Most "cheat sheets" are framework graveyards. Interviewers do not score whether you named Porter; they score whether you can structure, calculate, read a chart, recover when stuck, and close with a recommendation. That is what the firms publish: McKinsey frames the case as a problem-solving interview on analytical thinking; BCG says there is not always a single right answer and that structure, data, and clear reasoning matter; Bain asks you to show logical, structured thinking with precision and creativity. This page is the one-pager I wish candidates brought to preparation: stage timing, scripts, identities, exhibit protocol, recovery lines, and fit checklist — each row linked to a deeper CoachNed guide.
Print it. Drill from it. Do not try to smuggle it into the room.
What the scoring sheet actually rewards
Firms rarely publish numeric rubrics. They do publish dimensions. Deloitte assesses case interviews across multiple dimensions and spells out a five-step path: understand the issue, name assumptions, summarize findings, recommend, then outline next steps and impact (Deloitte US case tips). BCG's prep page lists the same spine: structure, useful questions, data, quick math, and what matters. Bain adds structured logic and creativity, without requiring one correct answer.
What I mark after a first round:
| Dimension | What I listen for | Common leak |
|---|---|---|
| Structure | Custom buckets tied to the ask, prioritized | Recited framework with dead branches |
| Analysis | Clean setup, units stated, sanity check | Silent arithmetic or uncorrected error |
| Judgment | Hypothesis that updates with evidence | Exploring every branch equally |
| Communication | Answer first, then proof | Narrating every calculation aloud |
| Presence | Coachable when redirected | Defending a dead path |
For the full scoring language, see the case interview scoring rubric.
Ned's rule. If your cheat sheet helps you talk longer, delete it. If it helps you decide what to do in the next 30 seconds, keep it.
Case flow with timing per stage
A typical oral case sits inside a 30–45 minute interview slot. Fit often takes the first 8–12 minutes; the case gets the rest. Timing below is Ned's working clock for a ~25-minute case block — not a firm rule. Use it as a metronome in practice so you never "run out of time" before a recommendation.
| Stage | Time | What you do | Deep dive |
|---|---|---|---|
| Clarify | 1–2 min | Restate objective, ask 2–4 scoped questions | Clarifying questions |
| Structure | 2–3 min | Write 3–4 MECE buckets, pick where to start | Structured thinking |
| Hypothesis | 30–60 sec | State a falsifiable guess that guides the path | Hypothesis-driven cases |
| Analysis / math | 8–12 min | Set up, compute, check, interpret | Case math tips |
| Exhibits | 3–5 min each | Title, axes, 2–3 takeaways, so-what | Data interpretation |
| Brainstorm | 2–3 min | Structured ideas, then prune | MECE |
| Synthesis | 60–90 sec | Answer first, 2–3 supports, risk, next step | Synthesis |
If the interviewer is driving (McKinsey-style), treat each prompt as a mini-case with its own 10-second structure. If you are driving (BCG/Bain-style), you own the clock: leave the last 90 seconds for synthesis even if a branch is unfinished.
Opening and closing scripts
Scripts are training wheels. McKinsey itself warns candidates to practice without becoming scripted; still, most first-rounders waste the opening on filler or close by summarizing the path instead of deciding. Memorize the skeleton, not the adjectives.
| Moment | Say this (skeleton) | Avoid | Deep dive |
|---|---|---|---|
| Opening restatement | "So the question is whether [Client] should [decision], measured by [metric], over [horizon]. Correct?" | Restating every fact they just gave you | Opening statement |
| Clarifiers | "I'd like to confirm [scope], [success metric], and [constraints]. Is [X] in scope?" | Asking for the full P&L before thinking | Clarifying questions |
| Structure handoff | "I'd look at A, B, and C. I'd start with A because [impact]. Does that work?" | Listing seven buckets "to be thorough" | MECE |
| Closing | "I recommend [action]. Three reasons: [1], [2], [3]. Main risk is [risk]; mitigate with [step]. Next: [owner/timeline]." | Path-summary endings with no decision | Synthesis |
Practice the closing cold: set a timer for 60 seconds and deliver a recommendation with numbers from a case you just finished. If you cannot finish in 60, your answer is not yet sharp enough.
Ten arithmetic identities
These are the identities that show up constantly. You do not need twelve "formulas" written as marketing copy; you need ten relationships you can rearrange under pressure. Full expansions live in the formulas guide and consulting math formulas.
| # | Identity | Rearrangement you will need | Deep dive |
|---|---|---|---|
| 1 | Profit = Revenue − Cost | Cost = Rev − Profit | Profitability |
| 2 | Revenue = Price × Volume | Volume = Rev / Price | Formulas |
| 3 | Cost = Fixed + Variable | Variable = Cost − Fixed | Profitability |
| 4 | Contribution = Price − Variable cost / unit | Break-even units = Fixed / Contribution | Formulas |
| 5 | Margin % = Profit / Revenue | Profit = Margin × Revenue | Formulas |
| 6 | Share = Company / Market | Company = Share × Market | Market sizing |
| 7 | Growth = (End − Start) / Start | End = Start × (1 + g) | Mental math |
| 8 | CAGR ≈ (End/Start)^(1/n) − 1 | Rule of 72: years ≈ 72 / % rate | Mental math shortcuts |
| 9 | ROI = (Gain − Cost) / Cost | Gain = Cost × (1 + ROI) | Formulas |
| 10 | Payback ≈ Investment / Annual cash | Years to recover at constant cash | Formulas |
Worked example (reproduce this)
HarborLine runs a regional ferry. Ticket price $48, variable cost $18 per passenger, fixed costs $1.2M per month, current volume 55,000 passengers per month. Management proposes cutting price to $42 and expects volume +25%.
Current profit: revenue $48 × 55,000 = $2.64M; variable $18 × 55,000 = $0.99M; profit = 2.64 − 0.99 − 1.2 = $0.45M.
After cut: volume 55,000 × 1.25 = 68,750; revenue $42 × 68,750 = $2.8875M; variable $18 × 68,750 = $1.2375M; profit = 2.8875 − 1.2375 − 1.2 = $0.45M.
Flat profit. Contribution fell from $30 to $24; the volume lift exactly offsets the contribution loss on the base (55,000 × $6 = $0.33M) with the new passengers' contribution (13,750 × $24 = $0.33M). I would not cut price unless the volume upside is larger than 25% or fixed costs fall. That is the judgment the sheet rewards — not whether you remembered "price elasticity."
Exhibit-reading protocol
When a chart lands, candidates either narrate every bar or freeze. Use a fixed four-beat protocol. BCG's public guidance emphasizes analyzing data and identifying the most important factors; Deloitte asks you to summarize specific issues and findings before recommending. The protocol below is how I train people to do both in under a minute.
| Beat | Seconds | What to say / do | Deep dive |
|---|---|---|---|
| 1. Orient | 5–10 | Title, units, axes, time range | Data interpretation |
| 2. Pattern | 10–15 | Biggest, smallest, slope, outlier | Charts drill |
| 3. So-what | 10–15 | Link pattern to the case objective | Hypothesis-driven |
| 4. Ask | 5–10 | "To test that, I'd want [data]" or move on | Clarifying questions |
Never invent a story the chart does not support. If the exhibit is noisy, say so and pick the two numbers that move the recommendation.
Recovery lines
Everyone stalls. The interviewer is watching what you do next. Bain and BCG both stress process over a perfect answer; recovery is process under stress. Keep these lines short and specific.
| Situation | Line | Then do | Deep dive |
|---|---|---|---|
| Blank on structure | "Give me 20 seconds to organize." | Write three buckets; start with the largest driver | Recovery playbook |
| Math error found | "I see the mistake — restarting from [setup]." | Recompute; sanity-check magnitude | Math tips |
| Wrong branch | "That didn't move the needle. Switching to [bucket]." | Name why the new branch matters more | Hypothesis-driven |
| Interviewer pushback | "Fair — my assumption was [X]. If [Y], then [Z]." | Update, do not defend ego | Scoring rubric |
| Out of time | "I'll close with a best-current recommendation." | Answer + top risk + one next step | Synthesis |
| Lost the objective | "Remind me — success is [metric], correct?" | Re-anchor, then prune | Opening statement |
Fit-story checklist
The case is half the interview. McKinsey's public interviewing page puts a personal experience interview before the problem-solving case and asks for detailed examples of impact, with your specific role and critical actions. BCG frames skills and case interviews as a two-way conversation about experience and problem-solving. Bain looks for curious problem-solvers who can explain why Bain specifically. Treat fit like a second cheat sheet.
| Check | Pass if… | Deep dive |
|---|---|---|
| Story bank | 4–6 stories covering leadership, conflict, failure, impact | Fit questions |
| "I" not "we" | Your actions are audible without a teammate narrating | McKinsey PEI |
| Numbers | At least one metric per story (%, $, time, people) | PEI workbook |
| Why firm | One firm-specific reason that fails a swap test | BCG fit |
| Follow-ups | You can go two layers deeper on any story | Behavioral practice |
If your case is strong and your fit is thin, you still fail. Build the bank before another weekend of frameworks.
Drill this today
Print this page. Run one 25-minute case with a timer on every stage. Force a 60-second closing even if analysis is incomplete. Then spend 15 minutes on quick math using identities 1–5. For a scored first rep without an account, use /start. For a full voice case with a seven-score debrief, use /interview. Everything on CoachNed is open for seven days, no card; then $120 for a recruiting season or $49 a month.
Frequently asked questions
Can I bring a case interview cheat sheet into the interview?
No. Notes you prepare beforehand stay outside the room. The point of this page is internalization: stage timing, scripts, and identities should be automatic so you can listen and think.
How many frameworks should I memorize for cases?
Zero verbatim. Learn building blocks — profit drivers, customer/company/competition, costs vs. willingness to pay — and assemble three to four buckets for the specific ask. Firms publish process and judgment, not a framework catalog.
What math do I actually need for case interviews?
The ten identities above cover most oral cases: profit, price-volume, contribution, margins, share, growth, and simple investment math. Accuracy and a spoken setup beat exotic formulas. Drill at /drills/math.
How long should my case recommendation be?
About 60–90 seconds. Lead with the decision, give two or three evidence-backed reasons, name the main risk and a mitigant, then one next step. Anything longer is usually a path summary, not a recommendation.
Should I prepare fit stories as carefully as cases?
Yes. McKinsey, BCG, and Bain all run a personal or skills conversation alongside the case. Four to six metric-backed stories, practiced with follow-ups, prevent the common failure mode of a strong case and a vague "we" narrative.
Sources
- Interviewing at McKinsey — problem-solving interview plus personal experience interview; analytical thinking on client scenarios. Checked 2026-09-24.
- BCG Case Interview Preparation — structure, questions, data, calculations, and reasoning over a single right answer. Checked 2026-09-24.
- Bain Interviewing — logical structured thinking, precision and creativity; case for consulting roles. Checked 2026-09-24.
- Deloitte US case and scenario interview tips — multi-dimension assessment; five-step approach from clarify through next steps. Checked 2026-09-24.
CoachNed is not affiliated with McKinsey, BCG, Bain, Deloitte, or any firm named above.