One problem with information systems in banks today is that most of the information flows uphill.

Advanced decision support systems are designed to pull data in from operational systems and push the information up to the business analysts who advise senior decision-makers.

As information rises up the chain, these knowledge experts apply powerful "analytic engines" to transform raw numbers into comprehensive views of risk, return, and profitability on products, business units, and customer relationships.

This upward flow is great for strategic decision-making. Modern decision support systems are beginning to give business analysts unprecedented insight into the fundamental economics of banking. And with the advent of more powerful data bases and analytic financial applications, we can now drive analysis down to the individual level.

Instead of focusing on representative customer categories, such as earnings and demographics, we can now measure each customer and each financial product a customer owns. These insights can lead to the formulation of more profitable business practices, and most banks are eagerly starting to adopt new sets of best practices.

The drawback is that the knowledge created tends to remain locked in the minds of financial and marketing analysts who are beholden to their executive patrons.

There is no mechanism for taking advantage of this vast output, collating it, and distributing it to the workers who must carry out the new ways of doing business. Knowledge trickles down to operating levels in the form of policy directives but seldom travels direct.

A chief executive officer's analysts may have access to profitability profiles for all the bank's customers, but the customer service representative, or CSR, will not know the profitability of the customer who asks to have a $1 automated teller fee waived.

Why can't banks use the same information that guides executive decisions to drive the execution of these decisions in the branch, call center, and electronic bank? Having pumped profitability information all the way up to the top floor, why can't we let it flow back down our networks to the places where profits are actually made, the "touch points" where customers interact with the bank?

The technology exists to begin two-way information flows. Terabyte-scale data warehouses and sophisticated analytic systems are being adopted in many large banks, as are advanced communications networks based on low-cost Internet technology.

The breakthrough comes in the form of rule-based messaging systems that can pull pertinent customer relationship information from a data repository and push it in real time to the places where it is needed.

This messaging technology can be readily deployed through existing bank networks, the Internet, or proprietary intranets.

The great promise of this technology is that it helps customer-contact agents make optimal decisions while doing business with the customer. Consider the simple example above in which a call-center representative is talking to a customer who wants a service fee waived.

The first step at any bank is for the CSR to call up the customer record in a centralized customer information file. This will probably list the customer's accounts with the bank, current balances, and most recent transactions.

But this file does not reveal the real value of the customer to the bank. The CSR can make a rough guess based on the account balances, but these can be deceiving.

Customers with high balances and multiple accounts can sometimes be quite unprofitable, and customers with only one or two low-balance accounts may be quite valuable.

The irony here is that the bank may already have a detailed profitability profile for this customer sitting in the chief financial officer's profitability analysis application.

It may even have predictive behavioral models that indicate whether the customer is likely to defect if a fee waiver is withheld.

Other pertinent information, such as the customer's role in a significant commercial banking relationship, may lie isolated in other information silos. It's all there somewhere, but it's not available to the person who has to make a quick decision about this customer.

Let's take our simple example a step further. Imagine an information network that not only gives the CSR the information needed to make the best waiver decision but also signals a significant marketing opportunity.

Assume, for example, that the bank's information network can tell the CSR that this customer has several profitable accounts, a good credit profile, and a substantial mortgage.

Assume further that someone in the bank had the wisdom to record information about the number and ages of the customer's children. A rule- based messaging system could cue the CSR that the customer's oldest child has recently turned 17.

The CSR can now call attention to the bank's college funding package, which might combine a home equity credit line, student loan, and student checking account with Internet banking access and monthly cash transfers from a parent's money management account.

Let's assume further that with a few strokes on a keyboard, the CSR can send the customer more detailed information, schedule a follow-up phone call from a relationship manager, and e-mail an application form. The bank may not only sell three or four products but also create a long-term relationship with the college-bound child.

This example illustrates a customer relationship management strategy that uses all existing information to drive customer behavior at a time when the customer has already chosen to interact with the bank. Most frequently, that is when the customer is making a transaction.

The rule-based messaging technology can be coupled effectively with multiple delivery channels. Most customers do not hesitate to use banks' wide variety of touch points. Systems that can track customer contacts and bank responses across these touch points can eliminate redundant messages and maximize the effectiveness of marketing efforts.

This strategy's major requirement is transformation of the traditional customer information file into an interactive, on-line customer information store. Banks can start doing this by expanding these files to consolidate all available customer information, including sophisticated profitability and modeling data now languishing in the minds and desktop computers of business analysts.

The second step is to use contact-manager applications to register every customer interaction through every channel.

The bank must have some collective memory of these encounters if it is to maximize the relationships' profitability.

The third step is to render this comprehensive customer information proactive through rule-based messaging. These systems can incorporate the bank's best practices in the data store's workings.

The more information a bank has on a customer relationship, the more powerful the interactive process can be. The processes outlined in this article would work even more effectively in commercial banking relationships. Cash management, for example, is complex and offers numerous options to both seller and buyer. Feedback from customer choices is almost continuous, and relationships often span a wider variety of products.

There has been much discussion in the banking industry recently about empowering employees to behave as shareholders rather than hired hands. Without the right information, it is impossible for employees to act consistently in their institutions' best interests.

If a bank truly wants to empower employees in this way, then it must supply the necessary tools. Two-way information flows are a fundamental requirement. If information cannot flow downhill, then earnings probably will.

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.