- Structuring is the moment, usually right after the clarifying questions, when you lay out how you will answer the client's question. It is also how you organise any open question later in the case.
- Start from the decision the client has to take, not from a framework you memorised. A structure that could fit any case tells the interviewer nothing about how you understand this one.
- Use three or four areas that are specific to the case, do not overlap and together cover the problem. Go one level deeper only where it matters.
- A structure must say where you would start, and why. Without a priority it is only an index.
- Present it top-down in about a minute, then use it: come back to it as data arrives and change it openly when the facts require it.
Most candidates know that they should "structure the case". Fewer know what the interviewer is actually looking for in those first minutes, and why a neat list of boxes can still score poorly. This guide starts from the basics, then works through three fictional cases where a weaker and a stronger structure sit side by side.
What structuring means in a case interview
A case interview gives you a business problem and about 25 to 35 minutes to reach a recommendation. Structuring is the step where you turn the open question into a plan: the few areas you need to understand, the questions inside each one, and where you would begin. Interviewers often call it a framework, an issue tree or simply "your approach". The name matters less than what it does.
A structure has three jobs. It shows the interviewer that you understood the question. It gives both of you a map, so that every later analysis has a place. And it makes your thinking visible, which is what lets the interviewer judge it.
When it happens
The main structure comes early. After the prompt and two or three clarifying questions, you usually ask for a moment to organise your thoughts, take 60 to 90 seconds in silence, and then present your plan. In an interviewer-led case, which candidates often report at McKinsey, the first question may simply be "how would you approach this?". In a candidate-led case, which is more common at BCG and Bain according to candidate reports, you will also be expected to use the structure to choose each next step.
Structuring does not stop there. The same skill is tested whenever the interviewer asks an open question mid-case ("what could explain this?", "how could they reduce costs?", "what risks do you see?") and again in the final recommendation, which is a structure in miniature: answer, reasons, risks, next steps.
What interviewers look at
Interviewers rarely score the structure on its own. They note whether it fits the problem, whether it holds together, and whether you use it. In practice, they look for five things:
Built on the question
It starts from what the client must decide and uses the facts in the prompt.
Specific areas
Three or four areas that belong to this case, not to every case.
No overlaps, no gaps
Each idea has one place, and together the areas would answer the question.
A priority
A hypothesis and a clear place to start, with a reason.
Clear delivery
Top-down, numbered, about a minute, and used again later in the case.
Official guidance from the firms points the same way. McKinsey's interviewing page says there is not one logical solution and encourages candidates to adapt their approach to the problem. Bain's case interview page says it is not looking for a single right answer but for how you think and structure problems, and BCG's preparation page asks candidates to structure their approach. That is a useful test for everything that follows: would your structure change if the problem changed?
Frameworks: a starting point, not the answer
Standard frameworks, such as profit equals revenue minus costs, or the familiar lists for market entry, are useful when you start preparing. They show which areas a business problem can touch, and they help you avoid forgetting something obvious. Knowing them is not the problem.
The problem starts when a framework replaces thinking. If the prompt says that the client's competitors are doing fine, a structure that spends half its time on "market trends" has ignored the most important fact you were given. If the client is a museum, a hospital or a ministry, a framework designed for companies may miss the real objective. Interviewers hear generic structures every day, and they recognise them within seconds.
A practical way to use frameworks is to remember only the first layer, as a checklist of what could matter, and to build the second layer fresh for every case from the facts in the prompt. Then check the result against the question: does each area help answer it?
For some questions a standard split is exactly right. A fall in profit does break down into revenue and costs, and saying so is fine. What makes the structure strong is what you put inside each branch, and whether it reflects this client's situation.
How to build a structure, step by step
The six steps below work for most prompts, including the ones that do not look like business cases. With practice they take less than two minutes.
1. Start from the decision the client has to take
Before you write any area, write the question in one line, in terms of a decision: "should they invest?", "how can they restore margin to 8% within two years?", "which of the two options should they choose?". Then add how the client will measure success, which is often what your clarifying questions were for. A target and a deadline change the structure: restoring margin in six months points to different levers than restoring it in three years.
2. Choose how to split the problem
Look for the logic that the answer has to follow. When the question has a clear arithmetic, such as profit, revenue or cost per unit, the split can follow it. When the question is a yes or no decision, a natural split is the set of things that must all be true for the answer to be yes: there is enough demand, the economics work, the client can do it, and the risks are acceptable. When the question is about a process, such as long waiting times, the split can follow the steps of the process.
3. Write three or four areas specific to this case
Name each area with words from the case. "Costs" is a heading; "the price paid to farmers for raw milk" is an area that shows you read the prompt. Three or four areas are a useful starting point. Choose the number that covers the question clearly, combining branches where that makes the approach easier to follow.
Then go one level deeper, with two to four questions in each area. These are the questions you would need answered, and they should point to data you could ask for.
4. Check for overlaps and gaps
Read the areas once more and ask two questions. Could any idea sit in two areas at once? If a point about price could go under "revenue" or under "competition", decide where it lives and move it. And if you had a full answer for every area, could you make the decision? If not, something is missing.
5. Say where you would start, and why
Look at the facts you have and form a hypothesis: the most likely explanation, or the most likely deal-breaker. Then say which area you would examine first. The best starting point is usually the one that could change the answer fastest, or the one the prompt already hints at. "If demand is not there, nothing else matters, so I would start with demand" is a complete and convincing priority.
6. Present it top-down, and adapt it later
Give the number of areas first, then name them, then go into each one briefly. Close with your hypothesis and where you would start. Afterwards, keep the page in front of you: when data arrives, say which area it answers, cross out what you have ruled out and add a branch if the facts require it. A structure that you change openly, with a reason, is a strength, not an admission of failure.
Example 1: a margin decline
The first example is the most common type of case, a fall in profit. The point is to show how much the facts in the prompt should shape a structure that starts from a standard split.
Your client, Linden Dairy, makes milk and yoghurt and sells almost all of it to four large supermarket chains. Its revenue has been stable at about £400m for three years, but its operating margin has fallen from 8% to 3%. The CEO asks: why has our margin fallen, and how can we get it back to 8% within two years?
The weaker structure
- Revenue: price, volume
- Costs: fixed costs, variable costs
- Market: size, growth, competitors, customers
- Company: capabilities, strategy
What goes wrong
Nothing in it is false, but it could be the answer to almost any profit case. It ignores the two facts the prompt gives: revenue is stable, and the customers are four powerful supermarkets. "Market" and "company" overlap with each other and with revenue (competitors affect price). And it has no hypothesis, so the interviewer cannot tell where you would go next.
The stronger structure
Margin fell by 5 points, about £20m a year, while revenue stayed flat. So either costs rose or the revenue mix changed while the total stayed similar.
- 1. Input costs per litre: what rose, and by how much?
- Raw milk price paid to farmers
- Energy for processing and cooling
- Packaging and transport
- 2. Price and mix: could Linden pass higher costs on?
- Contract terms with the four supermarkets: fixed prices, timing of price reviews
- Shift from branded products to cheaper supermarket own-label
- Promotions that Linden pays for
- 3. Efficiency of the plants
- Use of plant capacity
- Waste and spoilage
- Labour productivity
- 4. Is this Linden or the whole industry?
- Margins of other dairies over the same period
- Supermarkets' bargaining power across their suppliers
- Hypothesis and start: input costs rose across the industry, and Linden's contracts did not let it raise prices. I would start by comparing cost per litre and price per litre over the three years.
Why it is stronger
It uses both facts: flat revenue points the analysis to costs and mix, and the supermarket contracts become an area of their own because that is where the problem probably sits. Each idea has one home: competitor margins sit in area 4, not inside price. And it closes with a hypothesis and a first analysis that would confirm or rule it out quickly.
Notice that the stronger structure still contains revenue and costs. It did not throw the standard split away; it filled it with the client's facts and added the one area the prompt was pointing to. If you want to see a structure used through a whole case, from the first question to the recommendation, the complete case interview guide follows a fictional gym chain step by step.
Example 2: an investment decision
The second example is a yes or no decision that does not fit a profit framework. Candidates often reach for a market-entry list here, because the client is entering a new activity.
Alder Valley is a ski resort in the Alps. Almost all its revenue comes from December to April, and warmer winters have made the season shorter. The board is considering investing €20m in a summer mountain-bike park, using the existing gondola to carry riders and bikes uphill. Should it go ahead?
The weaker structure
- Market: size and growth of mountain biking
- Competition: other bike parks, their prices
- Customers: segments, needs
- Company: capabilities
- Entry mode: build, partner, acquire
What goes wrong
It treats the question as entering a new market from scratch. It misses the two things that make this case specific: the resort already owns the gondola, hotels and staff, and the reason for the project is a shrinking winter. "Entry mode" is barely relevant, there are five areas, and there is no view on what would decide the question.
The stronger structure
The board needs to know whether €20m spent on summer would earn an acceptable return and reduce dependence on winter.
- 1. Summer demand: would enough riders come?
- Current summer visitors, if any
- Comparable bike parks: visitors, prices, distance from cities
- Price per day pass, and spend on rental, food and lodging
- 2. Economics: does the investment pay back?
- Extra revenue per season
- Extra costs: trail building and upkeep, summer staff, insurance
- Payback compared with the board's target
- 3. Fit with the winter business
- Shared assets: gondola, hotels, staff kept all year
- Effects on winter: slope damage, maintenance windows
- 4. Feasibility and risk
- Permits and environmental limits
- Weather risk in summer, time to build
- Hypothesis and start: the project could work because the gondola and staff already exist, so most costs are shared; the open question is whether riders would come. I would start with demand, because without it nothing else matters.
Why it is stronger
It frames the decision (return and dependence on winter), uses the assets mentioned in the prompt, and turns "should we?" into four conditions that must all hold. The priority comes with a reason. Every area also points to data the interviewer could hand over next, such as visitor numbers at comparable parks.
For each area, ask yourself what number or fact you would request first. If you cannot name one, the area is probably too vague to be useful, and the interviewer will notice.
Example 3: structuring a mid-case question
Later in the Linden Dairy case, the interviewer asks: "What could Linden do to restore its margin?". This is a structuring question too, and it is where many candidates switch to a list of ideas.
The weaker answer
"They could renegotiate with supermarkets, cut energy costs, reduce waste, launch premium yoghurts, export, buy a competitor, automate the plants, reduce promotions, and switch suppliers."
What goes wrong
Many of these ideas are good, but they come as a list of nine. The interviewer has to sort them, cannot see which ones matter most, and has no sense of their size or timing.
The stronger answer
"I see three groups of levers. First, price: renegotiate the price review clauses so that milk costs pass through, and cut the promotions Linden funds. Second, cost: reduce waste and energy use in the plants, and buy packaging jointly. Third, mix: grow branded and premium products, where Linden sets the price. Within two years, price and promotions are probably the biggest and fastest lever, so I would size that first."
Why it is stronger
The same ideas, grouped into three named categories that do not overlap, with a view on which one matters most and a proposed next step. Fewer ideas, well organised, are worth more than a long list.
How to present a structure in about a minute
A good structure presented badly loses much of its value. The interviewer cannot see inside your head; they hear words in sequence. Top-down delivery gives them the map before the detail.
- Ask for a moment. "Could I take a minute to organise my thoughts?" is normal and expected. Use the time in silence, and write.
- Restate the goal in one sentence. "To get back to 8% margin in two years, I want to understand where the 5 points went and which levers can bring them back."
- Give the number of areas, then name them. "I would look at four areas: input costs, price and mix, plant efficiency, and whether this is Linden or the industry."
- Go through each area briefly. One sentence per area with its two or three key questions. Do not read every sub-point if time is short.
- Close with the hypothesis and where to start. "My hypothesis is that costs rose and the contracts stopped us passing them on, so I would like to start with cost and price per litre."
"So, I'd look at revenue. Revenue is price times volume, and on price I'd want to know about the contracts, and promotions, and own-label, and on volume I'd check if any supermarket reduced orders... Then costs, which are fixed and variable... and I'd also look at the market and competitors."
"I would look at four areas. First, input costs per litre: milk, energy, packaging. Second, price and mix: whether the supermarket contracts let Linden pass costs on. Third, plant efficiency. Fourth, whether other dairies have the same problem. Since revenue is flat, my hypothesis is that costs rose and the contracts blocked price increases, so I'd start by comparing cost and price per litre."
The weaker version expands the first area before the interviewer knows how many there are, so it is hard to follow and runs long. The stronger one gives the map first, numbers the areas, takes under a minute and ends with a priority.
On paper, write the question at the top and the areas side by side underneath, with space to add notes. In a video interview, you may be asked to hold your page up to the camera; follow the firm's rules on screens and sharing; neat, numbered areas make that easy. Ask your recruiter how written work is handled in your format.
Adapting the structure as the case moves
Data will answer some areas and close others. Say so: "This confirms that milk costs rose 15% while prices stayed flat, so area 1 explains most of the gap. I'd now look at area 2, the contract terms." If a new fact opens a question your structure did not cover, add it openly: "That's new; I'll add a branch for the loss of one supermarket contract." Interviewers value candidates who use their structure as a working tool rather than a speech.
Common structuring mistakes
The mistakes below are the ones that come up again and again in practice sessions with candidates. Most of them are easy to fix once you know to look for them.
- A generic structure. The areas would fit any case, and nothing in them refers to the prompt. Fix it by using at least one fact from the prompt in each area.
- Overlapping areas. The same idea appears twice, or revenue and cost points are mixed back and forth across areas. Fix it by deciding where each idea lives.
- Drivers placed at the top level. Good detailed points, such as "fuel prices" or "staff turnover", are presented as main areas, with no higher-level logic that groups them. Fix it by asking what the points have in common and naming that group.
- A list instead of a structure. Eight or ten ideas in a row, especially in mid-case brainstorming. Fix it by grouping them into two to four named categories before you speak.
- No priority. The structure ends with the last area and no hypothesis. Fix it by always closing with "I would start with... because...".
- Bottom-up delivery. You expand the first area before naming the others, so the interviewer loses the map. Fix it by giving the number of areas first.
- No thinking time. Starting to talk straight after the prompt often produces a weaker structure than taking a minute. Ask for the time.
- A structure that is never used again. You present it, then the case continues as if it did not exist. Come back to it after each analysis.
- Too much detail. Five areas with five sub-points each take three minutes to present. Keep three or four areas and the questions that matter.
None of these mistakes is fatal on its own, and interviewers expect some imperfection under time pressure. What weighs is a pattern: the same generic structure in every interview, or a structure that never guides the analysis.
How to practise structuring
Structuring improves quickly with short, frequent drills, because each attempt takes only a few minutes. Full cases are still necessary, because that is where you learn to use the structure under pressure.
On your own: the two-minute drill
- Take a prompt you have not seen, from a case book, a business article or the list below.
- Set a timer for two minutes and write your structure on paper: the decision, three or four areas with their questions, a hypothesis and a place to start.
- Present it aloud in about a minute, ideally recording yourself.
- Check it: is each area specific to the prompt? Do any overlap? Could you decide if every area were answered? Did you say where to start?
- Rewrite it once. The second version teaches you more than the first.
Ten minutes a day of this, over two or three weeks, usually does more for structuring than reading another collection of frameworks. Calculations and charts are the other skill you can drill alone in short sessions; the case maths guide has exercises and a drill set with answers.
Practice prompts
All the prompts below are fictional. For each one, a short note says what a strong structure would need to capture. Try your own structure first.
A regional bus operator's passenger numbers have fallen 15% in two years. Its board asks why, and what to do.
A strong structure separates the reasons people stopped travelling (fares, reliability, routes, new alternatives such as cycling or remote work) from what the operator controls, and asks whether the fall is spread across all routes or concentrated in a few. Starting with the routes that lost most riders is a sensible priority.
A family-owned chain of 12 hardware stores is considering launching online sales. Should it?
The structure should test demand from its existing customers, the economics of delivery for heavy and bulky products, the capabilities the chain would need, and the risk of moving sales from stores to a less profitable channel. Delivery economics is often the deal-breaker for heavy goods, which makes it a good place to start.
A city hospital wants to cut the number of patients who miss their appointments. What would you look at?
A good split follows the patient's path: booking, reminders, the day of the appointment, and what happens after a missed visit. It also asks which appointment types and patient groups miss most, because that is where to start.
A software company with 2,000 business customers is losing 20% of them every year. How can it reduce this?
The structure can separate customers who leave in the first months from long-standing ones, then look at product fit, onboarding, service and price, and at what competitors offer. Finding out when in the customer's life the losses happen is the obvious first step.
With a partner, then with feedback
A practice partner adds what drills cannot: interruptions, follow-up questions and a real listener. Ask them to stop you if they lose track, to ask "why that area?" and to give you new data mid-case so you have to adapt. Practice with other candidates has a limit, though. A partner can tell you whether your structure sounded organised, but rarely whether an experienced interviewer would find it generic or would doubt your priority.
That second kind of feedback is what an experienced interviewer can add. In 1:1 coaching with Franco, structuring and prioritisation is the first of five dimensions scored in the written report after every session, with a short explanation of what held the score back. You can read a sample report for a fictional candidate to see what that looks like.
1:1 interview coaching
Find out how your structures sound to an experienced interviewer
Franco Borgonovo spent ten years at BCG, from intern to Principal, has conducted more than 1,500 interviews, over 800 of them for MBB firms, and is currently an official BCG coach for new hires. Every session is a realistic interview with him, followed by a written report with your scores and priorities.
- Start with one session for a diagnosis and a plan (US$249)
- Packages of 3, 5 or 10 sessions add a progress dashboard, targeted self-practice materials and email support between sessions
- Money-back guarantee if the coaching does not meet your expectations
Not sure which option fits? Use the free call to tell Franco your target firms, interview dates and where you feel stuck.
Frequently asked questions
Should I memorise case frameworks?
Learn the common ones as a checklist of what can matter, then build each structure from the prompt. Remembering the first layer is useful; reciting a full framework without adapting it usually weakens your answer.
How long should I take to build my structure?
Around 60 to 90 seconds of silence is common and accepted. Presenting it should take about a minute. If you need a little longer for a complex case, ask; taking two minutes for a good structure is better than talking straight away.
What does MECE mean?
It stands for "mutually exclusive, collectively exhaustive": the areas do not overlap, and together they cover the problem. It is a useful check, but interviewers care more about whether the structure fits the question and leads somewhere.
Is an issue tree the same as a structure?
An issue tree is one way to draw a structure: the question at the top, branching into areas and then into more detailed questions. Interviewers do not expect a particular drawing, only a clear and logical plan.
What if I realise I missed something after presenting?
Say so and add it: "I'd like to add one area I missed: regulation." Interviewers see this as good judgement. What hurts is noticing a gap and carrying on as if it were not there.
How many areas should a structure have?
Three or four in most cases. Two can be enough for a narrow question; more than four is usually a sign that some areas should be grouped.
Checklist for your next structure
Your next step
Pick one of the practice prompts above, give yourself two minutes and present your structure aloud. Then check it against the checklist and rewrite it once. If you would like an experienced interviewer's view of how your structures come across, and which habits to change first, one coaching session gives you a diagnosis and a written plan to follow.
Coaching with Franco
Get precise feedback on your structuring
A realistic case interview, a debrief and a written report with your priorities. Start with one session, or talk it through for free first.
About Franco
Franco Borgonovo
Former BCG Principal · 10-year tenure · Official BCG coachFranco spent ten years at BCG, from intern to Principal, interviewing and selecting candidates as a global lead recruiter. He is still an official BCG coach for new hires, and today works one to one with candidates on applications, McKinsey Solve and interviews.
More about Franco →- McKinsey: Interviewing (careers page) · there is not one logical solution; adapt the approach to the problem; read 2 November 2026
- BCG: Case interview preparation (careers page) · structure your approach, think structurally; read 2 November 2026
- Bain & Company: Case interview (careers page) · no single right answer; structure your thinking; read 2 November 2026