Case Interview Drills: Six Dimensions, One Hour a Day

What firms publish about case interview drills, and what they do not

No firm we read publishes a drill programme, a repetition count or a scoring scale. Three publish something more useful: the material and the constraints.

McKinsey's interviewing page recommends reviewing four sample cases it publishes, and says they contain logic, thought processes and suggested answers. It adds that practicing them will help you become familiar with the types of questions the firm asks. Its published table on AI marks practicing interview questions as encouraged.

BCG's case interview preparation page names the five tasks inside a case: structure your approach, ask thoughtful questions, analyze data, perform quick calculations, and identify the most important factors. Five tasks is five drills, and the page hands you the list without calling it one.

Bain publishes the two most useful constraints. On its FashionCo case it tells candidates to take one to two minutes to think about how they would approach the problem, which is the only published structuring time we found. And it labels its own model answers: a good answer includes different types of ways to increase revenue, while a better answer includes the pros, cons and practicality of each idea. That gap is a drill specification written by a firm without being labeled as one.

The six drills, and the weights we score them on

These weights are ours. They are read from our own grading code, no firm publishes a mark scheme, and any page that shows you one has invented it.

Read the structuring row against Bain's published gap. Coverage at 45 out of 100 is the heaviest single weight in the set, which encodes a claim: most structures fail by leaving a branch out rather than by being untidy. Bain's own wording points the same way, since a good answer covers different types and a better one adds judgment on each.

The exhibits row encodes the opposite claim. So what carries 30 and Observation only 25, because reading a chart correctly and drawing nothing from it is the common failure there.

The case interview math drill, worked on a published case

McKinsey's Electro-Light case supplies a clean break-even with published figures. Its own page reminds candidates that calculators are not allowed and suggests writing out the calculations on paper. Do it that way.

The published inputs are a launch in a 16-ounce format, described on the page as one eighth of a gallon, at a price of $2 to retailers; total fixed costs of $40 million; and a cost of $1.90 to produce and deliver each bottle.

1. Contribution per bottle. $2.00 minus $1.90 is $0.10.
2. Break-even in bottles. $40,000,000 divided by $0.10 is 400,000,000 bottles.
3. Convert to gallons. At 16 ounces a bottle there are 8 bottles to the gallon, so 400,000,000 divided by 8 is 50,000,000 gallons.
4. Sanity check the margin. $0.10 on a $2.00 price is 5 percent of revenue, which is thin. A thin margin is why the volume requirement is large, and saying that aloud is the Interpretation weight earning its 20 points.

The same page later asks what would be needed to achieve a target 12.5 percent share of the electrolyte-drinks market. That figure lets you work backward: if 50 million gallons is 12.5 percent, the market is 50 divided by 0.125, which is 400 million gallons. That last number is our inference from the published target rather than a figure the page states in text, and we flag it as such.

The drill lives in step four. Most candidates reach 50 million gallons and stop, which loses 20 of the 100 points on our mathematics rubric and, more importantly, leaves the interviewer to supply the meaning.

The one-hour split

Sixty minutes, five days a week, is enough to move all six dimensions. The allocation below is ours and it is deliberately lopsided.

Structuring and mathematics take half the hour because they appear in every case regardless of type. Synthesis takes five minutes because it is short by nature, and skipping it is how a candidate who understands a case still finishes badly. Every lesson in the course maps to one of those dimensions, and the seven of them run 56 minutes in total, each paired with graded drills on the same idea.

McKinsey's assessment integrity expectations make one of those rules non-negotiable rather than stylistic. In its interviews you may use only pen and paper for notes and calculations, no calculator, no prepared notes, and it says it may ask to review your notes. Drilling arithmetic on a phone calculator trains a motion the room forbids.

Picking tomorrow's case interview drill

The five-minute review block is the part that makes the other fifty-five compound. Score one artifact a day against one rubric above, then follow the rule.

1. If a branch was missing, drill structuring. Coverage is 45 points and a missing branch takes most of them.
2. If the arithmetic was right and silent, drill synthesis. Answer first carries 35, and an unstated answer scores zero on it however good the working.
3. If the number was wrong, check units before speed. Units carry 15 on the mathematics rubric, and per bottle against per gallon is exactly the error the Electro-Light chain invites.
4. If the list of ideas was long and vague, drill brainstorming. Breadth and Buckets are 30 each and Specificity is 20, so ten shapeless ideas score badly.
5. If the chart was described rather than used, drill exhibits. So what is the 30-point weight there.

BCG's page supplies the behavioral half of the same diagnostic. Among the things it says to avoid are rushing your response, overcomplicating solutions and staying silent. Those are drill outcomes rather than personality traits, and the timed blocks above are what change them.

What case interview drills cannot do

They cannot substitute for full cases. A drill isolates a dimension, and an interview asks you to hold six at once for forty minutes. Both are required, and our mock case interviews page covers the second.

They cannot supply a firm's mark scheme. No consulting firm publishes one. Bain's good-to-better gap is the closest published thing, and it covers one question on one case.

They cannot legally be built from another candidate's exam. McKinsey's integrity page prohibits sharing information about the content of an assessment, including specific questions and problem-solving scenarios. Material that circulates from someone's sitting is material the firm has told you not to use.

Our own drills run the six rubrics above with the weights shown, so a session returns a score on one dimension rather than a feeling. The case interview math drill and the structure drill are the two that carry the hour, and the method page explains how each dimension is scored. They teach no business content, because none of the dimensions above is a knowledge test. For the material itself, the free prep resources page lists what the firms publish at no cost.

Which skill each drill isolates, and what its scorecard returns, is set out in the drill engine page.