Automatic teller machines (ATMs) have met with mixed success since their introduction in the early 1970s. Some 165,000 are now installed in the United States, handling 900 million transactions a month. Yet few of them earn their keep.

An average-sized bank might spend $20 to $25 million a year on a network of 1,000 ATMs—a net cost of $20,000 to $25,000 for each one after revenue from fees charged to banks and customers has been deducted (Exhibit 1). This sum has, on the whole, proved to be an extra expense, since ATMs have not produced the personnel savings or branch closures that were expected. On the contrary, the total number of transactions has continued to rise, while the number of branches has increased over the past decade by a compound annual growth rate of 2 percent.

Alarmingly for the banks, the economics of ATMs could deteriorate still further as cash payments are replaced by point-of-sale debit (with and without cash-back), smart cards, and credit cards. A fall in the volume of cash withdrawals would mean a fall in ATM revenue.

There is, nevertheless, scope to raise the profitability of ATM networks. Experience shows that by reducing operating costs, improving pricing, and broadening the functions of ATMs to stimulate customer demand, banks with large networks can boost the incremental profit from each machine by $15,000 to $35,000.

Fine-tuning the network

Operating costs

A network’s operating costs can be cut by 10 to 20 percent, or $4,000 to $8,000 per ATM, if close attention is paid to the following:

Function. The more complex an ATM is, the more it costs a bank. Function should therefore be tailored to the needs of each machine’s customer base, avoiding unnecessary complexity. This often means having a mix of machines in the network, offering different levels of function. Machines installed at airports, for example, do not need deposit and bill payment features or touch-screen technology; simple cash dispensers will suffice, since withdrawals represent the overwhelming majority of transactions that take place there.

Armored-car logistics. Many banks have scope to reduce the cost of loading ATMs and collecting deposits. Often, the two activities are performed separately and according to a fixed schedule. But it is more efficient to consolidate the two tasks, and have armored cars visit each ATM only when needed. In addition, stocking an ATM with bills of large denominations helps reduce the frequency with which it needs to be loaded, because customers withdraw fewer bills.

Preventative maintenance. Banks tend to forget that a broken ATM immediately affects the bottom line. It is therefore worth investing in preventative maintenance programs to help maximize revenue.

Servicing. Banks should employ dedicated ATM servicing units in order to release more expensive branch personnel for sales activities.

Premise costs. Banks routinely pay rent to locate their ATMs off site and reach other banks’ customers. They can reduce this operating cost by entering profit-sharing partnerships with premise owners. If an ATM in a convenience store is likely to attract extra customers, for example, the store owner might be prepared to waive the rent until the ATM breaks even, and thereafter share the profits.

Redeployment. Banks should consider reducing the number of ATMs in locations where capacity exceeds demand, perhaps moving them to locations with higher traffic and therefore more profit potential.

Pricing

ATM charges vary widely according to the type of transaction and the network owner (Exhibit 2). Eight of the ten largest US banks now charge non-customers for the use of their ATMs—an attractive source of revenue that has encouraged many non-banking organizations to enter the ATM business. The possibilities do not stop there. By recognizing that there is scope to grade charges according to the location, the product, and even the time of day, banks can quickly generate an extra $5,000 to $10,000 annual revenue per machine.

The guiding principle is usually price sensitivity. In the United States and Europe, it is becoming common to charge overseas visitors for the use of ATMs, because they have few alternatives if they want to draw out cash; in Japan, pricing is adjusted at peak hours; and at locations such as casinos, where price is less likely to be an issue, the charge for a cash withdrawal might be as high as $10 (Exhibit 3). And while individuals might balk at having to pay for a printout of their bank balance, they might accept paying a fee to settle a bill. Experience suggests that while customers are initially irritated at having to pay for a service that they are used to receiving free, they quickly become accustomed to the idea as long as the fee is not excessive.

Broadening the ATM’s appeal

By broadening the functions of ATMs and thereby expanding consumer demand for them, network operators can generate incremental pre-tax profits in the range of $5,000 to $15,000 per ATM per year. Given the fixed cost structure of networks, adding extra functions is often relatively inexpensive, and incremental revenue falls directly to the bottom line. (New functions can be added easily to ATMs with newer technology.)

Costs apart, thinking up further uses for ATMs will be critical if banks are to head off the threat posed by cash substitutes. The trick is to view them as machines that can dispense information and goods, not just cash and account information.

Some banks already dispense postage stamps, tickets for sports events, and promotion coupons, and show advertisements on screen while the customer awaits authorization for a transaction. We estimate that ATMs generated more than $35 million in revenues from stamp sales alone in 1996. Wells Fargo sells stamps at 3,300 ATMs, Star Banc at each of its 450 machines, and EDS at more than 1,000. These banks typically sell an 18-stamp sheet for the cost of 20 stamps, generating a margin of 64 cents. In addition, EDS rents monthly advertising space on its ATM screens during the transaction approval and receipt-printing times, at a cost of $200 a month. Wells Fargo even sells ski-lift tickets during the winter.

These kinds of enhancement are only the beginning. Stock trades, foreign currency exchange, travel insurance sales, credit card and loan applications, or even mutual fund sales could all be carried out at an ATM. And since the machines can dispense anything that is roughly the shape of a dollar bill, what about bus passes, gift tokens, lottery tickets, or city maps? If banks were to link with appropriate partners, there is no reason why they could not issue airline tickets, rail tickets, welfare checks, or weather reports, or why customers could not send telegrams, apply to renew a car licence, or do catalog shopping on screen. Through its stake in DocuNet, a distributor of document delivery machines, EDS already plans to print airline tickets and insurance policy statements.

A bank could earn revenue from a different source for each new function it provided. It could charge the convenience store in which an ATM is located for the increased traffic its new facility will bring; the post office for the sale of stamps through the ATM; the stamp purchaser for the convenience; and advertisers for displaying an ad while the customer waits. There may also be some interchange revenue (the revenue obtained by the ATM owner from the card user’s bank for facilitating the transaction) from the sale, and the customer may take the opportunity to withdraw cash, perhaps for an additional fee.

Creating growth options

If banks were to raise the profitability of their ATM networks by improving operating costs and charging strategies, and then use the machines in new ways, the networks would no longer have to be regarded as just another costly distribution channel. Instead, they could play an integral part in their owners’ growth strategies.

Machines with a state-of-the-art range of functions will have the power to increase market share by building brand equity. A network might offer any of the general services already described, but then go on to leverage customer information to develop highly customized offers for delivery through the ATM. A bank customer from Boston who is using the bank’s ATM in San Francisco, for example, could be presented with a screen showing entertainment and dining options in San Francisco. These attractions could be in the immediate locality of the ATM, and it might dispense a coupon that the customer could use at a nearby restaurant.

By offering value-added services such as these, banks could differentiate themselves from competitors and even attract new account holders. We believe that this kind of national network is likely to be profitable even without charges for cash withdrawals, as long as it offers enough functions to draw adequate transaction volumes.

To pursue this strategy, banks might adopt a franchise model whereby they would provide the ATM, the infrastructure, and a national brand, while the franchisee, with its local knowledge, would select attractive sites, sign up strong local partners, choose the most appropriate functions for each location, and service the machines.

A second growth option would be to use ATMs to create a foothold in new geographic markets. Bank One, an emerging national retail banking player, has announced plans to build a nationwide network of 20,000 machines, mostly for cash withdrawal, in order to win business in areas where it does not have branches. It will deploy 2,000 machines at Mail Boxes Etc. stores, outlets that attract small business customers with whom Bank One hopes to build a brand. If it then chose to develop a direct banking service, the ATMs could be upgraded to accept deposits. It also has the option of renting its ATMs to smaller banks whose customers would be able to use the network free of charge. In the longer term, Bank One may have to think about introducing new functions to retain market share, but for now, a simple, cost-efficient network of ATMs is being used as a platform for growth.

Banks with networks already up and running should consider maximizing their potential by organizing them as standalone profit centers. Hitherto the machines have been regarded as a free resource for customers’ convenience, rather than as a vehicle for making profits for the banks. Giving each ATM, whether at a branch or off site, its own profit and loss account would transform the way the business is run. A machine would be kept running only if profitable, and if a branch wanted a costly feature, it would have to pay the ATM business the cost of installing and operating it. Similarly, if a bank’s marketing department insisted that an ATM be sited at a university to attract student customers, that marketing department would have to foot the bill.

By viewing ATMs as a business in this way, innovative banks will be able to lock out competitors. Banks in general should be aware that some institutions are already taking measures to do just that.