# Market sizing can come up as a standalone case or as part of a larger consulting case study

## Majority of the consulting interviews will require you to do market sizing

You can be given a market sizing question as a standalone case (though less frequent) or as a part of a broader business situation case like “entering a new market”. Market Sizing can be tricky since there is not one correct answer so be prepared beforehand. In general, the interviewer will be less interested in the number you got for the market compared to the approach you used to get to that number.

This case type is used to test your **quantitative**** and reasoning skills**. The interviewer wants to evaluate whether you are **comfortable working with numbers **and whether you are capable of making **good assumptions** and **handle ambiguity**. Market sizing questions are not only about sizing markets but also other types of estimations, for instance, estimating the number of golf balls in a Jumbo.

Ideally your result should be “in the right ballpark”, meaning the answer must **not be off by more than 20% compared to the actual answer**. But more than the result, the crucial aspects of this problem solving process are the approach you took, the structure you applied and the way you thought about the problem. If you fail to share your approach and the thought process, it is unlikely the interviewer will be able to read your mind or give you a pass for the case interview.

Some examples for market sizing cases are:

- “
*How many mattresses are sold each year in the US*?” - “
*How big is the US diapers market*?” - “
*How many gas stations can be found in Paris*?” - “
*How big will the market for 3D TVs be*?”

Although, some of these questions might seem odd, it’s something **consultants are often asked by clients** in brainstorming sessions or just out of the blue. Having a structured approach and being able to analyze what the drivers are behind such estimation questions is essential for a useful and convincing response.

**much more important**than the result itself, you should still be aware that your interviewer might deliberately give you assumptions that lead to a much higher or smaller market size then what would be reasonable. In this case make an objection saying that the market seems to be much larger/smaller than you would expect. It could be that your Interviewer wants to test your ability to

**challenge your results**(AKA

**sanity checking**).

The challenges in market sizing questions are:

- To come up with a good structure. See examples in the "Replacement Concept" videos below.
- To calculate accurately without mixing up numbers. This is especially important if you are given numbers with lots of zeros (e.g. $B and $M).
- To make assumptions that are as defensible as possible. Only use gut feeling if no other options remain.
- To be quick and accurate while making calculations.
- To be concise and not missing any steps of the solution.
- To present the solution in a client-friendly and understandable way without repeating every step of the calculation.
- To know if you are in the right ballpark by doing a sanity check at the end.

To master the many challenges of a Market Sizing Case, we strongly recommend you to always use our **three golden rules** of market sizing discussed below.

The three golden rules of market sizing will help you to structure and present your approach easily, prevent you from making calculation mistakes and check whether you are right in the end.

The three golden rules of market sizing are:

- Use a tree to structure the problem.
- Find the right tradeoff between accuracy and pragmatism.
- Sanity-check your results.

In the following sections, we will look at each of these steps in more detail:

**1. Always use an issue tree to structure your problem**

In most market sizing cases, the interviewer has no clue what the exact solution is. The final answer does not matter as much as how you got there.

Always draw a calculation tree for the solution of a market sizing case: In the first step, layout your whole structure starting from the top box (your final objective) going down to the leaves (pieces of information that are easier to estimate).

You should go to the second step only after your tree is laid out **and **has been validated by the interviewer. The second step is when you start making assumptions from the leaves at the bottom and calculating the boxes of your tree towards the top (your final objective).

For example, if the market you are trying to estimate is the number of diapers sold in the US per year, first level assumptions could look something like this:

- # of babies in the US at a given day multiplied by
- # of diapers a baby uses per day multiplied by
- # of days on average in a year

As you can expect, the tree can have a **different number of levels for each of its branches**. While you will be able to estimate the average number of days in a year fairly accurately, estimating the number of babies in the US might require you to break down this box further.

The following illustration shows how we can draw a tree to estimate how many gas stations can be found in Paris. You can find the entire case here.

Splitting the process in two steps, first structure then calculation, simplifies the problem a lot. Why is that? Because your brain has two separate hemispheres: the right (creative) and the left (rational). While the right side is good at conceptually putting together all the factors that are needed for the estimation, the left side is good at the number crunching.

If you mix up structure and calculations, you will fail to focus on one of the hemispheres at a time and will not use your brain to its full potential. Therefore, first establish and validate a tree structure and only afterwards concentrate on crunching the numbers.

**2. Find the trade-off between pragmatism and accuracy**

Rounding numbers in a market sizing case is inevitable unless you are a genius or the data is very simple. But do not round them without asking the interviewer if he is fine with it. He might want to assess your accuracy when calculating something.

Always keep an eye on the right balance:

- Do not round more than 10%.
- If possible, round some numbers up and some down in order to cancel out your rounding errors.

Make use of averages if you do not have much time left or if the problem is already too complicated. Segment the data otherwise. For example, we took an average of 5 pumps for the Paris gas stations case above. Another option would be to segment gas stations into big, medium and small gas stations in a further tree level.

In case you decide to segment the data, try to come up with 3 segments, as three is the number that people like most. You'll find it in everyday life: First, business, economy class; children, adult, elderly; etc. Also, never forget to be MECE (mutually exclusive and collectively exhaustive) in your segmentation.

*ow deep do I have to go for each of the branches?*'. As a rule of thumb, you need to go as deep as you need to be able to make only

**defensible assumptions.**

**3. Try to do sanity checks along the way and definitely at the end of the case**

You might make mistakes throughout a market sizing case. However, with sanity checks, you may be able to detect them in time. For example, when you have to assess the yearly revenues of a company, compare them to the revenues of a competitor with similar size.

Know your facts! This will help you evaluate the correctness of your numbers. If, for instance, your calculation results in 350 million women wearing red pullovers in the USA, you will know that this is impossible because the population of the US is ~300 million.

Check out the 3 golden rules again in the video with new examples:

## The replacement concept will help you to estimate numbers that are hard to guesstimate

For some case problems, guesstimating the right numbers is even more complicated if you follow the reasoning induced by the interviewer. If, for example, you had to estimate the amount of babies being born per year, you would likely get into trouble when you reach the stage where you need to define the number of babies being born per woman per year. To avoid these pitfalls, think instead of the big picture. Think about how many years it takes for the population to be replaced. Watch the following videos in order to see the replacement concept applied in practice.

In this first video, PrepLounge's Co-Founder and Expert Fernando demonstrates how to use the replacement concept on the market-sizing case "How many babies are being born per day in the world".

For further practice, Fernando uses the replacement concept on the market-sizing case "How many cars are sold in the US per day".

Finally, see how Fernando applies the replacement concept on a tricky market-sizing case "How many incandescent light bulbs are sold in Germany per year".

## Market Sizing – Key takeaways:

- Quantitative and reasoning skills are the key to success. Train them!
- Make sure to ask the interviewer if you can round the numerical answers.
- Estimations off by 20% are good enough.
- When estimating markets of products that need to be replaced on a regular basis (refrigerators, laptops, cars, light bulbs etc.), use the replacement concept.
- Always use the three golden rules of market sizing:

1. Always use a tree to structure your problem.

2. Find the trade-off between pragmatism and accuracy.

3. Do sanity checks along the way.

@florian @karim- If you read the Case in Point book by David Consentio, that is how they approach solving this problem.

na it is correct unit wise. you actually divide (10m) people by (15k) people per gas station which gives you gas station as a unit

@em I see what you're trying to do here, but from an engineering perspective, you're just dividing people by people (if you look at the units), meaning you won't achieve an answer with "gas stations" as a unit.

That is the way I'd think of it tbh. Good luck!

Why can't you do the following approach for the gas station question?

Population: 10M people

There is 1 gas stations for every 15,000 people (found this in "Interview Math" book)

10M / 15k peole = 666 gas stations

This is a little more than the 20% standard of error. It's not as detailed, but would this approach be wrong to do in interviews?

Hello! I'm thinking of how one would do the market sizing for the following two questions:

1. Annual revenue of an average bike shop in London

2. Annual revenue of an optician (think Specsavers if UK, and Vision Express if US) in a busy area (think Oxford Circus for London, Fifth Avenue for NYC)

For both i have segmented as spend * number of customers annually. For spend I think I have a firm understanding how how I would segment and weight revenue streams, but for customers annually I would like to hear how others would approach such a question!

Thank you!

How do we estimated the market size for an amusement park in a country?

Hello everyone. About a year ago, before looking into cases, I got this interview question:

use the

top-downthenbottom-up approachesto estimate the total number of drills sold to private households in London every year.I did answer the top-down correctly, but couldn't find an alternative for the bottom-up. If any

differences at allin structuring the solution, what could it be? Thanks!How did he arrive at volume of golf ball as "diameter to the cubed"?

Volume of a sphere is 4/3 * pi * r * r *r. That certainly does not come close to his approximation?

How can we estimate the number of retail stores of a give software in an area?

@Jo Erika

No, you would be ignoring that the population growth rate is a net figure! So (Humans born - humans dead)/(humans alive in t-1) is the population growth rate.

I would think that we should also consider lightbulbs further in - commercial use (e.g. offices, small and big business) and public use (hospitals, schools etc).

Hi sam for empire state, I follow this approach

First, I try to figure out the perimeter of Empire State by multiplying the width and length of the area.

Second, you can find the perimeter of each brick.

So, you can assume around 80% of empire state is built from bricks ( minus glasses and other parts)

In the end, you divide the perimeter of empire state to perimeter of each brick, results to number of bricks.

For number of babies born in a year: if you know the population growth rate, wouldn't it be simpler to just multiply the total population and the growth rate?

I like the aspect of using a top-down issue tree, but I was wondering if it can also be turned "upside down". The reason I'm asking is because I've been using the book of Consentino - "Case in Point", which has the following step by step approach:

1. Total population of country or region

2. Choose type of estimation question (Population, Household, or Individual)

3. Divide by for example individual, and then segment further into groups (by region - urban/rural, etc) (and maybe gender)

4. Choose ownership percentage per segmented group

5. Choose replacement speed (how many one uses per given time).

So it starts broad and then narrows down until you have the estimated figure.

(this helps if you have many segmented groups)

How would you visualize this in the form of an issue tree?

The reason I'm asking is because what I miss in your approach is that if you need to segment in several different age groups (and associated ownership), one easily gets stuck.

Your approach is definitely VERY useful when dealing with non-traditional estimation questions like golfballs in planes, bricks in a building, moving a mountain etc, but I'm not sure which approach is best for dealing with groups of people. (So I'm looking for a way to use your approach and match it with that of Consentino's)

Hi. I was recently asked a question in an interview that left me rather stumped - How many bricks are there in the Empire State Building? Any help on how to approach this question would be appreciated. Cheers.

for the gas station example...would a bottom up approach based on total land area of Paris, split into approx sized suburbs, downtown etc filtered by population density and therefore no of gas stations calculated based on servicing a certain radius area work?

Hi Chaki, other reasons as you mentioned below probably do not form as big of a percentage compared to lifetime "ending". Since its an estimation question, you can ignore that and certainly mention why your numbers would be slightly off (due to assumptions etc,). Let us know if you have more questions.

Hi Clement,

Thanks for asking the question. Really good point and keep in mind in many market sizing questions, schools/hospitals/companies will be considered "commercial" use. You can clarify from the interviewer whether we are talking about "private use" or "commercial" or both. Most likely, the interviewer is interested in seeing your approach and the fact you know that "commercial" also needs to be taken into account, he/she will be impressed. I hope this answers your question!

What about the number of lightbulbs sold to companies ? Schools ? Hospitals ? Why aren't they taken into account ?

For the cars sold problem, is it relevant to state that cars are only sold due to lifetime 'ending' and not due to any other reason e.g financial problems/style trends?