Struggle Structuring Case Math

Strong candidates don’t “do calculations.” They first translate the question into a simple equation and only then touch numbers. If you can’t express the math in words, you’re not ready to calculate.

The discipline is always the same. Start with the output and decompose it one level at a time.

If the question is about profit, say it explicitly: profit equals revenue minus cost. Revenue equals price times volume. Volume equals number of customers times usage.  This logic generalizes. Almost every case math problem can be reduced to either: total equals units times value per unit, or change equals base times driver times rate. Market sizing, pricing, growth, capacity, utilization - they all follow this pattern once you slow down and impose structure.

Use round, defensible assumptions, state them clearly, and move on. If an assumption materially affects the answer, call it out. That shows judgment.

Finally, always signal intent. “Let me structure the math first,” then walk the interviewer through the equation verbally. That’s what interviewers associate with seniority

There are techniques for structuring your approach to calculation questions. Feel free to reach out and I can share a guide on this. 

High-level:

1 Validate the question and the data 

2 Take time then come up with the logic to solve the problem and share it with the interviewer. Once the interviewer validates it...

3 Take time to do the computations, sanity check them, then share them with the interviewer. Once the interviewer confirms the correct result...

4 Interpret the result within the context of the client's mission

However, it sounds like you are particularly struggling with step number 2, as in, coming up with the logic for the calculation. 

Reach out if you want a diagnostic on this. We could run through several examples and figure out what is holding you back. 

Without being able to diagnose it, the only thing that I can encourage you to do is to practice more, and gradually, with time, you'll get better. 

Best,
Cristian

That is an excellent question, and frankly, it highlights one of the most poorly covered areas in case prep. Many candidates focus too much on memorizing high-level frameworks and too little on the mechanics of building a clear, executable equation on the fly.

Here is the truth: there are no generic mathematical frameworks like a 4P model. The structure for the math is the equation itself, and it must be built MECE (Mutually Exclusive, Collectively Exhaustive) based on the specific drivers of the target metric.

When tackling a quantitative problem, your immediate first step should not be to calculate, but to define the final unit you need (e.g., dollars, percentage, headcount). Then, build the equation backward, isolating every variable required to achieve that unit. For example, if you need total revenue ($), you must decompose it into its core driver units: $\text{Revenue} = \text{Volume} \times \text{Price}$. Then, if you don't have Volume, you decompose that driver: $\text{Volume} = (\text{Customer Base}) \times (\text{Frequency of Purchase})$. This process of iterative decomposition creates your clear, logical structure.

For handling unknowns and approximations, remember that you are demonstrating commercial judgment, not perfect arithmetic. The rule is to use clean, round numbers (e.g., 25%, 10 million, 30 days) and state the justification for the assumption before you use it. If you are sizing the market for dog food and you assume a 40% adoption rate among urban millennials, tell the interviewer, "I'm going to approximate urban millennial dog ownership at 40%, slightly higher than the national average due to space constraints." This tells the interviewer you understand the trade-offs and are controlling the variables, rather than just guessing numbers randomly.

All the best!

Hello! 

Above answers from Cristian and Kevin are amazing but I would add that asking right clarifying questions is essential to build strong framework before doing calcs! 

Examples might be: 

1. The definition of the market (e.g. "Estimate the annual revenue for a hotel in NYC" -> "Do you mean a hotel chain or a single hotel"; or "What type of hotel is it (e.g. AirBnB, hostel) 
2. The unit of measure that we should provide the answer 
3. The time period (very often we will need to calculate yearly revenue in order to get market size) 

Why this is important? As in day-to-day job, you will literally do the same exercise, usually complex market sizing models start with simple assumptions and framework that you need to present to manager. Best idea is to practice these estimation drills, happy to connect and conduct to you dedicated session!

This is a great question and a common topic for those prepping. I'd suggest the following steps / elements:

-Always review your cases that you do. How did the answer suggest you should structure a quantitative analysis? When we review this we can better learn and digest and bring this across cases. At the end of the day there are only so many types of problems that will surface on cases.
-Do math drills as part of your practice - both quick math (add/subtract/multiply/divide/fractions/decimals) as well as key formulae (profit, ROI, ROR, BEP etc.)
-When you're casing practice asking to take a short moment to structure your approach before diving into the questions. This should include your formula and all necessary data you will need (so that you can ask for it, if it isn't given already)
-For assumptions/approximations. For assumptions always try to benchmark on something (so pick a known fact that you can leverage and then check the assumption with your interviewer). For approx. rounding is always the go to - to make numbers easier to work with, but be sure to align with the interviewer that it is ok to round before proceeding.
-Communicate clearly as you calculate - this will enabler your interviewer to stop you / correct you on the fly if you made a mistake!

Hope this helps!

A lot of people can do the math but struggle with setting it up. That's actually the harder skill, and it's what interviewers care about most.

Here's how I'd think about structuring math in cases.

Start with the end in mind.

Before touching any numbers, ask yourself: what exactly am I trying to calculate? Write it down in one sentence. "I need to find the annual profit from this product line" or "I need the break-even volume." That's your target. Everything flows backwards from there.

Break it into components.

Once you know what you're solving for, ask: what do I need to get there? Most case math comes down to a few building blocks:

• Revenue = Price × Volume
• Profit = Revenue − Costs
• Costs = Fixed + Variable
• ROI = Return / Investment
• Break-even = Fixed Costs / (Price − Variable Cost per unit)
• Market size = Population × Adoption rate × Frequency × Price

You don't need to memorize dozens of formulas. These basic relationships cover most case math. The skill is knowing which ones to combine.

Map it out before you calculate.

Say it out loud: "To get X, I need A and B. I have A, so I need to figure out B. To get B, I need C and D." This creates a clear roadmap. The interviewer sees your logic before you crunch a single number. That's what they're looking for.

On assumptions:

When you don't have a number, make a reasonable assumption and state it clearly. "I'll assume average household income is around $50k. Does that seem reasonable?" Don't be afraid to round aggressively. 47 million becomes 50 million. $18.99 becomes $20. Clean numbers make clean math.

The key is being transparent. Say what you're assuming and why. If the interviewer wants a different number, they'll tell you.

A simple process you can repeat every time:

1. What am I solving for?
2. What formula or relationship gets me there?
3. What inputs do I need?
4. What do I have and what do I need to assume?
5. State the setup out loud before calculating
6. Calculate step by step, narrating as you go

On practicing this:

Don't just practice doing math. Practice setting up math. Take a case question, and before you calculate anything, write out the full structure: what you're solving, what you need, and how the pieces connect. Do this 10-15 times and the setup becomes instinctive.

The interviewers aren't testing whether you can multiply. They're testing whether you can take a messy question and build a clean path to the answer. That's the skill to train.