The author James Thurber, once called upon by a bank manager to explain an overdrawn bank account, explained that he never kept a record of the checks he wrote. When the surprised manager asked how Thurber knew how much he had in the bank, an equally surprised Thurber replied, "I thought that was your business."
Today there is much more to the banking business than keeping track of account balances. And in an area to which banks may devote up to 25% of their overall spending by the year 2000, they must squeeze every last bit of relevance out of the data they compile on their most prized possessions - their customers.
Keeping track of account information is simple compared with some of the other challenges banks face. These other challenges include sharing data from multiple lines of business and mining those data for strategic information that may be used to attract and retain customers.
Most banks have seen their data bases grow by one-third a year, but - even for those that have access to all this information - two questions remain: What's important, and what's junk?
Assuming a bank could get at all its customer data, courtesy of parallel processors and other technology, what does it need to extract from this information?
The quick answer is to segment the customer base on a demographic model - for example, to predict what people 25 to 49 years old will be interested in buying - and then to tailor the marketing of financial products to that segment.
But as is often the case, the obvious answer is not always the correct one.
Banks that ask simple questions of their data bases will get simple answers. The search for market segments should be as complicated as the society we live in, yet easy of access to technology experts and marketing managers alike.
Data mining, when done best, is a higher form of strategic decision support that can be applied to marketing and segmentation, as well as to fraud control. The key to institutional success is to begin with a clear definition of the business problem, then to apply analytical mining techniques.
Why limit yourself to a search simply by age? What about taking affinity analysis to new levels? It is possible to predict trends from data that show, for example, not just how many males 30 to 39 years old are interested in buying life insurance but also how many of them would be likely to spend on a particular product if they are married, have children, or already have bought an investment product.
The average bank does not need just one look at a segment, but several - perhaps ways that will spawn investment products.
As banks move from so-called "dumb terminals" and become PC-centric, it is important to get the right data into the hands of the marketing and sales force as quickly as possible.
As hardware costs come down and the technology matures, banks must learn to rely less on gut instincts for their market analysis and build digital data warehouses that can be used early and often.
Financial institutions must move past their core operational systems to examine market segments, because the industry's consolidation means more competition for business than ever before.
Data mining is a new field, but the possibilities for financial institutions are as exciting as they are unlimited. The endeavor is especially attractive when the financial firm doing the mining has a broad portfolio of products.
Some catalogue retailers and credit card organizations have already built and leveraged integrated data warehouses. As data bases continue to grow rapidly, parallel technologies - both hardware and software - will become more needed by large and small banks alike.
The trick will be to install applications that can draw strategic insight from raw data - not just verify trends that bankers already suspect.
The days when bankers would ask for a market segment study then wait weeks for a computerized report run at off-peak hours soon will be just a memory. So, too, will be the days when keeping track of a savings account was all a bank or credit union had to do to satisfy customers.
Market segments are not waiting to evolve. The financial institutions that can "catch the wave" before it breaks are the ones that will succeed in the marketplace.
Mr. Disney is a manager of advanced technology in IBM's finance industry division.
