Bill Gates stated that banks were dinosaurs-then he said he meant it was their systems that were dinosaurs. Meanwhile, the U.S. financial services industry is spending $35 billion a year on technology.

Can we reconcile these statements?

The truth is that many bank systems do go back hundreds of years. Some methods, procedures, and functions have never changed. They never had to.

Consider as an illustration the data processing arrangements of the British Post Office Savings Bank. Started in 1861 by William Gladstone, the POSB quickly became the world's largest savings bank; it had 12.4 million accounts by 1920. Depositors could use any of 10,000 post offices, and these people averaged 2.5 transactions per year.

(My source is Martin Campbell-Kelly, "Data Processing & Technological Change: The Post Office Savings Bank, 1861-1930," published in Technology and Culture, University of Chicago Press, Vol. 39, No. 1, Jan. 1998.)

Transaction processing was done entirely on paper and by hand.

Yet many systems for organizing and processing data were remarkably similar to today's. Understanding which ones were, and why, gives an insight into the dinosaur.

We live in a world where half or more of banks' paper has been replaced with electrons. Some things have changed since Victorian days, and some have not.

All POSB accounts were maintained centrally in hand-written, fixed-page ledger books, one ledger per post office, one account per page. Deposits and withdrawals were manually copied into the ledgers daily, having been transmitted among the post offices on large "account sheets."

The analogy to today's centralized batch-processing is fairly clear.

In the Victorian era, signature verification was used on warrant requests. It still is used on high-value checks.

Clerks also did all the correspondence, copying, sorting, balancing, reconciling, and interest-posting manually. At its peak, the POSB's 4,000 employees handled 7,000 to 10,000 transactions/ year/employee. The 400,000 back-office employees in U.S. banking today handle 100 billion to 200 billion transactions a year, a productivity rate 25 to 70 times greater.

Low productivity in the old days required some bizarre operational practices. For example, the POSB did not keep a running account balance. Clerks would simply inspect the ledger entries when approving a warrant.

The interest algorithm of 0.5 pence per month per whole pound was designed to allow easy figuring. Depositors had to send their passbooks in to have interest posted-which happened once a year. There were never any customer statements.

The POSB's net interest margin was only 50 basis points, so costs were of paramount concern. They were addressed through numerous process refinements, use of low-cost labor, and elimination of work steps where possible. Yet the bank resisted the then-available mechanization in strikingly modern ways.

Loose-leaf or card-based ledgers were a huge improvement on fixed-leaf ledgers. (In 1910, the POSB had recopied by hand more than two million accounts in 35,000 fixed-leaf ledger books, simply to purge them of old accounts.) It was a change POSB management had delayed for 30 years, worried about the growth of fraud, conversion costs, and general disruption. When the conversion finally occurred between 1925 and 1930, POSB chose elaborate proprietary card-based systems, unique tables, specially designed posting machines, and so forth, locking itself into a technological dead end.

Of course, there have been numerous improvements since. The POSB offered just one product for 68 years. And the deposit rate remained at 2.5% for more than 40 years.

A withdrawal took three to seven days to process, with the exception of a higher-cost telegraphic withdrawal. Payments could only be made by buying a postal money order. The minimum deposit was a shilling, and small transactions were not confirmed.

The moral is that today's legacies are not just computer systems. All banks have concepts, processes, and functions with roots in the Victorian era-or earlier. For all its advances, the electronic world will continue to evolve from these fundamentals.

Take delivery channels. The POSB piggybacked on 10,000 local post offices, which essentially served as branches. Today, up to 40% of new information technology investment goes to nonbranch delivery, including call centers, automated teller machines, voice response units, and Internet banking.

The imperative of geographic ubiquity requires that we solve a host of issues never imagined in the Victorian era.

But individual customer accounts will remain the heart of banking systems, even if they are viewed as part of a wider relationship. Centralized control of accounts will remain, even if a distributed computer architecture is used. Two-stage transactions-initiation and consummation- will remain; we see this today in trade order and execution, clearing and settlement, and card authorizations.

Electrons move instantaneously, and so should financial transactions. Float, but not spread, will probably disappear. Electrons have no marginal cost, so all users and sources of funds could be connected directly. The bank balance-sheet "black box" should ultimately disappear, replaced by conveyances that are more direct. We see this already in the mutual funds and asset-backed securities of our disintermediated financial marketplace.

A century ago, the POSB put all its funds in the United Kingdom. Treasury and was the only viable investment option for England's lower middle class. Today, every investor has thousands of choices. And with electronics, those choices could multiply into the millions.

Transaction volume growth reflects more complex, interdependent, wealthier lifestyles. A POSB depositor relied on cash and averaged only 2.5 account transactions a year. Today it is common for a retail bank account to see 1,000 transactions per year. In the future, after elimination of cash and growth of debit cards and Internet micropayments, 20,000 transactions a year could flow through an account.

A new role for banks is as the front-end integrator or access point to the multiplicity of financial electrons. This role defends, defines, and manages the customer. Examples range from Treasury workstation software to personal financial managers and personal banking Web sites. Someday, no bank customer will be without one.

Some bankers fear Mr. Gates' Microsoft, believing he wants to bypass the dinosaur. He could skip the legacy systems and functions and focus instead on the new opportunity in customer control.

While this may happen, there is also opportunity for banks. Customer relationship management is clearly among banking's strongest trends. The dinosaur will evolve. T. Rex survived for millions of years.

In financial services, there has been no cataclysmic meteor strike to wipe out old forms of life. The business is stronger than ever, despite some old methods. But only by adding technology where and when needed can banks make sure they maximize their potential.

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.