Practicing Cases Alone: What Works and What Degrades

The verdict on practicing cases alone, dimension by dimension

The six dimensions and their weights below come from our own grading code, published on our method page. No firm publishes a mark scheme.

The two failures are specific rather than general, which is the useful part. Brainstorming weights Breadth at 30 and Buckets at 30 out of 100. Alone, most candidates stop at the point where recall becomes effortful, which is the exact point where the score starts accruing.

Synthesis weights Answer first at 35. You can practice delivering an answer first, and a recording will show whether you did. What you cannot practice is the question that follows it, which is where a weak answer is exposed.

The material that is genuinely self-workable

Two firms publish cases with answers, which is what makes solo work possible at all.

McKinsey's interviewing page says its four cases contain logic, thought processes and suggested answers. A suggested answer is a marking key, and having one changes solo practice from guesswork into checking.

Bain goes further on one question. Its FashionCo case labels a good answer as one that includes different types of ways to increase revenue, and a better answer as one that adds the pros, cons and practicality of each idea. Two grades on the same question, published. Score your own attempt against both and you have a reading of your own ceiling.

The Electro-Light case is the cleanest quantitative one to work alone, because every input is on the page. A 16-ounce bottle, described as one eighth of a gallon, priced at $2 to retailers; $40 million of fixed costs; $1.90 per bottle to produce and deliver. Contribution is $0.10 a bottle, so break-even is $40,000,000 divided by $0.10, which is 400 million bottles, and at 8 bottles to the gallon that is 50 million gallons. Check the margin as your sanity step: $0.10 on $2.00 is 5 percent.

One warning about everything else. McKinsey's assessment integrity expectations prohibit sharing the content of an assessment, including specific questions and problem-solving scenarios. Cases circulating from someone's sitting are outside what the firm permits, whatever the site hosting them says.

The recording protocol for practicing cases alone

A recording is the closest substitute for a listener, and it is the only one that catches delivery faults. Run it this way.

1. Set a timer for forty minutes and start recording audio. Video adds nothing here; the faults are auditory.
2. Read the prompt once, aloud. Then take Bain's one to two minutes and say the structure without looking at the page again.
3. Do the arithmetic out loud, with units in every line. McKinsey's rule is pen and paper only, so keep the calculator in another room.
4. Deliver a thirty-second synthesis at the end, cold. Answer, reasons, risk, next step.
5. Listen to it the next day rather than the same hour. Same-day listening hears the effort; next-day listening hears the answer.

Three faults show up on the recording and nowhere else. Silence longer than four seconds while you compute, which BCG names as something to avoid. A structure that arrives as five minutes of hedging around ninety seconds of content. And a synthesis whose first sentence is context rather than an answer, which is the Answer first weight going unclaimed.

Substitutes for the two dimensions that degrade in a case alone

These are ours. No firm publishes advice on practicing without a partner, and we are not going to pretend one does.

For brainstorming, use a forced count and a hard timer. Write the number before you start. Twelve ideas in three named groups, two minutes, and keep writing past the point where it stops being pleasant. The last three ideas are where Specificity is won, and alone they are the three you skip.

For synthesis, write the follow-up question yourself, then answer it. After each synthesis, write the single question you would least like to be asked. Set the paper down, wait an hour, then answer it in thirty seconds. Predicting the question is a weaker exercise than being asked it and it is a real one, because the question you least want is usually the risk you failed to name.

For both, use a vendor practice test where one exists. Sova's candidate preparation hub publishes a free practice test covering situational judgment, video interview and three types of ability question, and it states that practice answers do not count toward your score. It also notes that the activities in a real assessment are selected by the employer, so treat it as familiarity with the interface rather than a preview of content.

The self-marking sheet

Twenty minutes after each case, score four artifacts. Nothing else.

Write the score and the date. One session is noise. Four sessions over two weeks show you which dimension is moving, and that is the only thing the sheet is for.

The twenty-minute gap before you score matters. Immediately after a hard case your judgment of your own performance tracks how difficult it felt, which is uncorrelated with what you produced. Twenty minutes is enough for the artifacts to look like artifacts again.

Score one rubric per session rather than all four. Scoring everything produces a page of numbers and no decision, and the point of the sheet is to pick tomorrow's drill. If the structure was the weak artifact, tomorrow is a structuring hour, and the structure drill returns a Coverage score without a partner.

One habit belongs here rather than in the protocol. Keep the papers. A structure written three weeks ago, reread cold, is the clearest evidence you will get about whether your default framework has actually changed or whether you have only got faster at writing the same one.

What practicing cases alone leaves open

Two things cannot be closed alone, and it is better to know which.

The first is interruption. A real interviewer redirects you mid-structure, hands you an exhibit you did not ask for, and rejects your recommendation. BCG's case preparation page lists ignoring feedback among the behaviors to avoid, which presupposes feedback arriving while you work.

The second is the defense of a number. Bain's page tells candidates to ask the interviewer questions when confused, which is a two-way act. Alone, every number you produce stands unchallenged, and a candidate who has never been asked why becomes fluent at asserting rather than at reasoning.

There is a third gap worth naming because candidates mistake it for nerves. Alone, you never experience the cost of an interviewer waiting. The silence in your recording is uncomfortable to hear and it is nothing like the silence of a person holding a pen across a table. Simulation closes part of that and only part.

So the honest allocation is most sessions alone and some with a person. Our case practice partner page covers how to make a non-expert genuinely useful, which is easier than it sounds because Bain published the marking key. Our mock case interviews page covers the full-condition version. And the case interview math drill returns a score on the mathematics dimension without needing anybody, which is the one place a tool substitutes cleanly for a listener.