Business intuition in a case interview is calibrated judgment: knowing well enough what a plausible gross margin, market size, cost structure or growth rate looks like that you notice when a number is wrong. It is not instinct and it is not a substitute for analysis. Its practical function is error detection and hypothesis generation, and both are visible to an interviewer within seconds.
Candidates usually try to build it by reading business news, which is slow and poorly targeted. It is built more efficiently by learning a small number of reference structures and then practising prediction against them.
What it looks like in a case
Three moments, all of which an interviewer registers:
Catching an implausible result. You calculate a net margin of 40 per cent for a grocery chain and stop, because grocery runs on low single-digit net margins. The error is a slipped decimal or the wrong denominator, and you find it before you say the number. A candidate who reports 40 per cent confidently has demonstrated the opposite of judgment.
Choosing where to look first. Told that a manufacturer’s costs have risen, you ask about raw materials and labour before overhead, because in most manufacturing those two dominate. That is not a guess; it is knowing the shape of a cost base.
Sizing a claim. Told a proposal will save $50 million, you register that this is a fifth of the client’s total cost base and therefore a very large claim requiring very good evidence. Without the reference point, $50 million is just a number.
None of this replaces the analysis. It directs it, and it catches the errors.
Learn margin structures first
The single highest-return thing to learn is what money looks like in different sectors, because it is stable, compact, and it drives most recommendations.
The pattern worth internalising is the relationship between gross and net margin, and what sits between them:
| Sector | Gross margin | Net margin | What eats the difference |
|---|---|---|---|
| Grocery retail | Low | Very low, single digits | Store labour, rent, distribution, shrinkage |
| Fashion retail | High | Low to moderate | Markdown, rent, unsold inventory |
| Packaged goods | Moderate to high | Moderate | Marketing, trade spend, distribution |
| Software | Very high | Varies widely | Sales and marketing, R&D |
| Airlines | Thin and volatile | Thin, cyclical | Fuel, fleet, labour, high fixed base |
| Professional services | Moderate | Moderate | People — the cost base is almost entirely salary |
| Heavy manufacturing | Moderate | Low to moderate | Materials, energy, capital intensity |
The exact figures matter less than the shape. What you need is the ability to say “that seems high for this kind of business” and be right often enough to be worth listening to.
Two consequences follow directly and are worth knowing as facts rather than deriving each time:
A thin-margin business is leveraged to small changes. If net margin is 2 per cent, a one-point improvement in cost of goods roughly doubles profit. This is why grocery and airline cases so often turn on operational detail that looks trivial.
A high-fixed-cost business is leveraged to volume. Airlines, hotels, software and manufacturing all share the property that the marginal unit is far cheaper than the average unit, which is why capacity utilisation is usually the first question in those sectors.
Build reference points, not general awareness
Reading widely produces the feeling of business knowledge without the calibration. What produces calibration is a small set of anchors you can actually recall under pressure.
Worth knowing roughly:
- Population and household counts for the markets you might be asked about — your own country, the United States, major economies. Canada around 40 million and roughly 15 million households; the US around 340 million. Market sizing answers are built on these.
- Typical cost splits. Restaurant: roughly a third food, a third labour, a third everything else including rent. Airline: fuel and labour dominating. Retail: cost of goods dominating, then rent and store labour. Professional services: salary as the overwhelming share.
- Order-of-magnitude revenues. What a single grocery store, restaurant, hotel or bank branch turns over in a year. These make bottom-up estimates possible.
- Growth rates that are normal. A mature consumer category growing low single digits; a developing technology category growing double digits. A claim of 40 per cent growth in a mature category should prompt a question.
Thirty of these, held roughly, are worth more than three years of reading the financial press, because they are retrievable in the moment.
The practice that actually builds it
Predict before you check. This is the whole method. When you read that a company has revenue of some amount, estimate the margin before reading it. When you pass a business, estimate its daily revenue. Write the estimate down, then find the real figure. The gap is the feedback, and feedback is what turns exposure into calibration.
Reading without predicting produces almost none of this, which is why the years of business news do not deliver what candidates expect.
Work backwards from familiar businesses. Take a coffee shop you know. How many customers an hour? What is the average spend? What is the rent likely to be? How many staff at what wage? Build the P&L, then ask whether the profit you arrive at seems plausible for that business. When it does not, find which assumption is wrong. This exercise builds the mechanism-level understanding that lets you reason about a business you have never encountered.
Read the industry primers rather than the news. The industry primers on this site are written for exactly this purpose: each covers the economics of a sector — what drives cost, what drives price, what the recurring analytical traps are. The forest products primer, for instance, explains why recovery rates and mill utilisation drive profitability and why comparing across different measurement bases produces wrong answers. That is the kind of knowledge that changes a case answer.
Ask why after every case. Not “was I right” but “why is this business shaped this way”. A case answer you understand mechanically becomes a reference point for the next one. A case answer you only remember does not.
Using it without over-reaching
The failure mode is asserting rather than hypothesising, and it is a serious one because it looks like the thing being rewarded.
| Over-reaching | Using it properly |
|---|---|
| “The problem is clearly labour cost” | “I would expect labour to be a large share here — could we see the cost breakdown?” |
| “Margins in this industry are 8 per cent” | “I would expect single-digit net margins in this sector. Is that roughly right?” |
| “They should raise price” | “Price looks like the largest lever given the margin structure, but I want to check volume sensitivity first” |
The right-hand column does three things: it states the prior, it makes it checkable, and it commits to testing rather than concluding. Interviewers reward this pattern specifically, because it is what a consultant does with a hypothesis.
And when the data contradicts the intuition, say so out loud and follow the data. “I expected labour to dominate and it does not — the largest line is distribution, which is unusual for this kind of business and worth understanding” is a stronger moment than either being right or quietly abandoning the prior.
Where it fits
Intuition supports structure; it does not substitute for it. Structure is what a case interview assesses, and a candidate with sharp instincts and no framework fails. The value of calibration is that it makes structured analysis faster and catches the arithmetic errors that would otherwise reach the recommendation.
Build it alongside the mechanics rather than instead of them. The case maths techniques guide covers the sanity-checking habit this feeds, the industry primers supply the sector economics, and the case interview hub sets out where both sit in a full preparation plan.
