Consulting interview scorecard

The consulting interview scorecard, dimension by dimension.

McKinsey, BCG and Bain do not publish their grading sheets. This scorecard is CasingLab's rubric, built by former MBB consultants from what interviewers actually evaluate and from what candidates consistently report. It is the same rubric CasingLab scores you against after every AI case.

Written by Fouzia A., ex-McKinsey, founder of CasingLab, who wrote the rubric.

For the story of how interviewers use a scorecard like this, read the consulting interview scorecard article. This page is the reference: print it and keep it next to your practice cases.

The full scorecard

Bands follow the 1-5 scale candidates most often describe firms using; a 2 or a 4 sits between the anchors.

Weights are CasingLab's weighting of the six dimensions, not any firm's official numbers.
Dimension	What a 1 looks like	What a 3 looks like	What a 5 looks like	Observable behaviours
Structure 20%	Recites a memorised framework. Buckets overlap or stay generic, no hypothesis, dives into analysis before laying out the approach.	A workable MECE breakdown, mostly case-specific, but no prioritisation and no explicit hypothesis. Needs a nudge to pick a starting branch.	A MECE, case-specific tree in the first 60-90 seconds, with a prioritised first branch, a stated reason for the priority and a falsifiable hypothesis.	Restates the core question in one sentence Names 3-4 case-specific buckets before analysing Picks a branch and says why States a hypothesis that data could disprove
Quantitative 20%	Jumps into arithmetic without a setup, loses track mid-calculation, presents the number without saying what it means.	Correct setup and result with narration, but rounding is unstated and the interpretation needs prompting.	Writes the equation first, states rounding choices, estimates before computing, narrates cleanly and closes with the business implication of the number.	Lays out the equation before calculating States rounding and assumptions out loud Sanity-checks the result against an estimate Translates the number into an implication
Business judgment 20%	Chases low-impact branches, asks for data that could not change the answer, conclusions fail a commercial smell test.	Sound instincts on the main drivers, but prioritisation needs prompting and assumptions go unchallenged.	Investigates the highest-leverage branch first, asks for exactly the data that could change the recommendation, and challenges assumptions without derailing the case.	Starts with the branch that moves the answer most Requests data that could flip the recommendation Questions a doubtful assumption at the right moment Connects findings to operational reality
Communication 15%	Bottom-up rambling, heavy hedging, long monologues with no check-ins, the listener has to assemble the point.	Clear and mostly structured, but signposting is inconsistent and some answers run long before landing.	Top-down throughout: answer first, then numbered support, pausing for acknowledgment, concise under pressure.	Leads with the answer, then the reasons Signposts: 'three points; first...' Stays under about 60 seconds per answer Avoids hedging language
Creativity 15%	A short list of obvious ideas, all from the same angle, dries up quickly.	A solid list across a couple of categories, but the structure emerges late and the non-obvious ideas need prompting.	A structured scaffold stated first, then a broad, diverse set of ideas across distinct categories, including several non-obvious ones.	Names the categories before listing ideas Covers several distinct angles, not one Goes past the first obvious ideas Keeps ideas relevant to the client's constraints
Synthesis 10%	Recaps the case chronologically, no recommendation or one that appears at the end, no risk, no next step.	Recommendation-first but thin support, the risk or the next step is missing.	In about 60 seconds: the recommendation first, supporting reasons with numbers from the case, one risk, one next step, delivered with conviction.	Opens with the recommendation, not the story Backs it with numbers from the case Names one risk honestly Closes with a concrete next step

How the emphasis shifts by firm

Candidate accounts describe the same six dimensions everywhere, with the emphasis shifting from firm to firm: McKinsey interviews are often described as leaning on structure and quantitative rigour, BCG on creativity, and Bain on business judgment and synthesis. Formats and emphasis vary by office and recruiting cycle, and candidate reports are not official firm policy. For the full picture of each process, see the McKinsey interview guide, the BCG interview process guide and the Bain interview guide.

Before the interviews, several firms screen candidates with online assessments that use their own scoring, separate from this scorecard. If your invitation mentions one, the practice pages for the McKinsey Solve game, the BCG Casey chatbot interview and the Bain SOVA assessment cover what to expect.

Self-assessment checklist

Run through this immediately after every practice case, on screen or on paper. No account is needed. Tick what you actually did, not what you meant to do; a dimension where you tick fewer than half the boxes is your prep focus for the week.

StructureRestates the core question in one sentenceNames 3-4 case-specific buckets before analysingPicks a branch and says whyStates a hypothesis that data could disprove

QuantitativeLays out the equation before calculatingStates rounding and assumptions out loudSanity-checks the result against an estimateTranslates the number into an implication

Business judgmentStarts with the branch that moves the answer mostRequests data that could flip the recommendationQuestions a doubtful assumption at the right momentConnects findings to operational reality

CommunicationLeads with the answer, then the reasonsSignposts: 'three points; first...'Stays under about 60 seconds per answerAvoids hedging language

CreativityNames the categories before listing ideasCovers several distinct angles, not oneGoes past the first obvious ideasKeeps ideas relevant to the client's constraints

SynthesisOpens with the recommendation, not the storyBacks it with numbers from the caseNames one risk honestlyCloses with a concrete next step

How CasingLab scores the same six dimensions

Self-scoring builds the habit, but it tends to be generous, and communication in particular is hard to judge from the inside. In a CasingLab session, an AI interviewer runs the case out loud and then grades these same six dimensions from your transcript, on a 1-10 scale, with the evidence quoted back to you: you see which sentence earned or cost the score.

The two work together. Use this printed scorecard after peer practice, and use the scored sessions to check your self-assessment against an interviewer that applies the rubric the same way every time.

See your own scorecard, free

Scorecard questions, answered

Do McKinsey, BCG and Bain publish their scorecards?

No. The firms keep their grading sheets internal, which is why this page presents CasingLab's rubric, as explained at the top. The six dimensions are the ones interviewer and candidate accounts consistently describe being evaluated.

Do all firms weight the dimensions the same way?

The dimensions are similar across firms but the emphasis shifts, and the weights on this page are CasingLab's weighting, not any firm's official numbers. Formats and emphasis vary by office and recruiting cycle, and candidate reports are not official firm policy.

What score do I need to get an offer?

There is no published passing bar. What accounts consistently describe is that one very low dimension usually sinks the evaluation regardless of the others, because interviewers ask themselves whether they would put you in front of a client. Lifting your weakest dimension matters more than perfecting your strongest.

How should I use this scorecard after a practice case?

Immediately after each case, go through the checklist on this page and band yourself 1 to 5 on each dimension, honestly. Any dimension below a 3 is your prep focus for the coming week. Repeat after every case so you can see the trend, not one data point.

Is self-scoring reliable?

It is a good habit but it tends to be generous, especially on communication, which is hard to hear from the inside. A second opinion helps: a practice partner using this same rubric, or an AI-scored interview that grades the six dimensions from your transcript.

Can I print this scorecard?

Yes. The page has a print layout: use your browser's print function and the rubric table and the self-assessment checklist come out as a clean paper reference, with the navigation and buttons removed. No account is needed.