CaseKit
07Playbook · Quantitative

Doing the math without freezing

Most candidates know the frameworks cold. The ones who lose points aren't stumped by the analysis — they freeze on the arithmetic.The candidates who handle this well don't do it faster. They do it more visibly.

10 min read4 worked problems5 techniques
Mental Math Toolkit
~50MRound first
2×10Decompose
¼Use fractions
✓?Sanity check

Mental math in a case interview isn't about being a human calculator. It's about staying calm, working systematically, and communicating your approach so the interviewer can follow — even when the numbers are ugly.

Round aggressively. Communicate constantly. Never go silent.

Interviewers are not checking your arithmetic. They are watching how you think under pressure. A candidate who rounds and narrates beats a candidate who tries to compute exactly and goes silent for 90 seconds.

01

The 5 techniques that actually matter

1

Round to the nearest clean number — always

Simplify before you compute. This is not cheating.

$47M → $50M. 13% margin → ~15%. 2.3x growth → ~2x.

The goal is to get to the right order of magnitude. Whether the answer is $48M or $52M doesn't change the recommendation. Whether it's $5M or $500M does.

When to round more aggressively: Market sizing. You're estimating anyway — precision is false confidence.

When to round less: Breakeven calculations, margin analysis, or anywhere the interviewer has given you exact figures on purpose.

2

Break multiplications into steps

Don't try to multiply large numbers in one move. Decompose them.

$240M × 18% margin
$240M × 20% = $48M, then subtract 10% of that = $4.8M
~$43M

The technique: multiply by a round number, then adjust. Always easier than doing it straight.

6.4M × 35%
6M × 35% = $2.1M, then 0.4M × 35% ≈ $0.14M
~$2.2M
3

Use percentages as fractions

Some percentages are easier to compute as fractions.

PercentageFractionUse when
50%1/2Always
25%1/4Always
10%1/10Always
33%1/3Dividing into thirds
20%1/5Market share, margins
12.5%1/8Unit economics
$360M × 12.5% = $360M ÷ 8 = $45M
Faster than computing 12.5% directly.
4

Sanity check every answer before you say it

Before you give the number, ask: does this make sense?

If your market sizing produces a $4 trillion US coffee market, something went wrong. If your breakeven shows a 3-month recovery, double-check the inputs.

Two sanity checks that catch most errors:

Order of magnitude check: Is the answer in the right ballpark? ($40M? $400M? $4B?)

Direction check: If I change one input, does the output move the way I'd expect?

Saying “Let me just sense-check this — that implies each customer is worth $1,200 annually, which feels reasonable for a B2B SaaS product” shows judgment, not just arithmetic.

5

Flag uncertainty without undermining your answer

Name assumptions — but don't make it sound like you're guessing.

Weak: “I'm not sure about this, but maybe around $2 billion? I might be off.”
Strong: “I'm getting approximately $2 billion — that's using a 15% margin assumption which I'd want to verify, but the order of magnitude should hold.”

The second version flags the same uncertainty. It doesn't sound like you're hedging your whole analysis.

02

4 situations where candidates freeze

Learn to recognize these moments — each has a specific recovery move.

📊

The interviewer hands you a messy data table

Don't stare. Don't process everything at once.

What candidates do wrong: Stare at it, try to absorb everything at once, go quiet.
What to do instead:

1. Spend 10 seconds scanning the structure — rows, columns, units.
2. Identify the one or two numbers most relevant to your hypothesis.
3. Say out loud what you're looking for: “I'm going to focus on the COGS line here to test whether the cost increase is in production or distribution.”
4. Do the math on those numbers only. Don't process the whole table.

🔢

The numbers don't add up

You're mid-calculation and the result feels wrong.

What candidates do wrong: Try to paper over a wrong answer or hope the interviewer doesn't notice.
What to do instead:

Stop, say so, reset. “Let me step back — I don't think I set this up correctly. I want to make sure I'm calculating operating margin, not gross margin. Let me redo this.”

Catching your own error and correcting it out loud is a green flag to interviewers. It shows self-awareness and rigor.

💡

The interviewer gives you a number you weren't expecting

Your hypothesis is contradicted. Now you need new math.

What candidates do wrong: Pretend the original number still works.
What to do instead:

Acknowledge the new data, integrate it, recalculate.

“Interesting — if fixed costs are actually $80M, not $60M, then the breakeven volume goes up significantly. Let me recalculate: $80M ÷ $40 contribution margin = 2 million units. That's a 30% higher bar than I had estimated.”

🌐

You're asked to size something with no obvious starting point

Pick any reasonable anchor and be explicit about it.

What candidates do wrong: Freeze because there's no "right" answer.
What to do instead:

“I'm going to start from the US adult population of about 260 million. I'll work down from there by segment.”

Three valid approaches: Top-down (macro number, carve out your segment), Bottom-up (unit-level math scaled up), or Comparable (known market, scale it). Pick whichever feels most natural and say which one you're using.

03

Do this. Not that.

DoDon't
Round aggressively and say you're doing itTry to compute exact numbers in your head
Narrate your math as you goGo silent for more than 20 seconds
Sanity check the output before stating itState a number that's clearly impossible
Flag your assumptions and move onHedge every number with "I'm not sure"
Catch and correct your own errors out loudTry to hide a wrong answer
Ask for a piece of paper if you need itDo everything mentally when paper is available
04

Worked problems

These aren't timed drills — the full solution is one click away. Attempt each one out loud (or on paper) first, then reveal the walkthrough and compare your approach, your rounding, and your final number to the model.

1

Round and compute: $183M × 23% margin

Step 1: Round: $180M × 25% = $45M

Adjustment: We rounded revenue down slightly and the margin up. $183M × 23% is a little less than $45M. Knock off ~7% → ~$42M

Stated answer: “Approximately $42 million in operating profit.”

Actual answer: $42.09M. Your rounded estimate was within 0.2%.

2

Market sizing: How large is the US gym membership market?

Approach (bottom-up):

US adults (18–65): ~180M people
Estimate ~20% have gym memberships → 36M members
Average annual membership cost: ~$600/year ($50/month)
Market size: 36M × $600 = $21.6B → call it ~$20B

Sanity check: IBISWorld puts this at ~$35B including premium gyms and boutique fitness. Your estimate is in the right ballpark — and you'd defend it by noting you excluded boutique/premium.

3

Breakeven: $12M launch cost, $80 price, $50 COGS

Contribution margin per unit: $80 - $50 = $30

Breakeven volume: $12M ÷ $30 = 400,000 units

Sanity check: At 400K units × $80 = $32M revenue, $30 margin = $12M recovered. Checks out.

What to say: “At a $30 contribution margin per unit, the company needs to sell 400,000 units to recover its launch investment.”

4

Messy table: Revenue by region with mixed currencies

Don't: Try to convert everything and add it all up before saying anything.

Do: Identify which region the question is actually about. “Given that we're looking at where profitability is declining, I'm going to focus on the North America and Europe columns — those are the two largest revenue contributors. I'll convert Europe's €240M to roughly $260M at a ~1.08 rate and compare that to North America's $310M.”

Key move: Scope the problem before doing any arithmetic.

💡

The real test: Can you do arithmetic while also explaining your reasoning, integrating new data, and maintaining your structure? The math itself is almost always simple. The challenge is doing simple math under pressure while managing everything else at once.

Rate your mental math

After your last math-heavy case, check each habit you demonstrated.

0 / 5

Test your math skills in a real case.

The best way to get comfortable with case math is to do it live — with time pressure, new data, and structured feedback on your process.

Try a case