When automated teller machines first became popular more than 20 years ago, banks and their customers gained clear advantages. Distribution costs fell while returns on investment grew rapidly. Customers were happier too, as their money became more easily accessible.
Although the market for ATMs has matured, there are plenty of ways to increase their profitability. We recently analyzed two banks—one in the United Kingdom and one in the United States—and discovered that ATM location had a significant impact on revenue.
By using more detailed data and borrowing advanced concepts from outside the financial services industry, including game theory, banks can optimize ATM networks to boost profitability and increase competitive advantage.
Simply put, moving poorly performing ATMs to better locations creates significant value. For a U.K. bank with 2,500 ATMs, we found that improving the location of its least-used ATMs generated the equivalent of 45 extra withdrawals per day per machine, creating $19 million more in revenue per year.
Revenue attributable to ATMs is not always obvious. While foreign fees (those charged when a customer uses a different bank's ATM) are clear, "on-us" transactions (a customer using the home bank's ATM) offer advantages such as attracting new customers and keeping existing ones happy. By applying a nominal value to on-us transactions and combining them with foreign fees, we found that top-performing ATMs generate between $30,000 and $60,000 more per year than the lowest performers.
Few banks have explored the science behind ATM locations. Part of this is practical. When determining ATM locations, most banks use more general geographic and demographic divisions such as ZIP codes and municipal boundaries, which don't get down to the granular level necessary to predict traffic flows.
To help banks get more from their ATMs, we recommend a four-step approach.
First, know your objectives. ATM success is about more than maximizing profits. It requires defining and weighing competing priorities, including generating more fees, adding new customers and improving service to existing customers.
Second, identify key "catchment" locations. Demographic, geographic and commercial data, from readily available public and commercial sources, will help determine whether you should add, move or eliminate ATMs and where there is the greatest unmet need.
Many banks already use elements of these first two steps to determine ATM locations. However, the quantitative analysis often stops here, and project teams are left to scout for appropriate sites, negotiate rights and recommend a short list of locations for a distribution executive to approve.
Third, create a short list of potential sites. Identify these locations by talking with branch managers, creating focus groups of existing or target customers, and using location scouts. Then borrow from outdoor advertising techniques to gather empirical data about existing ATMs in those locations.
Finally, pinpoint sites. The information from the last step is adjusted to account for site characteristics such as access issues, and is then grouped with other sites with similar characteristics. Game theory, in which mathematical techniques are applied to real-life competitive situations, can now be used to improve decision-making. By predicting how competitors will react to different situations, you can make better business decisions.
Final site selection of ATMs requires combining game theory modeling with other local conditions that your distribution and frontline teams have discovered in their research. For example, some ATMs at low-traffic locations may generate more customers than usual because card users are accustomed to using them; these locations may require a different strategy.
ATMs may seem like old news as the market is almost completely saturated across the globe, and one might think that nothing could be done to generate new revenue from them. But our research has proven otherwise. By adding science to intuition and using lessons from other business disciplines, banks can further optimize their ATM networks.
