Over the last decade, banks around the world have developed "alternative distribution"-a range of channels beyond traditional branches and the mail through which a bank can communicate with its personal and small-business customers and vice versa.

The first alternative channel was the automated teller machine, followed by the telephone, and now the computer. There are others, such as smart phones, specialty branches (workplace, supermarket, mobile), kiosks, and interactive television, but they are basically variations on a theme.

Each channel was developed with the same basic expectation: that it would be cheaper to operate. In the late 1980s and early 1990s, anticipated reductions of 10% to 30% were commonly stated in the press.

Some excited commentators mentioned the possibility of savings as high as 70%. "Expensive bricks-and-mortar" became a commonplace characterization of branches, and everyone foresaw their demise, using phrases like "bringing down the branch network."

But in examining the economics of our clients' distribution channels, we have not found the expected benefits of alternative distribution. In fact, in several cases, we found aggregate costs had risen.

In early 1998, to test whether these findings were anomalous, we conducted a study among major banks in the United States, Canada, and Europe. Respondents were asked about their distribution channels, the rough timing of their development and capabilities, and the original motives for building them. They were also asked to assess how well the original objectives had been met and how their outlook on the whole matter had changed.

We received responses from nearly 50 banks, from the largest down to about $20 billion of assets, or the equivalent if not in U.S. currency.

Cost reduction was by far the most frequently cited reason for building alternative distribution. Second was "protecting the customer base" and a distant third was "attracting new customers."

Unfortunately, the other part of our surmise also turned out to be correct: 87% of respondents said their aggregate distribution costs had risen, not fallen. And the lucky 13% said costs had fallen only 3%.

On a brighter note, 50% said their expectations with respect to defending the customer base had been met and over 80% said the same thing with respect to attracting new customers.

The responses clarified why the hoped-for reductions have not materialized. One important reason is that the number of transactions has increased substantially: Respondents estimated an average increase of 25%.

Seemingly encouraged by the accessibility of the new channels, their ease of use, and perhaps the zero marginal cost (to the customer) of executing transactions, customers have significantly increased the number of times they transact.

A second reason why costs have not come down is that while "transaction migration" has occurred, "customer migration" has not. Customers are now highly selective in their use of channels: they use branches for deposits, applications, and complaints; they use ATMs for withdrawals, transfers, and balance inquiries; they use the phone for inquiries and bill payment; and they use personal computers for checking balances, watching for check clearances, and paying bills.

Very few customers have moved all their banking transactions to the new channels.

Third, the few customers who have shifted all their activity to nonbranch channels do not fall into any one category, so there has simply been no way to "take the branch network down." Though new distribution channels have grown rapidly and gained a share of all transactions, branches still execute more transactions than they did in 1992.

Finally, the newer channels have not actually been as cheap as advertised. Their marginal transaction costs are much lower, but the fully loaded costs at the channel level are not as low-not low enough to compensate for the increased number of transactions. Some banks say the fully loaded cost of a transaction in a phone center is actually the same as it is in a modern branch.

We sorted the surveyed banks into four categories: Laggards (35%), Median Banks (55%), Advanced Banks (10%), and Bank 2003 (still a concept).

The Laggards were most likely to have a cost-reduction objective-and no others-for their alternative distribution efforts. They had implemented voice-response units in phone centers only within the last one to two years, on average. Their transaction activity was centered in the branch and ATMs, with phone centers accounting for 10% of the total. Their focus going forward was solidly on building their relatively new alternative distribution capabilities.

The Medians had had the voice-response units in use for three to six years, and claimed nearly 50% of total transactions were nonbranch. About one new-product application in seven was taken outside the branches, and they derived 5% of revenue from remote, direct-banking activity. Their current focus was less on building new channels than on enhancing the functionality of those already built. They were optimistic about setting cost-reduction goals for the next five years, but gave equal prominence to continued defense of customer bases.

The few Advanced banks had implemented voice-response units seven to nine years ago. For the immediate future they said they were concentrating on rationalizing their branch networks, using alternative distribution to field active sales efforts. Their emphasized goals were defending their customer base and increasing customer convenience.

The Bank 2003 category was notional-a profile of the most progressive statements or statistics accumulated from various individual banks' responses. Bank 2003 would have installed voice-response unit technology more than 10 years ago, with 20% of customers using PC banking and only 20% of transactions in the branches. Two-thirds of new-product applications would be by phone, PC, or other nonbranch modes.

This composite bank will probably become a reality over the next few years, oriented mainly toward attracting new personal and small-business customers and adding revenues.

The banks in our survey were generally optimistic about the future. Despite their unmet expectations on cost reduction, nearly 60% said they could and would achieve that goal by 2003. And every respondent said it expected to meet its revenue goals for alternative distribution by then.

At the same time, they said they expected no change in the current pattern in which transaction growth, especially in the new channels, outstrips the migration to new channels.

Although banks are now in a position to set more realistic goals for their distribution channels, we sense our respondents' expectations may still be too optimistic on the cost front. Banks must undertake three key initiatives to reach their objectives: align costs and revenues through pricing; exploit alternative distribution's strengths, such as the ability to send individualized messages to customers; and make one branch do the work of two.

Customer behavior is of course influenced by price, but it is critical to note that individual customers, under any given price structure, will exhibit widely varying behavior that results in widely different revenues.

Similarly, on the cost side, a particular product or service, offered to any group of customers, will result in a wide variation in the individual customer behavior that actually drives costs.

The wide skews in customer profitability-sometimes 40%, 50%, or even 60% of customers may be unprofitable, even in a portfolio of customers with attractive aggregate returns-carry over to alternative distribution. Bank products and services are priced in ways that give rise to distortions of the revenue and cost dynamics of distribution.

The challenge for banks is to find a price structure-broadly defined- that can resolve 80% of the dysfunctional revenue/cost misalignment with only 20% of the complexity of a fully rational, unit-priced structure. And then implement it without traumatic PR exposure!

It is possible that better, more rational pricing could make banks indifferent to the expense of operating bricks and mortar. If so, some customers will happily use branches and pay for the privilege, while others will go "direct" and enjoy both convenience and lower pricing, and still others will mix and match and the bank will not care.

In practice, things are unlikely to be so neat, partly because there are limits to what can be achieved with rational pricing.

But it is worth asking: If a particular geography used to be served with 24 branches and now it is also served with 36 ATMs, a phone center and a PC banking capability, is it not possible to be as effective with only 12 of the 24 branches?

The obvious benefit would be a direct reduction in bricks-and-mortar costs. The hidden benefit is that one of the ways that a branch can be made effective over a wider radius (which is what making one branch do the work of two necessarily implies) involves using data-driven marketing techniques. A branch can enhance its physical presence and appeal by tailoring communications to customers and prospects in the surrounding area, thus adding an improvement in customer mix to the more obvious benefit of cost reduction.

Banks need two important capabilities to take these tactical steps. The first is the ability to relate the economics of different distribution channels to customers' likely behavior, under alternative price structures. This calls for internal economic analysis, model building, and research and testing on price alternatives.

The second is information-based decision support: the ability to harness customer and prospect data in support of key decisions in solicitation, cross-selling, retention, and customer service. Armed with these capabilities, banks can look forward to the realization of their ambitious goals for alternative distribution. Mr. Carroll is managing director of retail financial services at New York- based Oliver, Wyman & Co.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.