In the beginning there were no branches and no mass-market retail customers. Banks were self-contained. Customers were businesses, traders, merchants, kings, and landowners.

By the late 1800s branching began, along with efforts to serve ordinary consumers and to centralize processing.

By 1900, Bank of America had our country's first large-scale branch network, and the Post Office Savings Bank had the United Kingdom's first.

After World War II the United States went on a branch-building binge. The total swelled from 17,287 in 1946 to 35,350 in 1970. The ratio of branches to main offices reversed-from one branch for seven main offices in 1934 to 6.5 branches per main office in 1996. And from 1956 to 1966 the industry's occupancy expense grew 15% a year, reflecting a buildup of drive-through branches.

Branch networks are very expensive, constituting 50% to 60% of the industry's $160 billion cost base, despite the migration of processing to central facilities. Even the costs of branch automation are high, some $14.7 billion globally, though rising much more slowly than those of call centers and on-line systems.

Many observers have predicted the death of the bank branch. Yet in 1996 the United States had 67,316 commercial banking locations, a number still slowly increasing.

Branches are being built. Customers - not necessarily unprofitable ones- still visit branches. So the event here isn't really death, it's transmogrification-the assumption of a new form.

A branch simply isn't what it used to be. Today it might be described as a quasi-attended, quasi-automated outpost for sales and servicing. The purpose of the teller or customer service representative is to sell, to advise, or to assist in using the automated equipment. The technology is for the transactions.

Charles Schwab, the epitome of electronic and indirect channels, has opened over 270 brick-and-mortar branches. Why? The key to customers with over $250,000 of assets is financial planning, and that requires a personal presence within reasonable driving range.

Only about 20% to 25% of all U.S. bank locations still contain some processing or decision-making functions. And this number is diminishing.

Supermarket banking exemplifies this transmogrification. It brings banking to where retail commerce is occurring. Other high-volume locations are possible, but 90% of branches that are integrated with another entity are where people buy food. From 675 in 1989 the number of in-store branches is expected to rise to more than 9,000, or 10% of total locations, by 2000.

Supermarket branches are an attempt to make the functioning, cost, and organization of customer facilities meet technological realities. All routine transactions or inquiries not involving paper can and should be done through a remote electronic channel such as call centers, pre-standing arrangements, or on-line. All routine transactions or inquiries involving paper can and should be done through on-premises technology, such as ATMs and kiosks.

The technology used in a supermarket branch is oriented to support these points. The standard platform system must have a user-friendly graphical user interface. There must be a single system for the staff to work from, not two or more. And that system must handle off-line transactions when the host is down at night.

Since the skills of supermarket branch employees are limited-they may have sales skills but not product expertise-and the branches may be open 24 hours a day, additional automation is increasingly prevalent. This automation may include:

One or more ATMs, some with advanced functions.

Telephones to reach the call center.

Video kiosk technology.

Internet access to the bank Web site.

Data base technology to link bank systems with retailers' loyalty schemes.

Coin-collecting machines, money-order stands, and check-cashing machines to serve the unbanked and the underbanked.

And banks are only half of the supermarket banking equation. The other half-the retailers-have their own needs and points of view.

Financial services are hardly new to them. They've cashed checks in high volumes. Debit and credit cards are accepted in over 80% of all chains. Most take coupons, some have loyalty cards, a few have frequent-shopper programs, and over half sell lottery tickets.

Retailers require from their bank tenants both traffic stimulation and brand building. The rents are hardly enough.

Most U.S. in-store branches are bank-branded. In a twist, Ukrop's and National Commerce Bank started a new in-store bank brand, First Market Bank.

In the United Kingdom, however, the large supermarket chains Tesco and Sainsburys have started their own brands. Tesco Personal Finance and Sainsburys Bank both partnered with Royal Bank of Scotland. Through high rates and simple products they have already attracted several hundred thousand deposit accounts. But all servicing is telephone-based, not in- store. In essence they are repeating the Sears experiments of a decade ago.

Canada's Loblaw has found a middle ground. The brand is President's Choice, the same as the supermarket's private-label food. The bank partner is Canadian Imperial Bank of Commerce. In-store servicing and kiosks are planned for 133 locations.

Supermarket banking has some economic drawbacks. Despite sales of $3,500 per household per year, supermarkets will neither replace drive-through windows nor meet every definition of convenience. In-store branch deposit growth may lag that of more traditional branches. The certificate of deposit mix may be higher, causing a higher cost of funds. Entry to new markets may be problematic.

Wells Fargo is the leading example of wholesale conversion to supermarket branches. At the end of 1996 it had 527 "banking centers," consisting basically of attended ATMs, and 247 fully staffed in-store branches. It also maintained 325 stand-alone locations. The ratio of in- store to total locations currently varies from 64% in San Diego to 0% in Seattle.

Wells had some trouble with its banking centers because they were a bit small. However, account openings were higher at the banking centers: 17.4 per month, versus 10.3 in the fully staffed in-store branches.

So the branch is not dead, but transmogrifying. It is no longer a place where physical monetary value is held or processed. Decisions should not be made there. And it need no longer be a place where members of a geographically defined customer set come to process transactions.

In place of the old-style branch is a complicated network of "touch points": assorted technologies and people to serve customers and sell to them as needed. To provide convenience, the number and variety of touch points will expand.

Physical touch points will be driven by the need for personal contact in complex situations or to touch and handle paper. Electronic touch points will grow, especially through the Internet. Indirect touch points are already quite prevalent and will continue.

What results can we expect?

More need for branding to define and identify the provider through different touch points.

More relationships with those who control customer convenience, such as supermarkets.

More investments in back-office middleware, data warehouses, and customer relationship management systems to manage multiple touch points.

More experiments to reduce the share of routine transactions still handled with personal intervention.

More use of technology to develop new options and make the existing ones better and less costly.

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.