US ATM Network Analytics; Rising Fees and Growing Need for a More Optimized Network

The LTP Data Analytics team sat down again to analyze the footprint of ATMs in the U.S. against several relevant parameters. ATM networks in the U.S. (for all banks across the nation) were studied against two major factors, viz. population and GDP, to understand their correlation. The idea was simple - we wanted to know how well the network is laid out relative to where people need them.

LTP is thankful to Factual, a location data company, for providing us bank branch data from their Global Places database in order to complete this analysis. Although Factual data is typically available for commercial consideration, they graciously shared it with us (under NDA) for performing this analysis.

This analysis was performed from two different perspectives: first, at an overall state level (using GDP and population data from the US Census), and second, at a zip code level (using Income Level data from What was most interesting was that we observed a major change in the pattern between the state level correlation graphs and the zip code level ones. The following charts reveal the insights that we unearthed:

Let's start with the simpler observations. Those who are not familiar with the ATM network in the US, here you go...the following two charts show the current footprint of ATM networks in the country:

Chart 1: Distribution of ATMs by State

We mapped the number of ATMs based on the states, as highlighted by the heat-map below. The ATM density in a state is indicative of the commercial activity and population of the state. As observed from the charts below, states like California, Texas and New York show the highest numbers in terms of ATMs.

Chart 2: Distribution of ATMs in US by Zip Code

We mapped the number of ATMs based on their postal area code, as highlighted by the heat-map below. From a density perspective, the eastern region of the U.S. seems to be more populated in terms the of number of ATMs. The presence of ATMs in the postal code areas also indicates the commercial activity in areas. Areas with a more active commercial business environment witness higher numbers of ATMs in comparison to other areas.

The following charts offer further insights as we compare the number of ATMs with macro-economic factors:

Chart 1: ATMs vs Population (State Level)

We mapped the number of ATMs by state against the populations of various states. From a statistical standpoint, the following chart indicates that there is a high level of correlation (R-square value of 0.9107) between the two. This high correlation indicates that ATMs have been well laid out depending upon the population density and everything is as it should be, with no surprises.

Chart 2: ATMs vs Population (Zip Code level)

At LTP, our DNA is research and analytics. We love doing this stuff at a granular level, asking more questions and presenting more queries about the database. So, after the state level analysis, we went all the way down to the Zip Code level, and this is where things got really interesting. Typically there should have been 1 ATM per 1000 (981) people in the US on average. From a statistical standpoint, the following chart indicates that there is a low level of correlation (R-square value being only 0.5612) between the number of ATMs and the population at the Zip Code level. As such, there is la ack of correlation and that troubled us. A number of Zip Codes with lower populations have relatively larger numbers of ATMs, while the scenario is totally opposite for certain Zip Codes with higher populations where there are fewer ATMs.

It seems that the logic in place to set up the ATM networks at the zip code level is not that great.

Chart 3: ATMs vs GDP (State level)

This time we plotted the states based on number of ATMs against state GDPs. From a statistical standpoint, again, the following chart indicates that there is a high level of correlation (R-square value of 0.8878). There are larger numbers of ATMs in states with higher GDP values, which is what you would generally expect. Also, since the population and GDP both have been considered in the ATM network planning, everything seems good and logical so far. However – and you knew this was coming! – there’s more to the story, so please read ahead…

Chart 4: ATMs vs Median Income (Zip Code level)

We thought the mismatch in ATMs vs. population (at a Zip Code level) was because some banks considered economic activity to be more important. From a statistical standpoint, the following chart indicates that there is a very low level of correlation (R-square value being only 0.0115) between the number of ATMs and the median income levels at a Zip Code level. The median income is being considered as a factor here to represent the GDP, or wealth, in the zip code’s area (this is understood to be an acceptable assumption). There is quite a bit of irregularity considering the overall scenario of distribution of ATMs using this lens. Certain areas with lower median income levels have a larger number of ATMs and the scenario is opposite for certain areas with higher median income levels.

There doesn’t seem to be any real correlation between ATMs and economic activity.

Is there a need to increase the current base of ATM networks? Can we bring more logic to the ATM network footprint?

Before we go ahead and criticize the ATM network, we must give credit where credit is due. The ATM network has a better correlation with population at a Zip Code level than that of bank branches. In the U.S., around 52% of the 412,000 strong ATM base is set up by non-bank operators. Some of these non-bank ATM operators include large organizations, such as Concord EFS and Genpass. Considering this, it seems that third party players are laying out ATM networks in a more informed way based on diligence, or perhaps simply in response to market forces.

Setting up additional ATMs is not a good solution; optimization is what we need. A primary reason behind this is the rising operating expenses for ATMs, attributed to supplying cash, maintenance, processing and security.

Although Visa and Mastercard have brought down the net interchange rate according to a bank fees survey by, ATM fees have seen a significant upward shift. For use of out-of-network machines, the average fee being charged to customers has risen by 9 cents, to $1.61. For non-customers who use a bank’s ATMs, the surcharge for them has risen by 14 cents to $2.65. These costs have to be brought down at an industry level. All banks have to work together to do this. Digital and mobile is the way forward, but as of now people still need cash, and they don't want to go to a bank teller to avoid maintenance fees.

With the rising ATM fees, the ideal strategy at a bank level should be to highly optimize the ATM networks based on demographics, income data and other factors. Place the ATMs where people need them even if it means relocations. Customers should be able find and use the ATM of their issuing bank the majority of the time.

This optimization can further involve conversion of ATMs into new age branches. For example, Lake State Credit Union had installed videoconferencing technology in a number of ATMs last year. Moreover, this optimization is important as consumers are increasingly relying on more than one channel for banking. This can be highlighted by the following chart:

(Source: ClickFox)