As more and more customers of financial-services companies turn to low-cost virtual-distribution channels, the higher-cost physical channels—traditional bank branches, in particular—will no doubt have a harder time earning their keep. Yet rumors of the death of bank outlets are exaggerated (Exhibit 1). Indeed, banks should find them a source of substantial value for years to come if they are carefully contoured to fit the rest of the distribution system and local market opportunities.¹

For bank outlets offer customers something that the Internet can never match: a secure physical location for transacting complex financial business with real people. In fact, their use has recently been increasing—from 54 transactions per US household in 1993 to 62 in 1998. More than 80 percent of consumers visit a physical outlet at least once a month, and bank outlets still generate 80 to 90 percent of new deposit, investment, and loan accounts.

Thus it should hardly be surprising that consumers prefer financial institutions offering services both on the Internet and in physical outlets to institutions that offer them only on-line. Although 40 percent of on-line customers say they would consider opening an account with an on-line-only banking institution, some 70 percent say they would open an account with a bank that had physical outlets as well. Similarly, only 28 percent of brokerage customers say they would open a pure on-line brokerage account; 42 percent say that they would open an on-line account if the broker also had physical locations.

Bank executives too favor physical outlets. In a series of recent interviews, several top executives of leading US banks identified physical channels as the most defensible source of competitive advantage over attackers. Executives confirm that while the Internet and phone channels are good for meeting the service needs of existing customers (by, for example, providing their account balances), physical outlets are better at bringing in new business. Charles Schwab, the pioneer on-line broker, reports that 70 percent of its new accounts are opened in its branches. It plans to increase its branch network, now approximately 350 units strong, by 15 to 25 percent a year over the next several years. Many conventional banks find that when they shut down a physical outlet, a new-economy competitor like Schwab moves into the space.

But physical networks definitely need careful tending to flourish. Overall, they account for 50 percent of the cost base of a typical retail bank. During the past few years, banks have been busy developing lower-cost channels: automatic-teller machines (ATMs), telephone centers, and the Internet. Paradoxically, these often increase total distribution costs. If banks want to raise the productivity of their total distribution systems, they must cultivate their branch networks too—pruning in some places, planting in others.

It isn’t hard to find and close the worst-performing 10 to 20 percent of a bank’s outlets. Although getting the rest of the network into shape without losing customers can prove much harder, it is worth the effort. Our experience suggests that optimizing the physical network can improve a retail bank’s cost-to-income ratio by 5 to 8 percent.

Principles of distribution

Many bank executives would agree that optimizing a retail distribution system involves providing customers, at a minimum, with comparable convenience at a lower cost. Distribution systems today comprise not only physical channels but also remote, or virtual, ones. Banks need to understand how their physical outlets fit with these other channels. Before making ground-level decisions about specific outlets and customers, banks must develop broad guiding principles for serving their target segments.

Banks often start work on their distribution strategies by developing value propositions for many microsegments. But since the challenge of developing channel variations to serve all of them would be paralyzing, most banks should stick to confirming value propositions for only four or so segments—say, small-business, mass-market, affluent, and rural customers. That would give banks plenty of traction in designing their networks, and they could save microsegmentation for tactical marketing programs (such as targeted cross-selling), since programs of this sort can tailor value propositions for microsegments more economically than a dedicated channel can.

Within the distribution network as a whole, a channel that fulfills an important function for one segment may be less useful to another. Moreover, different channels may play different strategic roles, so a bank must also determine the functional and strategic roles that each channel should play (Exhibit 2). If branches are greatly esteemed by high-value small-business customers and play a critical role in distinguishing a bank from its competitors, for example, those branches have a strategic role and thus warrant attention and investment. An ATM network, by contrast, might play a "parity" role: that of keeping a bank even with its competitors rather than distinguishing it from them. In parity channels, a bank should aim to deliver services that match those of its competitors at the lowest possible cost.

With an understanding of each channel’s role, banks can identify the ideal location, format, and staffing pattern of an outlet serving a particular segment—and thus what to aim for. But in most neighborhoods or micromarkets, banks have a "one size fits all" legacy. Usually, it isn’t worth revamping their outlets, since major format changes rarely overcome the target customers’ reluctance to switch.²

A new geographic market or "virtual" channel such as the Internet, however, does permit banks to target discrete segments. In a new suburb, for example, one bank has designed branches with the feel of an upscale café to attract the affluent young.

A single channel may serve two or more target segments successfully. But doing so mustn’t compromise the channel’s strategic role in a particular segment; service levels must be consistent across segments, and banks must be able to cut costs or boost revenue by combining them. Channels used mainly for transactions—ATMs, mini-branches, and teleservice channels—are more likely to surmount all these hurdles.

Applying the theory to micromarkets

A few banks still try to optimize physical outlets one by one, but most understand that it is the convenience of the network to the places where customers live and work that matters to them when they choose a bank. The bank with the most convenient network within a particular micromarket typically wins market share greater than the sheer number of physical outlets would explain. But even banks that manage their networks as micromarkets rather than individual outlets often make the mistake of managing each micromarket as a simple, closed system of physical resources that supply all the needs of local customers. Many banks also treat all segments and types of interaction as equally valuable, which has the effect of making networks fairly uniform across micromarkets.

Yet today, an increasing number of customers use remote or mobile channels for transactions. As a result, physical channels are adding value less by effecting transactions than by generating new accounts. Moreover, customers from some segments create more value than customers from others. To guide network investment and cost reduction priorities, banks need a strategy that takes such factors into account for each micromarket.

A strategic-direction matrix is a powerful tool for developing micromarket strategies (Exhibit 3). It works by comparing a forecast of the bank’s performance in a micromarket under a business-as-usual strategy with the micromarket’s growth prospects. If, say, the bank’s business-as-usual performance were only average in a micromarket that had a dim future, the right strategy would be, "milk hard." But if the micromarket were set to grow, "invest to get better" would be more appropriate. Banks that apply these differentiated strategies across their micromarkets have realized improvements of 200 basis points or more in their overall cost-to-income ratios.

Each micromarket’s place on the matrix determines how the capacity of the physical network serving the micromarket should be adjusted. In a market that must shrink, decisions about which branches to close depend on how customers use the network, the value of different outlets, and the alternative outlets available. Banks may wish they could exit some micromarkets. But the results of recent experiments with alternative forms of outlet ownership suggest that banks could keep a profitable presence even in some of the most unpromising places (see sidebar, "Letting go").

Prune the branch, not the customers

Say that Trustbank must close one of its five outlets in Middletown. The bank measures its target segments’ use of these outlets and weights the results by the value at risk should any one of them close—that is, the value rather than the absolute number of customers. By this reckoning, the outlet with the smallest number of customers at risk isn’t necessarily the one to go.

After deciding which branch to close, Trustbank has to migrate its valued customers to alternatives. Many banks find it hard to shut a branch and keep its customers. But banks that adopt a comprehensive program for migrating them and shutting down capacity within micromarkets (Exhibit 4) can realize an improvement of 100 to 300 basis points in overall cost-to-income ratios. The trick is to find and move the heavy transactors rather than to cut out a class of transaction. If all of Trustbank’s customers at the branch slated for closure switch to ATMs to withdraw cash but still make deposits there, Trustbank can’t close it. Banks must also take extra care to migrate change-resistant customers—elderly people, for instance—perhaps by moving a staff member popular with that group to the alternative outlet. With thoughtful migration levers like this, banks can actually retain 99 percent of their valuable customers from a closed branch.

The pricing of alternative channels is, however, crucial because costs can increase if customers start using them much more than the old branch, even if the cost per use is much lower. But while price structures in alternative channels must reflect their anticipated use and the cost to the bank of providing them, the price shouldn’t discourage migration.

Taking these principles into account, a bank can begin to adjust its network to reflect its changing business environment. Rolling out such a program across an entire physical network might take one or two years. But banks can’t afford to stop there: retail banking changes so much that perfecting the physical network is a never-ending task. Banks that continually cultivate their networks will encourage the greatest growth in profits.