Schedule mock interviews on the Meeting Board, join the latest community discussions in our Consulting Q&A and find like-minded Interview Partners to connect and practice with!
Back to overview

Estimate the ATM market size in the U.S. 

Not sure at all how to approach this question. 

I know the number of ATMs are decreasing as more people switch to digital forms of payments. I could start with a top down approach. Like 320 million people in the U.S. —> minimum age to open a bank account without parent permission is 18. We should consider maybe 10% of that demographic has a bank account. I read somewhere that 20% of Americans either don’t have a bank or are underbanked. 

So we could see that —> 

14.4 million under 18 minus 1.4 million (who have bank accounts)

18+ we estimate that 80% have a bank account —> 248*0.8= 199 million have bank accounts 

So approx 200 million have bank accounts. 

We then need to figure out how many bank accounts are needed for ATM but this approach doesn’t fully convince me. 

Any thoughts? 

Thank you!

Answering the user question: 

Are we talking the installed base of ATMs, the yearly market revenues of all manufacturers of ATMs, the yearly withdrawal fees generated by ATM owners or total money withdrawn/inserted?

We are talking about the he yearly withdrawal fees generated by ATM owners. 

Although calculating the number of ATMs in a country is also pretty interesting 

3
2.4k
20
Be the first to answer!
Nobody has responded to this question yet.
Top answer
Anonymous B
edited on Jul 10, 2021

First of all, could you clarify what you mean with the ATM market size?

Are we talking the installed base of ATMs, the yearly market revenues of all manufacturers of ATMs, the yearly withdrawal fees generated by ATM owners or total money withdrawn/inserted?

The first two would probably have a similar approach. The last two would start from quite the different approach. Either way, estimating the total amount of people with bank accounts definately makes sense and your number seems sensible enough (probably on the low side).

I would think about the following two questions:

What determines how often people use ATMs?

What determines how many ATMs are needed? Availability/proximity or capacity?

Given that you are focused on operator revenue. The following approach should give a fairly robust estimate:

DOMESTIC

Number of bank account holders

Average number of withdrawals made per account holder

Average fee per withdrawal

FOREIGN

Total number of foreign visitors to thr US

Average number of withdrawals made per visit

Average fee per withdrawal

What drives the average number of withdrawals? Are there segments that differ in behavior that makes it easier to estimste?

What is the typical fee structure?

How high are fees?

Does it matter how much money is withdrawn?

Does it matter if the fee is paid by the withdrawer or their bank?

Does it matter if withdrawals are made on or off a network owned by the card issuer?

Using a demand side approach like this makes sense since the supply side estimation requires estimating the utilization of an average ATM which is probably low (hence easy to get wrong) and total number of ATMs likely quite uncertain. The demand side parameters such as number of bank account holders will be easy to estimate accurately like you did, average frequency and fee you should be able to use your everyday experience for.

14
Anonymous A
on Jul 07, 2021
just answered your question
Anonymous
on Jul 07, 2021
Allright! Added my suggested overall approach and a few more questions to consider.
Ian
Coach
on Jul 07, 2021
Top US BCG / MBB Coach - 5,000 sessions |Tech, Platinion, Big 4 | 9/9 personal interviews passed | 95% candidate success

Hi there,

You've gone about this wrong.

If you're really going to do top-down, you should think about how frequently individuals need to access an ATM (no need to measure bank accounts. This is a waste of time). Then, think about how many are needed to meet demand.

That said, I'd prefer a bit more of a bottom-up approach where you think about how many ATMs you see near you, then extrapolate that out to the cities, towns, etc. of the US.

Anonymous B
on Jul 07, 2021
He is not talking about number of bank accounts but people with bank accounts. I think it fair to assume that people without bank accounts do not use ATMs. Would love a sensitivity analysis on the assumptions you need to make to do the bottom up approach to get to a more robust number
Anonymous C
on Sep 11, 2022

Here is a good way to solve the problem:

> Segregate the US market in sectors:

1. Urban population → 70% of total US population → 80% have bank accounts → 70% of banked population has debit card → 80% of cardholders use it frequently (estimate 3 transactions/month because usually banks allow 3 free transactions/month) → Divide by 30 (ignoring seasonality - usually more transactions would be made during the weekend) ==> This gives you the total number of transactions in a day.

Average transaction value = $20 (assumption) → most living expenses are done through POS machines/online purchases etc. $20/transaction is for small expenses.

Now using the approach of ATM capacity. 4 cassettes in each ATM - each can store 500 notes. Assuming all $10 notes are stored in the ATM → 4 x 500 x 10 = $20000/ATM

usually cash replenishment is done once in 2 days. So an ATM needs to have enough cash for 2 days.

==> Number of ATMs needed = (# of daily transactions x 2 x avg transaction amount ($20) ) / ATM capacity, i.e. $20,000

Now assume that in urban areas, there are usually 20% more ATMs than required.   

 # of ATMs required 

2. Rural population - Use the same approach with realistic assumptions.

30% of total US population is rural → 50% have bank accounts → 60% of banked population has debit card → 50% of cardholders use it (estimate 1 transactions/month for rual users as ATMs is not their first preference to get cash - daily wage earners would get cash in hand everyday, etc.) → Divide by 30 to get average daily transactions

Use the ATM capacity approach to solve for number of ATMs required. 

I would assume cash replenishment frequency in rural areas to be lower - once a week - as it is more expensive for banks to reload cash in far flung areas.

Use this approach to calculate # of ATMs in rural areas.

Hope this helps! 

2
Similar Questions
Consulting
Just did the Mckinsey Solve Game (January 2025) - got some questions/insights
on Apr 24, 2025
Global
5
3.3k
Top answer by
Hagen
Coach
#1 recommended coach | >95% success rate | 8+ years consulting, 8+ years coaching and 7+ years interviewing experience
38
5 Answers
3.3k Views
+2
Consulting
Employment Gap on Resume and How to talk about it during Interview
on Apr 14, 2025
Global
9
8.0k
Top answer by
Ariadna
Coach
BCG | Project Leader and Experienced Interviewer | MBA at London Business School
110
9 Answers
8.0k Views
+6
Consulting
How should I explain a change in course at university? Will it be asked of me?
on Apr 14, 2025
Global
10
3.5k
Top answer by
Alessa
Coach
xMcKinsey & Company | xBCG | +200 individual & group coachings | feel free to schedule a 15 min intro call for free
81
10 Answers
3.5k Views
+7
How likely are you to recommend us to a friend or fellow student?
0 = Not likely
10 = Very likely
Thanks for your feedback! Your opinion helps us make PrepLounge even better.