Case Practice Partner: How to Make a Non-Expert Useful
Bain publishes a good answer and a better answer for one question. That is a marking key, and it lets someone with no consulting background run a case on you.
- The common objection to a non-consultant partner disappears once you realize Bain publishes graded model answers. Your partner reads the key; they do not need judgment of their own.
- McKinsey's four published cases carry suggested answers, so a partner can run them from the page.
- BCG publishes twelve behaviors, six to show and six to avoid, which is a usable observation sheet for someone who cannot assess content.
- Bain's published one to two minute thinking window gives your partner something to time, which is the single most valuable thing an untrained partner can do.
- The role sheet, the three partner jobs and the rotation below are ours, and no firm publishes advice on practice partners.
Your partner does not need to know consulting. Bain publishes a good answer and a better answer for the same question, McKinsey publishes suggested answers for four cases, and BCG publishes twelve behaviors to watch for. A partner with those three pages can run a case that is genuinely worth your time. Sources checked 30 September 2026.
Why the experience objection collapses
The standard reason for not asking a friend is that they cannot judge your answer. That is true, and it has stopped mattering.
Bain's FashionCo case states two grades for the same brainstorming question. A good answer includes different types of ways to increase revenue, it says, and a better answer includes the pros, cons and practicality of each idea. Your partner does not have to know which answer is better. They have to read whether you added the pros, cons and practicality, which requires no background at all.
McKinsey's interviewing page supplies the same thing at greater length, saying its four cases contain logic, thought processes and suggested answers. A partner reading from a page with the answer on it is functionally an interviewer with a script.
| What your partner needs | Where it comes from |
|---|---|
| A case with figures | McKinsey's four, Bain's two, or our case library |
| An answer to compare against | Published with all six of those cases |
| A list of what to watch | BCG's twelve behaviors |
| Something to time | Bain's one to two minute window |
| A reason to interrupt | The brief you give them, below |
Reading a worked example is not producing one.
Score my answerThe three jobs a practice partner gets, and nothing else
Give your partner three jobs. Adding a fourth makes them try to coach, which is where an untrained partner does damage.
| Job | What it looks like | Why it matters |
|---|---|---|
| Timekeeper | Says nothing except the elapsed time at 2, 20 and 38 minutes | Bain's published window is the only firm-sourced clock in a case |
| Interrupter | Cuts in once, at a moment you have pre-agreed | This is the dimension solo practice cannot reach |
| Observer | Ticks a behavior sheet, no commentary until the end | BCG publishes the sheet, so no judgment is needed |
The interrupter job is the one you are paying for. Alone, nobody redirects you mid-structure or asks you to defend a number, and BCG's page assumes feedback arrives while you work, since it lists ignoring feedback among the things to avoid.
Agree the interruption in advance. Two useful forms: after you name your first branch, they ask why that one first. After you give a number, they ask what it means. Neither requires them to know whether you were right.
The brief you hand your practice partner
Write this once and reuse it. Five lines, and it is the difference between a useful session and a polite one.
- Read the case prompt aloud once, then stop talking. If I ask a clarifying question, answer it from the page, and if the page does not say, tell me it does not say.
- Start the timer. Tell me when two minutes have passed. Nothing else at that point.
- Interrupt me exactly once, at the moment we agreed, with the question we agreed.
- Tick the behavior sheet while I talk. Do not react, do not nod encouragingly, and do not explain the case.
- At the end, read me the published answer. Then tell me only what you ticked, in the order you ticked it.
Bain's page supports line one directly, since it tells candidates to ask the interviewer questions if they feel confused or need more information. A partner who improvises an answer to a clarifying question has changed the case, which is why the brief says to say when the page is silent.
The observation sheet, from BCG's own list
BCG's case interview preparation page publishes six behaviors to show and six to avoid. Hand your partner both columns and ask for ticks.
| Show | Avoid |
|---|---|
| Listen actively | Panic under pressure |
| Be yourself | Rush your response |
| Think structurally | Overcomplicate solutions |
| Communicate clearly | Ignore feedback |
| Clarify assumptions | Stay silent |
| Show your thinking | Interrupt |
Twelve boxes, no opinions required. Two of them are especially valuable from an outside ear, because you cannot hear them yourself. Staying silent is a real fault with a real threshold: a partner can tell you that you went quiet for eleven seconds while computing, which you will remember as two. And interrupting is invisible from the inside entirely.
The one box that needs a rule is Clarify assumptions, because a candidate can tick it by asking three pointless questions. Ask your partner to note what you asked rather than whether you asked.
Can you hold the practice under a clock?
The session with a practice partner, and the rotation
Sixty minutes for a pair, split so both of you get something. Berkeley's career office puts an informal professional conversation at 20 to 30 minutes and tells people to respect the agreed timeframe, which is a good habit to import: a session that runs long stops happening weekly.
| Minutes | What happens |
|---|---|
| 0 to 25 | You take the case, they run the brief |
| 25 to 35 | They read the published answer, then their ticks |
| 35 to 60 | Swap, if they are also preparing |
If your partner is not a candidate, the second half is yours to use for the repair work: redo the worst moment three times while they watch the clock. That is a smaller ask than a full case and it is the block most people skip.
Twice a week is our working rhythm, and it is ours rather than any firm's. Two partnered sessions plus four solo hours covers all six dimensions we score on, which our method page sets out with Coverage 45 as the heaviest structuring weight.
The material rule nobody mentions
One boundary applies to what your partner is allowed to bring. McKinsey's assessment integrity expectations ask candidates to refrain from sharing information about the content of their assessment or interview with others, including specific questions and problem-solving scenarios.
So a partner who interviewed last month and wants to run you through what they were asked is offering something the firm has explicitly asked them not to share. The published cases exist precisely so that this is unnecessary. Six of them, free, with answers, from two firms.
That rule also cuts the other way and it is worth saying plainly. If you sit an assessment and your partner asks what was in it, the correct answer is that you cannot say.
There is a practical argument alongside the rule. A recalled question is a distorted question. Someone reconstructing an exhibit from memory a month later produces numbers that do not reconcile, and a case whose figures do not reconcile teaches the wrong lesson: you spend the session hunting for an arithmetic path that was never there. The published cases have the advantage of being correct.
The same applies to material your partner finds online with no attribution. If nobody will say where a case came from, you cannot tell whether it is a firm's published exercise, a school's copyrighted casebook, or someone's recollection of a sitting. We link to the published sources and to school casebooks rather than reproducing either.
What a case interview practice partner cannot give you
They cannot give you a grade. No firm publishes a mark scheme, Bain's two labels cover one question on one case, and a friend's impression of your performance is an impression.
They cannot give you a new case once the six published ones are used. That is a supply problem rather than a partner problem, and our case library exists for it.
Finding the partner at all is the step before any of this, and a market exists for that step. Our review of the peer matching board that pairs candidates with each other prices its free and paid tiers, states how many meetings a free account actually gets, and sets out what a board can and cannot promise about the person on the other side of the call.
They cannot give you consistency either. A partner who is also a candidate has their own round next week, and the sessions stop when their calendar fills. Plan for the gap instead of being surprised by it, which in practice means keeping the solo protocol running underneath.
And they cannot give you the pressure of a stranger. A friend across a kitchen table is a different physiological event from a manager across a desk, and the gap shows up in the first ninety seconds. Our mock case interviews page covers building a session that gets closer to real conditions, and practicing cases alone covers the weeks when nobody is available.
One last thing a partner is unexpectedly good for. Ask them, at the end, to repeat your recommendation back in their own words. They have no stake in it and no vocabulary to hide behind, so what they give you is the version that actually landed. If it differs from what you meant, the Answer first weight on our synthesis rubric, 35 of 100, is where you lost the points.
See your answer scored, dimension by dimension.
Reading a worked example and producing one are different skills. Work a full case and see it scored, dimension by dimension.
Sources
Primary sources only: the firm, the test vendor, or the institution concerned. Each carries the date we last read it.
- bain.com/careers/hiring-process/interviewing/fashion-case-study/the published good and better model answers, the one to two minute thinking window, the instruction to ask clarifying questions, and the step by step walkthrough instructionchecked 2026-09-30
- mckinsey.com/careers/interviewingthe four published cases containing logic, thought processes and suggested answerschecked 2026-09-30
- careers.bcg.com/global/en/case-interview-preparationthe six behaviors to show and the six to avoid, and the five named tasks inside a casechecked 2026-09-30
- mckinsey.com/careers/assessment-integrity-expectationsthe prohibition on sharing assessment content with others, which bounds what material a partner may bringchecked 2026-09-30
- career.berkeley.edu/explore-careers/informational-interviews/the 20 to 30 minute norm for an informal conversation and the instruction to respect an agreed timeframechecked 2026-09-30
Edits the firm and case guides on this site. Reads each claim about a recruiting process back to the firm's own page before it is published, keeps the source ledger dated, and owns the corrections.
How this guide was made
- Written against a published editorial process, not from memory. Every claim about a firm's process is either linked to that firm's own page or absent. We say what we could not verify rather than filling the gap.
- Each page carries a dated source ledger, below. When a firm changes a programme or a deadline, the date tells you how far to trust what you just read.
We look for first-hand candidate accounts for every firm guide. Where we found none we could source, the page says so and carries the firm's own words instead. Nothing here is dressed up as testimony, and if your own experience contradicts what you read, tell us and we will correct it and date the change.
Something wrong or out of date here? Tell us and we will correct it and date the change.
