Getting the most out of customer information.

By using statistical models and well-mined data in careful combination, a company cannot only determine a customer's "profitability potential" - but boost it.

One of today's most popular business mantras is "customer driven." It seems every company aspires to become more customer focused, but for most, the lens is distorted by the company's own structure and the competing ambitions of its departments. Few businesses have a clear vision of either their customers' true value to them or the value customers themselves ascribe to their relationship with the company. This gap leaves the field - and the future - wide open for those companies whose chief executives can see beyond their own customer rhetoric and offer their people both the vision and the means to sustain win/win customer relationships.

At American Management Systems (AMS), we've found that the stakes are particularly high for those CEOs whose companies provide varied services originating from different business units within the company, such as banks and telecom providers. Consider these examples:

* A financial institution repeatedly solicits an existing customer for a credit card, then denies the application once it's returned. At the same time, the customer is charged an overdraft fee when he bounces a check. Yet this customer has more than $50,000 in a money market account. Will this customer consider the bank when shopping for new investment options?

* A long-distance provider continually offers usage incentives to an existing customer, unaware the customer generates high costs and is at high risk of default. Does the company really want to encourage additional usage?

How is such illogical and self-defeating activity possible? The problem, simply put, is that most customer relationships aren't managed from either an enterprisewide view or with a long-term perspective. The vast stores of detailed customer information now available thanks to major advances in data warehousing and networking technology - can be overwhelming for organizations that are structurally unaccustomed to handling it all, and that includes most of us.

Banks in particular have grown up as stovepipe organizations, and their various units - credit, mortgage, deposits, bank cards, etc. - routinely don't share information. Instead of building whole relationships with customers, corporate fiefdoms compete for ownership of their piece of the customer, unwilling or unable to make the internal tradeoffs that maximize the value of customer relationships. And even when companies take some measure of overall customer value into account, they often have a short-term view and can't perceive long-term benefits.

For the customer, the result can be a disjointed and low-value experience. For the corporation, viewing the customer as a series of unrelated transactions rather than a long-term annuity means lost opportunity to enhance each relationship's real profit to the bottom line. For companies with very large customer bases and multiple products, it is no simple feat to determine the most profitable relationship strategy. But a few corporate leaders are successfully doing just that by creatively weaving sophisticated technology together with old-fashioned customer relationship skills.

Transforming U.S. Bank

At U.S. Bank, most of our major technology projects were focused on efficiency; indeed, this has been the primary goal for many financial institutions over the last decade. While this continues to be very important, we realized that improving efficiency goes only so far and that we needed to find a way to improve the revenue side of the equation and position the company for growth while continuing to keep costs down.

We saw a great opportunity to improve performance by leveraging the wealth of information that we have about our customers better. So we launched a two-year program to transform the bank into a customer-focused enterprise, and made that the bank's top technological priority. We began with a considerable leg up: strong executive sponsorship for the program as well as a culture accustomed to change; experience using technology to help make certain customer decisions; and a centralized operating environment with a common technology infrastructure.

But, technology aside, there were numerous business and organizational challenges to be addressed. U.S. Bank previously lacked measurement tools and incentives for maximizing the value of each customer's relationship to the whole organization. We were too narrowly focused on performance objectives related to individual product groups and internal functions, such as marketing, credit, and customer service. So, for example, in the past, marketing and sales officers might be aggressively targeting an individual for a loan product - only to have that individual later turned down for the loan by the credit department. This counterproductivity affected us negatively at all levels of the company.

The basic premise behind the U.S. Bank initiative is to use statistical models of customer behavior to routinely scan customer data and, based on each customer's unique characteristics, to predict the customer behaviors that drive relationship profitability. The bank runs all customers through models of customer value, risk, attrition, and the propensity to buy specific products on a regular basis. The graph on the opposite page, for example, demonstrates results of three test programs in attrition management. While group B delivered the best performance in terms of attrition rate - the lowest, at 12 percent - it did not deliver the highest customer value, expressed in terms of annual profitability, within that segment.

By managing only one-dimensional metrics, such as attrition, rather than by customer value, companies often sub-optimize the value of customer relationships, and they may consequently spend too much money to keep an unprofitable customer or use inappropriate methods to deal with a more profitable customer. By better understanding the true dynamics, the bank can develop effective strategies for managing each customer relationship in a way that maximizes long-term value, both for the customer and the company.

The results of this analysis are then translated into a set of decision strategies tailored to boost the profitability potential of each individual customer. These decision strategies extend across the relationship lifecycle. They determine which products or services are best marketed to each customer. They also establish how much credit to extend, how to service the customer, and what strategies to use to retain the customer relationship.

Using technology, the decision guidance is communicated to front-line staff, who then can focus more on how to build lasting relationships through more efficient service and less on trying to decide the best way to handle each customer transaction. For example, in the past, a banker could not be sure that a customer would be approved for a loan without initiating a credit evaluation; now an evaluation is performed proactively, and the banker is presented with a set of pre-approved products. This has a very favorable impact on the front-line staff, who now feel more confident in cross-selling the bank's products, and can do so with greater efficiency.

The tools can also be used to evaluate not only a customer's present value to the company, but his or her expected potential value in the near future. If a customer, who has a particularly high potential value, bounces a check, the system may cover the check amount and waive the fee to win this customer's loyalty.

Overcoming the Challenges

As might be expected, implementing a system this complex requires a substantial commitment by the company to manage customer relationships effectively. The process introduces four highly challenging requirements:

* The business must systematically assess each customer's total relationship value. This is typically accomplished through some form of enterprise-level value equation. While it may not be possible to tie some cost elements down to the customer level, reasonable approximations can suffice if they are analytically sound.

* The organization needs a decision analysis platform to make customer decisions across whatever internal departments and business units control parts of the customer relationship. Specific decision strategies are then defined to support all the areas that drive relationship value, such as marketing, customer acquisition, customer service, and retention programs. On the technology side, this requires a customer-level database for decision making as well as a decision engine to perform customer analysis, select the right decision strategies, execute them systematically, and communicate them to each customer contact function.

* Cross-enterprise processes, technology, new competencies, incentives, and, in some cases, cultural change must be developed and implemented to enable these decisions to be carried out across product silos, distribution channels, and internal functions.

* Results must be captured and continuously analyzed to find the right combination of decision actions for maximum value to the institution and customer.

It is critical to deliver initial results in months, not years - and these projects can certainly take years if you let them. A smaller pilot version of the larger roll-out can go a long way toward helping to identify tangible financial benefits to the business early in the cycle. U.S. Bank conducted a pilot before full rollout, which resulted in early benefits and a better understanding of the project's real impact. This accelerated buy-in among key stakeholders and allayed fears that a complex enterprise could not realize such a far-reaching vision.

Implementing the Customer Vision

At AMS, we have worked with a number of organizations in the area of customer value management. Most organizations simply don't know where to begin in developing an all-inclusive vision for customer management. Another financial institution, one of the largest retail banks in North America, used a three-step process to develop its own concept and is now in the early phases of implementing a customer management program.

Although the leader in its market category, this company was not content to sit on its laurels. Despite its past success, it was clear to executives that in order to achieve its three-year business targets, the bank required a breakthrough approach that placed the customer squarely as the centerpiece to generate significant increases in revenue.

First, working with consultants, the bank developed a strawman vision of what "customer focus" really meant to the organization. The words are slippery enough that without a clearly articulated vision, everyone has a slightly different view of what it means; without a common understanding, implementation will likely fail. As a first step, the team mapped the ultimate vision of how the bank wanted to manage its customer relationships - in this case defined as using information to profitably deliver a customized, high-value, and consistent experience to each customer, every time.

Up to that point, the bank had embarked upon a series of initiatives to position itself for its transformation to a customer focus. For one thing, the organization was fairly advanced in its development of a customer data warehouse to capture customer information; executives were also developing a customer segmentation approach, using statistically based models to measure current and future expected profitability, credit risk, propensity to buy, and other dimensions of customer behavior that drive value. Further, the bank had been investing heavily to inject a sales culture among front-line staff. By injecting several unique twists into the vision as well as leveraging capabilities already in motion, the team focused on how the bank could achieve a market advantage the competition could not easily replicate.

Then the real work began. Up until that point, getting agreement from key players across the bank that a customer focus was desirable was akin to getting buy-in that environmental concern is a "good" thing it's only when the impact on timber versus the spotted owls starts to be understood that the rubber meets the road.

The team followed by initiating a broad stakeholding and education process, since some customer management concepts - such as quantitative modeling and decision-support tools - were new to key people involved. Then they conducted an implementation readiness assessment, looking at the "how to" phase in the prerequisite processes, organizational elements, and technology enablers.

The organization is now developing an implementation strategy that will require coordinated teamwork. By painstakingly developing a shared vision and marching step by step to achieve understanding and buy-in among the stakeholders, the bank is positioning itself for successful transformation.

While a customer management vision can only be driven from the top of the organization, the implementation itself can be staged in manageable chunks. For example, a large international telecommunications provider has begun this process by differentiating service levels for its business customers in accordance with a customer value model - or more simply put, treating the valuable customers like valuable customers.

By knowing who its most profitable customers are, the company can make service and responsiveness to those customers a higher priority. This means several departments across the enterprise - sales, marketing, customer service, to name a few have to coordinate to create a customer database and apply advanced data mining tools to develop a predictive model for customer turnover. With this in place, the company as one unit can let unprofitable customers go and, at the same time, keep customers happy before they ever think of leaving.

Philip G. Heasley is vice chairman of U.S. Bancorp, a Minneapolis-based bank holding company with assets of $70 billion.

Patrick W. Gross is a founder and vice chairman of American Management Systems, Inc., a Fairfax, VA-based international business and information technology consulting firm with 1996 revenues of $812 million.