What to take away
- Analytical thinking carries half the score, so structure, maths and your recommendation matter more than polish
- Each of the fourteen skills is marked 1 to 5, where 3 means "met the first-round standard", not "average"
- Scoring is criterion-referenced: you are compared to a fixed bar, not ranked against other candidates that day
- A 4.5 weighted average reads as outstanding, 3.75 as a clear hire, and below 3.0 as not yet there
- Hedging costs more than being wrong: a reasoned answer that turns out incorrect outscores "I'd need more analysis"
- Asking the interviewer to repeat a number is never a deduction, so stop treating it as one
What the interviewer is filling in while you talk
Every MBB interviewer walks out of the room and completes a scorecard. It is not a single number and it is not a gut call written up afterwards. It is a grid: a set of named skills, each with its own score and a line of evidence justifying it. The debrief that follows is a conversation about those cells, not about whether the interviewer liked you.
This matters for preparation because it changes what improvement means. You do not get better at "case interviews". You get better at framing, or at synthesis, or at rigor, one cell at a time. Candidates who plateau are almost always drilling the skills they already have.
The rubric below is the one we mark against at MBB Ready, and it is modelled on the scorecards the two of us used as interviewers. Firms differ in vocabulary and in weighting at the margins, but the shape is remarkably consistent: a heavily weighted analytical block, a communication block, and a behavioural block assessed from your stories rather than from the case.
| Dimension | Weight | What it covers |
|---|---|---|
| Analytical thinking | 50% | Framing, commercial judgement, rigor, creativity, synthesis |
| Presence & communication | 25% | Presence, precision, listening, relationship-building |
| Track record & drive | 25% | Self-direction, collaboration, depth, growth mindset, results delivered |
The fourteen skills, and what each one is really testing
Fourteen skills sounds like a lot to hold in your head. In practice they cluster, and knowing which cluster a moment belongs to tells you what the interviewer is writing down.
Analytical thinking (50% of your score)
| Skill | What earns the mark |
|---|---|
| Framing | A structure that fits this problem, broken into parts that do not overlap, with a stated hypothesis or a named place to start |
| Commercial judgement | Business instinct: knowing which branch matters, what a number implies for the client, when an answer is too good to be true |
| Rigor | Maths that is right, or right in order of magnitude, with assumptions said out loud and a sense-check at the end |
| Creativity | Ideas beyond the obvious list, generated from a named cut rather than free-associated, with trade-offs attached |
| Synthesis | A recommendation first, then the reasons with numbers, then the risk, then the next step, inside sixty seconds |
Presence & communication (25%)
| Skill | What earns the mark |
|---|---|
| Presence | Calm under pushback, comfortable with silence, thinking out loud without losing the thread |
| Precision | Answering the question actually asked, in numbers where numbers are asked for, without padding |
| Listening | Catching and using the interviewer's nudges and clues rather than talking over them |
| Relationship-building | Reading like a colleague a partner would put in front of a client, not a candidate performing |
Track record & drive (25%)
This block is assessed from your behavioural answers, not from the case, which is why a case-only practice session cannot score it. It covers self-direction (what you start without being asked), collaboration (how you work through other people), depth (whether your stories survive three follow-up questions), growth mindset (what you did with hard feedback), and results delivered (whether anything measurably changed).
What a 1, a 3 and a 5 actually mean
The single most common misreading of a case scorecard is treating 3 as "average" and 5 as "perfect". Neither is right. The scale is criterion-referenced: each number describes a standard, and you are measured against that standard rather than against the other candidates interviewing that day. There is no curve and no forced distribution.
| Score | What it means |
|---|---|
| 5 | Did this as well as a strong hire does it: fully realised, or a genuinely insightful non-standard move that achieves the same end |
| 4 | Clearly above the first-round bar, with at most a minor rough edge |
| 3 | On the bar: met the first-round standard. Solid and usable, would not hold you back |
| 2 | Developing: a real attempt with partial signal, but a gap a coach would flag |
| 1 | Not yet shown: no usable evidence of this skill in what you attempted |
Two consequences follow, and both are worth internalising.
First, a 5 is reachable. It does not require brilliance, only doing the thing to the standard a strong hire does it. Candidates routinely leave 5s on the table because they assume the top of the scale is reserved for something extraordinary.
Second, a 1 means absence, not weakness. If you attempt the maths and get it wrong while talking through your assumptions, that is a 2, sometimes a 3 if the approach was sound. A 1 is for the skill never appearing at all. The candidate who says "I'm not great at maths, could you walk me through it" scores a 1 on rigor by never attempting it, where the candidate who tries and fumbles scores higher.
How the scores become a hire recommendation
The fourteen skill scores average into three dimension scores, and those three combine on the 50 / 25 / 25 weighting into one overall number between 1.0 and 5.0. That number maps to a recommendation with published cut-points.
| Overall | Reads as |
|---|---|
| 4.5 and above | Outstanding |
| 3.75 to 4.49 | Strong, clearly above the bar |
| 3.0 to 3.74 | Solid, keep building |
| 2.0 to 2.99 | Developing |
| Below 2.0 | Early days |
We treat 3.5 as the readiness line rather than 3.0, and the reason is worth understanding. A 3 on a single skill means that skill met the standard. But a real first round wants you comfortably above the standard across the board, because the interviewer is extrapolating from thirty minutes to three years. An overall sitting exactly at 3.0 usually means a few 4s propping up a 2, and that 2 is what gets discussed in the debrief.
The five things that lose the most points
Across the transcripts we mark, the same handful of moves account for most of the lost ground. None of them are about intelligence.
Hedging instead of recommending
"There are several factors and I'd want to run more analysis" is the classic level-one close. A partner can work with a wrong-but-reasoned view and cannot work with no view. Being decisive and wrong genuinely outscores being vague and safe.
Reciting a framework at a problem it does not fit
Framing marks whether the structure fits this case. A memorised profitability tree bolted onto a market-entry question scores worse than three home-made branches that actually address the question.
Silent maths
Rigor is assessed on an auditable chain, not just the final number. Work done silently and announced as a result is unscoreable, and if the number is wrong the interviewer cannot see where it went wrong.
Answering the adjacent question
Asked for a number, candidates often supply a risk or a strategy instead. Precision marks whether you answered what was asked. Answer it first, then volunteer the caveat.
Talking over the nudge
Interviewers plant clues, and sometimes plant deliberately wrong summaries to see whether you notice. Listening marks whether you caught them. Candidates who monologue through pushback lose this cell every time.
Two things that are not deductions, whatever you have been told
Case-prep folklore is full of rules that were never on any scorecard. Two are worth killing explicitly, because the anxiety they cause does real damage.
Asking for a number to be repeated is not a deduction. It never was. Numbers are said once, out loud, often over a video call. Asking for a figure again is what a competent colleague does, and interviewers who bother to write a note about it write it under listening as a positive, not a negative.
Taking time to think is not a deduction either. Sixty to ninety seconds of silence to build a structure is normal and expected. What loses points is filling that silence with narration of your own process. Ask for the moment, take it, then speak.
“Could I have thirty seconds to lay out how I want to approach this?”
“Sorry, could you give me that revenue figure again? I want to make sure I have it right before I use it.”
How to mark your own practice case
Self-assessment is where most solo preparation fails, because "that went well" is not a score. Use the grid instead. After a practice case, write the five analytical skills down the page and give each a number with one line of evidence for why.
- Framing: did my structure fit this case, and did I name where to start? Would a stranger reading it know what the problem was?
- Commercial judgement: did I chase the branch that mattered, or the one I was comfortable with?
- Rigor: was every calculation audible, and did I sense-check the answer against something?
- Creativity: did my ideas come from a named cut, or did I list until I ran out?
- Synthesis: did I open with the recommendation, and did I name a risk and a next step?
Two rules make this honest. Score the transcript, not the feeling, so if you cannot point to the moment, the score does not stand. And write the evidence line before the number, because a number written first is a number you will justify.
Common questions
Do McKinsey, BCG and Bain use the same case interview scoring rubric?
Not word for word, but the structure is close enough that preparing against one prepares you for all three. All three weight problem-solving most heavily, assess communication separately, and evaluate track record from behavioural answers rather than from the case. The visible differences are in format rather than criteria: McKinsey leans interviewer-led, while BCG and Bain lean candidate-led.
Is a case interview scored on a curve against other candidates?
No. Case scoring is criterion-referenced, which means you are measured against a fixed standard rather than ranked against the people interviewing on the same day. There is no quota of passes per session. Two candidates in the same round can both score above the bar, or both below it.
What score do you need to pass a first-round case interview?
As a rule of thumb, an overall around 3.75 reads as a clear hire and anything at or above 3.5 puts you in credible territory. We use 3.5 as the readiness line because a real first round wants you comfortably above the standard on every skill, not exactly at it on all of them.
Does getting the maths wrong fail the case?
Not on its own. Rigor is scored on whether the chain of reasoning is auditable and sense-checked, so an arithmetic slip you catch and correct costs very little. What does real damage is silent maths, because the interviewer cannot see the thinking, and a wrong number stated with total confidence and no sense-check.
How much does the behavioural part count?
Track record and drive carries about a quarter of the overall score in a full interview. It is assessed entirely from your stories, so it cannot be assessed at all in a case-only session, and the remaining two dimensions reweight when it is absent. Candidates consistently under-prepare it relative to that 25%.
Can I see a real scored case interview report?
Yes. We publish a full sample report showing per-skill scores, the evidence quotes behind each one, and the single highest-leverage fix, at mbbready.app/sample-report. Running your own case produces the same report on your transcript.