Forced Fit?

By Kenneth Cline and Lauri Giesen

As banks categorize their customers, they need to make sure they're measuring the right things and using the data pragmatically.

It looks great on the rack, but does it really fit the customer? That's the debate brewing over segmentation, a sophisticated computer-driven process that attempts to identify the most salient drivers of behavior within major customer groups and then use that information to better serve the individuals within those groups.

While some banks report a measurable performance lift, challenges remain with these programs, which are a subset of the larger process known as customer relationship management.

There are sharp disagreements over which customer characteristics are the most useful. And translating abstract findings into productive work tools for marketers and front-line personnel has proven difficult.

Nor is segmentation strategy a magic solution that can transform a sluggish performer into a star overnight. To capitalize on the concept, practitioners say, an institution first must organize itself around the total customer relationship, as opposed to product silos, and establish a robust sales management system. "Segmentation is not a stand-alone exercise," says James Rager, vice chairman of Royal Bank of Canada in Toronto. "You have to connect it into the way you run your business."

Retail financial services executives are under pressure to address these challenges. In the current austere economic environment, institutions are straining to justify their substantial investments in data collection and analysis technology. In 2001 alone, U.S. banks spent an estimated $701 million on the "analytics," or segmentation, component of customer knowledge technology, according to Tower- Group. That figure is expected to reach $902 million by 2005.

Segmentation also goes to the heart of competitive differentiation in banking, being the primary method by which large banks emulate the personal feel and customer knowledge of community banks. When the process works as it should, call center and branch employees can use the data on their PC screens to interact with customers as individuals, despite the lack of any prior contact. "Segmentation is our tool for creating a personalized approach to customers," says Joseph Guyaux, an executive vice president and head of retail for PNC Financial Services Group in Pittsburgh.

Leading practitioners suggest that segmentation's payoffs hinge on the mastery of some critical details. There are three main factors: the type of customer information that is gathered; the manner in which this data is
categorized; and the way in which the output is fashioned into work tools
for back-office marketers and front-line staff.

Each of these processes must function correctly if segmentation is to yield the intended results. "I don't think we'd be able to do what we're doing in retail banking without segmentation," Guyaux says. "But I'm also sure that if we didn't execute well, segmentation would be worthless."

Banking Buzzword

Segmentation became a banking industry buzzword during the '90s as a surge in computing power and data storage capacity enabled institutions to build "data warehouses" for all the customer information they had previously maintained in separate files. Banks then combined this data with credit and demographic information acquired externally to compile an in-depth view of each customer.

The information by itself is of little value unless it is analyzed and categorized, which is where the segmentation comes in, aided by analytical technology such as decision engines and predictive modeling software. By viewing each customer as part of a type, or segment, marketers can design products and services that appeal to broad categories. And service personnel can deal more effectively with individual customers.

The payoffs are potentially tremendous. It's not cost-effective, for example, for institutions to market home equity loans to their entire customer base. Segmentation identifies which customers might be receptive to that particular product and even helps marketers design the most appropriate way of approaching them.

The issue then becomes, what's the most useful way to segment customers and what types of information should be included? In the early going, institutions relied mostly on profitability data. By combining their own account and transaction history information with demographic and credit data sourced externally, banks were able to categorize their customers by the degree to which they contributed to (or detracted from) institutional profitability. This led to the famous "profitability skew," the recognition that 20% of customers contribute about 80% of a typical retail bank's profits.

Segmenting by profitability alone has its drawbacks, however. While it may help retention, by allowing institutions to cull their most profitable customers for special coddling, it doesn't provide any insight into financial needs and buying patterns. Today's unprofitable customers can become tomorrow's profitable ones, moreover, so it's unwise to pigeonhole them. And a profitable customer may not necessarily be predisposed to purchase an additional product. Upscale clients, in fact, are usually the ones most resistant to seeking investment advice from a bank — they already have their broker for that.

Segmentation practitioners now accept that profitability is only part of the puzzle. The new approach is to combine demographic and behavioral information describing the current financial profile of customers with attitudinal data that pinpoints their financial needs and aspirations. In a 2001 study, BAI and Cambridge Group argued for the primacy of attitudinal data in generating "predictive insights" into consumer demand for financial services.

"Attitudinal segmentation was critical for us," says PNC's Guyaux, citing an internal survey of consumer attitudes toward large and small banks that his institution did three years ago, including responses from both established customers and prospects. Ascertaining that somewhat less than half of bank customers preferred small banks no matter what, PNC was able to incorporate those insights into the design of its account packages, which are customized for various customer segments. The data also allows PNC to avoid expending marketing effort on people unreceptive to large banks, such as those who recently opened accounts in response to the company's "free checking" offer.

When customers signed up for those accounts, PNC was able to take the information gleaned on that occasion and then match it up with the profiles generated by the attitudinal survey. "We knew we were going to get a lot of people who like small banks, which is no problem," Guyaux says. "But we don't try to go back and cross-sell them a lot of other stuff."

Getting Traction

Once the customer data has been collected, an institution must organize it into categories, or segments, that can be usefully applied against its established customer base. Care must be taken at this stage to limit the number of categories to a manageable level.

"Banks will drive themselves crazy if they try to accommodate every
customer segment they identify," says Kathleen Khirallah, senior research analyst at Needham, Mass.-based TowerGroup. "They have to decide which segments they want to attract — which ones will give them the profit they want."

Royal Bank of Canada, for example, uses six strategic codes for segmenting its customers. These are: current profitability, current potential, life stages, credit risk, channel preference and vulnerability to attrition. The Toronto-based company ranks nearly all of its 10 million customers by their placement within each segment.

A centralized staff of about 50 people runs the analyses, electronically generates product or service recommendations for each customer, and then sends that information to the company's call centers and branches in the form of sales or referral prompts. Front-line employees can call up a screen and find that a particular customer is eligible to receive, say, a pre-approved credit card, line of credit and mortgage.

There is some debate in the industry as to how much information representatives should receive. Royal Bank's policy is to provide representatives with "tactical" information, such as a customer's propensity to buy a particular product, while retaining the profitability analysis with the back-office statisticians and marketers. "We try to keep the analytical complexity away from the people who deal with clients," Rager says. "Because if you give that information to people, they have to figure out what to do with it."

PNC recently Web-enabled its call center and branches, which facilitates the delivery of segmentation data to front-line employees in both areas. This data, organized into several "folders," or screens, provides the sales reps with a customer's account balances, transaction history, and demographic profile. One of the screens presents sales opportunities considered relevant to that client, as well as a history of marketing approaches already made by various units of the company — direct mail, for example.

As is the case at Royal Bank, branch and call center reps at PNC don't see the profitability codes assigned to a particular customer, since that might cause the employee to provide a higher or lower level of service than is appropriate. "Every customer should have an expectation of a certain level of service quality," Guyaux says.

The service differentiation comes in other ways. The routing of a call center contact to a particular type of rep, for example, is determined by segmentation data that includes the types of products and balances held by the customer. Thus, a high-balance, multiple-account customer will be routed to a senior service rep who can handle complex questions about both banking and non-banking products. A customer with only one certificate of deposit, by contrast, would more likely experience a longer wait time.

"It saves lots of expense not to go after everybody the same way — to match cost-to-serve against the value provided by various types of customers," Guyaux says.

Does this cost-to-serve match provide institutions with a quantifiable bottom-line benefit? It's hard to say. Most banks cite internal metrics that don't show up on quarterly earnings reports and can't be verified by outsiders. PNC, for example, claims significant improvement in checking account growth, fueled by an 11% improvement in retention and a 32% boost in new customer acquisition. Royal Bank credits segmentation with improved marketing efficiencies and a nearly 14% increase in the number of highly profitable clients over the last two years.

Prioritizing Service

Cross-selling understandably gets a lot of attention in any discussion of segmentation strategy, since diversified financial institutions are keen to get their banking customers into in-house investment and insurance products, or at least hook them up with in-house financial advisors. But some of the most beneficial aspects of segmentation might actually lie in the area of customer retention.

First Tennessee National Corp. uses behavioral and attitudinal data to identify customers considered to be at risk of leaving the Memphis-based bank. Those in the profitable category then receive tailored product or service offers: for example, free financial planning guidance. While this is expensive for the bank, First Tennessee has found that customers who use the service tend to be the most loyal, says David Miller, vice president of strategy and consulting.

Another service enhancement, used by both PNC and Royal Bank, is to recommend appropriate packages. When customers come in to discuss their personal finances or open an account, their information is matched against the profiles generated by the bank's segmentation data. Such an analysis might show, for example, that Customer "A" is better off in a no-fee, low-balance account, while Customer "B" could gain by earning interest on large balances.

PNC, which began this account-modeling process in 1998, aims to give its customers a sense that the bank is actively looking out for their best interests, thereby strengthening the relationship. "If a bank can't be helpful to customers on their basic checking accounts," Guyaux says, "why should customers expect it to help them with their investment and insurance needs?"

By establishing such linkages, PNC and Royal Bank use segmentation to fortify their brand promise of personalized service. But Guyaux and Rager both stress that carrying through on this promise requires more than just an effective segmentation strategy implemented by the data-collection staff and back-office marketers. It also takes some institutional support in the form of a retail operation structured around the customer.

When Guyaux took over PNC's retail operation in 1997, for example, he grouped all the product and marketing managers under one head of marketing. Distribution channel managers likewise came under one manager. Both of these executives now report directly to Guyaux. This type of reporting arrangement helps keep line managers focused on the total customer relationship, Guyaux says, as opposed to just maximizing revenues for their individual units.

As is the case with so many other areas in banking, however, savvy tactics still can fall woefully short if there's no overarching strategy. Consultant Charles M. Bruney, executive vice president for Atlanta-based Speer & Associates, says banks have better databases at their disposal than ever before, but still don't always use this data to best advantage. "The problem is that most institutions do not use segmentation data as part of a long-term strategic plan. Rather, they use the information on a product-by-product basis or a test cell-by-test cell basis. They lose sight of the
bigger picture."

This turns the pressure back on senior managers to think clearly about what their institutions stand for and the distinctive ways in which they intend to serve prime customer segments. In other words, segmentation contributes to strategy but doesn't replace it.

Guyaux reinforces this point by saying segmentation is simply "another tool" in his strategy of creating a "personalized feel" at a large bank. "As long as it's contributing to that, I have the freedom to use it in any way that makes sense. But if I get stuck on saying, 'segmentation itself is my strategy,' I'll end up spending too much and being disappointed."

Mr. Cline is senior editor of Banking Strategies; Ms. Giesen is a freelance writer based in Libertyville, Ill.

Copyright © 2003 by Banking Strategies, published by BAI.

back to top