Market sizing case: how many petrol stations are there in the US?

Hi Anonymous,

As mentioned earlier, there are many different ways one can approach this problem. I typically approach this following way:

1. What is the total demand for gas stations?

2. What is the demand for one gas station?

3. -> Find the total number of gas stations

To start with number 1, while I have no idea about the real numbers, let's assume that population is 400M, and there is 1 car per 4 people. Thus we would have 100M cars.

Now we need to estimate the total demand for a day/week/year, let's go ahead with day. From my personal experience, I visit gas stations once every two weeks. Assuming this is the case for the entire population, 100M cars would visit gas stations every two weeks. This means that about 7M cars visit gas stations every day. -> This is the total demand.

Now we are at number 2, we need to understand the number of cars visit a gas station during one day. To keep it simple, we can assume (again, based on personal experience) that there are always 3 cars occupying the gas station for 5 minutes throughout 18 hours (and the remaining of the time, the gas station is empty). This would give us ~600 cars every day. -> This is the demand for one gas station.

We can find the number of gas stations by diving total demand (7M) by local demand (600), and reach ~12000 as our final answer. 

I have quickly checked on google, and the real answer is about x10 higher than my answer. HOWEVER, it really is not important to get a precise answer to impress the interviewer. It is all about sharing your train of thought and laying out your assumptions very explicitly. 

I hope this was helpful.

Have a good day,

Deniz Ahmet

Hi there,

Population and Car Ownership: I would assume a population of approximately 330 million in the US. To estimate the number of cars, I would assume that, on average, there is one car for every 2.5 people. This would give us roughly 132 million cars in the US.

Daily Visit Frequency: I would assume that, on average, a car visits a petrol station every three days. This assumption is based on typical driving habits and the need for refueling.

Capacity of a Petrol Station: Assuming each petrol station can serve 500 cars in a day, given the number of fuel pumps and the station's operating hours.

By multiplying the number of cars by the visit frequency and dividing by the capacity of a petrol station, I can estimate the total number of petrol stations needed. Using these assumptions, the calculation would look like this:

(132 million cars / 3 days) / 500 cars per station per day = 88,000 petrol stations

I'm glad you brought up this market sizing case. It's a classic case interview question that tests your structured thinking, quantitative skills, and ability to break down a complex problem into manageable components. Let me walk you through how I would approach it.

First, it's important to understand the context and the objective of the case. The interviewer is likely assessing your ability to size a market or estimate a figure using logical reasoning and data-driven assumptions. The key is to take a structured approach and communicate your thought process clearly.
 

Here's how I would tackle this case:

1. Clarify the question: Before diving into calculations, I would ensure that I understand the scope of the problem. In this case, I would ask for clarification on whether the question refers to the total number of petrol stations across the United States or a specific region or state.
 

2. Identify the key components: To estimate the number of petrol stations, I would break down the problem into smaller components that can be estimated individually. The main components could be:
a. Total population of the United States (or the specific region)
b. Average number of vehicles per household
c. Average fuel consumption per vehicle
d. Average capacity of a petrol station
 

3. Gather data and make assumptions: For each component, I would try to gather relevant data or make reasonable assumptions based on logical reasoning or industry knowledge. For example:
a. Total population: I could look up the latest census data or make an assumption based on general knowledge.
b. Average vehicles per household: I could use industry reports or make an assumption based on urban vs. rural demographics.
c. Average fuel consumption: I could use data from automotive associations or make an assumption based on average mileage and fuel efficiency.
d. Average capacity of a petrol station: I could research industry standards or make an assumption based on the typical size and throughput of petrol stations.
 

4. Calculate and sanity check: Once I have gathered the necessary data and made assumptions, I would perform the calculations to estimate the total number of petrol stations. This could involve dividing the total fuel demand by the average capacity of a petrol station.
 

5. Communicate and refine: Throughout the process, I would communicate my thought process, assumptions, and calculations to the interviewer. This allows them to understand my reasoning and provide feedback or additional information if needed. I would also be open to refining my assumptions or approach based on their input.
 

6.Consider additional factors: Depending on the interviewer's prompts, I might also consider additional factors that could impact the estimate, such as regional variations, urban vs. rural differences, or trends in alternative fuel vehicles.

Remember, the interviewer is not necessarily looking for a precise answer but rather assessing your problem-solving approach, quantitative skills, and ability to think logically and communicate effectively.
 

By breaking down the problem, making reasonable assumptions, and clearly communicating your thought process, you demonstrate the structured thinking and problem-solving abilities that consulting firms value highly.

There are infinite ways to size it so the following is by no means an exact answer. My preferred method would be to initially segment by locality, given that the drivers of demand for a petrol station are significantly different between rural areas and population centers.

Stations in rural areas are driven by the average range of a car when they need to refuel, so you could derive a number through the station density and area of America.

Stations in population centers are driven by the demand (i.e. the number of cars) in each city. I would actually segment this down further into Metropolises/Cities/Towns given the large difference in population size between them. Also, you would need to factor in that these stations are nowhere near running at full capacity.

I appreciate any feedback on this approach.

I would think how many petrol stations are in a specific region. For example, you know that in the city you live in there are 6 petrol stations. Let's say the city has 30 000 people, so that results to about 5000 people per petrol station. US has about 300million people, so divide 300m/5000 and you get 60 000 petrol stations.

Hi OP,

This is a great question; I was asked exactly this case in a mock interview with a BCG consultant. 

Most people have approached this from the demand-side perspective. One could also approach it from the supply side as follows:

• # of total gas stations in the US = density of gas stations (# of people served by a gas station) x # of people in the US
• For a more nuanced approach, one could segment the density of gas stations and people into the following categories: (a) urban cities, (b) towns and (c) rural areas
• I'd make a sensible assumption for each of the metrics and sanity check my answer at the end

And voila, you've cracked the case!

I like answer 1!

There are more detailed ways but I like this approach because you give a quick answer and then you can walk your interviewer how you refine these high-level assumptions.

Hi!

To add on top what´s been said, you can also do it from: 

# vehicles > # liters each of them consume > averge need for a gas station. 

You can also use 1 method and the 2nd one to proof it. 

Cheers, 

Clara