How many are there? - Which method is better?

Question

I came upon a question regarding market sizing:

Whenever there is a question like "how many schoolbuses/ commercial airplaines etc. are there in the USA?

I saw two ways of doing it:

a) easier way: the amount of students in the USA taking the school bus divided by how many students fit in a bus

b) harder way: how many rides are demanded a day (or month or year) divided by how much rides one bus offers a day (or month or year)

Preplounge generally suggests using the later one (demand = supply). However, often this takes too much time during a bigger case.

My question: where is the logically difference between the two?

Same for this question: how many airplaines are there in the US?

b) passengers/year divided by how many passengers one plane transports a year

a) Could we say: how many people are there in the US who use commercial airplaines divided by how many people fit in an airplaine (assuming 100 occupation) ?? Here we would have to assume the full capacity in the airplaine is used?

Thanks a lot,

Best

Answer 1

Hi Anonymous,

you should always use the second method (b). The first method (a) is not as accurate for two reasons:

it neglects the frequency people have using a vehicle, as Vlad mentioned
it assumes that a single transportation vehicle can be used only once in a specific period of time

If you find how many people in the US use commercial planes, for example, you are not considering that they may use a commercial plane more than once in your selected period of time. Also, you are not considering that the airplanes can be used more than once.

Let's see that with an example.

Let's say that in a generic country you have 100M people taking a flight every year. For simplicity, let's say airplanes just go from point A to point B and the other way round. Let's say that there are 200 people per airplane as possible occupancy and that you have a 75% occupancy, thus 150 people per airplane on average.

Application of method a

Applying 100M/150, you find 670k airplanes approximately

Application of method b

For this method, you need to know

Frequency of usage - let's assume 4 flights per year per person in this particular country
How many times the airplane can be used again by someone else - let's say the airplane goes from A to B to A in 24h

Let's assume for simplicity 50% of the trips are from A to B and 50% from B to A.

The yearly number of trips is equal to 100M*4=400M, let's assume equally distributed during the year

This means that daily you have 400M/365 = 1.1M trips, of which 550k from A to B.

Thus, the number of airplanes needed in this country is 550k/150= 3.7k

As you can see, using the correct approach b, which takes into account frequency and multiple usage of the vehicles, you get a completely different number compared to approach a.

Hope this helps,

Francesco

Answer 2

Hi,

As a general comment, keep in mind, multiple approaches are possible to size a market (top down vs. bottom up, demand vs. offer). Your goal is to find the most appropriate one to the current situation, meaning the one that will requires the less sensitive hypothesis, and that will be likely to be more accurate

Bus question

The second is simply a more advenced version of the first, bit basically in both cases you'll consider demand as a starting point. So depending on how much time is dedicated to market sizing you should oriente your structure in more / less complex direction

Airplanes question

You approach could be used, but of course would require an assumption on the plane occupation (which can't be 100%). Again you're building your sizing from the demand side

Best

Benjamin

Answer 3

Hi,

- how many people are there in the US who use commercial airplaines divided by how many people fit in an airplaine (assuming 100 occupation) ?? - you are missing the frequency in that approach

- the amount of students in the USA taking the school bus divided by how many students fit in a bus - again missing the frequency

Best

Answer 4

The second option requires more assumptions (for example, what's the % of students needing a bus ride, which itself is a fraction of the number of students that you estimated anyone in the 1st option).

It comes down to how intricate the market sizing exercise needs to be. Are we talking something high level, 2 minutes back-of-the-envelope exercise, you can do in your head type? Or are we looking at a significant chunk of the case which can take a full 10 or 15 minutes on its own.

In my experience, MBB is more likely to ask about the 1st type, but that's not an absolute. You may also look to your interviewer for guidance if needed.