
That recommendation leads with the answer, gives the number, names the next lever, and states the risk and how to manage it. That is the shape interviewers listen for, whichever case type you are in.
Where the profit tree goes wrong
The tree is a tool for one kind of question, and most framework mistakes come from using it everywhere.
Using it for a different case type. Should a client enter Brazil? Should it buy a competitor? What should a new product cost? Those are market entry, M&A and pricing cases. Profit shows up in them, but as the last step, not the opening structure. Starting a market entry case with “revenue and costs” tells the interviewer you have one framework.
Skipping the market branch. A candidate who diagnoses the client’s ingredient costs without asking whether competitors face the same increase will recommend fixing a supplier relationship when the real answer is pricing.
Treating every branch as equally likely. Listing ten cost lines at equal depth wastes minutes. Say which branch you expect the problem in, and why, then test it.
Confusing margin with profit. A margin can fall while profit dollars rise, if volume grew. Say which one the client cares about.
Holding variable costs fixed when volume changes. In the bakery fix above, forgetting that ingredients fall with volume would understate the gain by $0.3 million. Interviewers check for exactly this.
Practice the diagnosis out loud
Reading a profitability case is quick. Running one is slower, because the interviewer only gives you data when you ask for it, and the order you ask in is part of the score. The way to build the habit is to run full cases where someone holds the data back.
Several of the 38 interviewer-led cases on Tough Tongue AI are profitability or cost cases with real exhibits: Events.com (market share rising while profit falls, and a pricing root cause), Fruit Drink Co. (one per-unit cost exhibit, an easy start), Garthwaite Healthcare (an insurer whose costs outrun the market), High Q Plastics (rebuilding a plant P&L), Race to the Bottom (an airline losing profit faster than its rivals) and Pink Cross (a payer-provider with two chained break-even problems). The AI interviewer waits for your structure, answers only what you ask, puts each exhibit on screen, makes you find your own arithmetic errors, and grades you against the case’s answer key.
The case interview prep hub lists every case by casebook and difficulty, and the McKinsey practice cases page maps each question type to the cases that drill it. Market sizing often sits inside a profitability case as well; the guide to market sizing questions works six of them end to end.
