“Know thyself,” wrote William Shakespeare half a millennium ago. Shakespeare recognized that we all have elusive wants, feelings, and frustrations that set each of us apart from everyone else. So when Dallas Teachers Credit Union wanted to grow its assets two years ago, it adapted a lesson from the Bard. Rather than manage its database with “neither rhyme nor reason,” the DTCU invested in customer segmentation software to really know its customers — all 150,000 of them.
By creating 150,000 distinct customer profiles, the DTCU was able to determine customers’ individual banking needs. Each customer was profiled, based on more than 100 different data points — things like age, level of schooling, annual income, type of home, and location. All customers also were “householded” to reveal who else in the nest might need one of the bank’s services, such as a checking account, an IRA, a credit card, or a mortgage. Only then was targeted marketing material delivered to individual customers. The upshot: the DTCU increased its assets by more than $500 million in the last 18 months — to $1.45 billion — making it the fastest-growing credit union in the nation.
After rolling together its own internal customer data with an array of external demographic information provided by Acxiom, the DTCU ran it through sophisticated modeling algorithms created by IBM Global Services. This rigorous process enabled the credit union to “rank order” each of its myriad services by customer — that is, to identity which individuals had the greatest inclination to use each service. “The customer segmentation strategy took the guesswork out of cross-selling,” says Patricia Korioth, senior vice president and CFO.
“Instead of sending marketing material to all our [150,000] members and getting a single-digit response rate, we’ll now send out 15,000 pieces of mail and get a 10 percent response rate,” adds Korioth. “The strategy gave us the confidence to re-charter as a community bank.” Now called the Credit Union of Texas, the bank has expanded its potential customer base — it’s no longer just for teachers only.
“We’ve even used the data to determine where our new branches should be located,” says Jerry Thompson, the bank’s chief information officer. “Internal surveys indicated our customers like to bank at a facility within a 10-minute drive. The internal data pinpointed where our most profitable customers resided, and the demographics indicated others of similar banking needs. Up went the branch.”
Customer segmentation strategies using CRM software are almost a matter of course in the financial services industry, which retains an enormous amount of internal transactional and personal data on customers and customer behavior, enabling relatively easy segmentation for purposes of marketing and customer service. In recent years retailers have been discerning their customers’ buying preferences by scanning their credit cards or frequent-shopper cards, and even by using radio frequency identification (RFID) technology to pick them out through their cell phones and other devices as they enter a store.
(Thereby hangs a tale: It’s been reported that at least one retailer has explored the possibility of embedding RFID chips in its clothing, so salespeople could easily identify repeat visitors on arrival and give priority service to these more loyal — and hence more valuable — customers. Though it’s not nearly as intrusive as the iris-scanning shopping-mall come-ons of Minority Report, privacy concerns seem to have put such initiatives on hold.)
Customer segmentation can offer both revenue-building and expense-lowering benefits. Top-line revenues can be increased through more-targeted cross-selling, as well as through superior service that builds customer loyalty. Bottom-line expenses can be pared by relegating lower-margin customers to less expensive self-service options — or perhaps even by steering those customers to another service provider altogether. Self-service is considered the most important CRM initiative by 62 percent of Global 3500 firms interviewed recently by research firm Forrester.
Yet while success stories like the Credit Union of Texas are legion, analysts warn that customer segmentation strategies are hardly risk-free. That same Forrester study noted that 41 percent of interviewees found no return on their investments in Web-based and interactive voice response (IVR) self-service. “It’s a sticky thing,” says Jill Griffin, president of The Griffin Group, an Austin, Texas-based consulting firm specializing in customer loyalty issues.
One sticking point, explains Griffin, is that a customer who is initially categorized as low-end — and then provided low-end service — will be less likely to stay with the provider when his or her material wealth increases. “The Internet has changed customer perceptions of corporate responsiveness,” says Griffin. “People now expect instantaneous conclusions. If they’re required to endure interactive voice response systems that lead nowhere or Web sites that are so labyrinthine that they get lost, their loyalty to your company breaks down.”
Others agree. “You need to be careful when you introduce new customer service options that you don’t completely cut off traditional familiar communications channels,” says Karen Smith, research director at Aberdeen Group, a Boston-based consultancy. “Think about the poor fellow standing on line at the bank for a half-hour who watches someone come in and get preferential, immediate service. If he hits the lottery next week, he may decide he wants service elsewhere.”
A Little Something for Everyone
To segment customers fairly as well as profitably, Stockholm, Sweden-based Nordea Bank sought a customer-segmentation strategy that would allow it to reduce its call-center staff and associated expenses without lowering customer service. The key was a system that ensures each customer is assisted by a service representative with the appropriate skills.
For example, Nordea wanted its high-end customers to be transferred to representatives who had been trained to provide high-touch customer service — and to make a very persuasive cross-selling pitch at the conclusion of the interaction. Low-end customers would be routed to reps trained in the bank’s online service system, to encourage those callers to use the Web for service issues in the future. That channel “is critical to our success,” says Martin Karlsson, Nordea business IT manager, “since it is more cost-efficient, and also very convenient for customers.”
Nordea Bank implemented customer segmentation software provided by Daly City, California-based Genesys two years ago. In Nordea’s 14 contact centers across Sweden, Finland, and Denmark, the bank has increased the time spent talking to customers by 24 percent, yet it has increased its number of employees by only 15 percent. “We use the interactions to provide added value for the customers and for the bank, which means simultaneously advising on and selling products that would benefit the customer,” says Karlsson. “It’s very useful to match customer segments to specially trained and qualified agents. A call from a customer that is more profitable to the business deserves more high-level attention and service.”
Brian Bingham, manager of customer care research at research firm IDC in Framingham, Massachusetts, puts it even more strongly. “Losing a high-end customer can have a profoundly negative effect on a company’s profitability,” says Bingham. “If they have urgent needs requiring satisfaction immediately, and they’re forced to go through the same channels as low-end customers, they will become dissatisfied. There’s an old expression, ‘it’s significantly more expensive to acquire a new customer than to retain an existing one.’ “
Many CRM analysts recommend a value-based segmentation model similar to the one in place at the Credit Union of Texas. “You want to create a target list for a sales campaign based on the attributes of the customer,” says Steven Bonadio, senior analyst at Meta Group in Stamford, Connecticut. “Knowing the customer is everything. You want to understand who they are and what is their relative value to the enterprise, not just here and now but into the future. Then you tailor the appropriate level of service or customer experience.”
But Smith from Aberdeen warns that “appropriate customer experience” should not, in the case of low-end customers, translate into inferior service. “It goes back to the primary question — is the segmentation strategy predicated on serving the customer better, or just corporate interests and the bottom line?” she says. “When you say you’re improving customer experience, are you really? Technology is not a panacea; it must be managed.”
Even among high-end customers, continues Smith, “there are cases where a husband and wife each have checking accounts at a bank, and even though the wife is the actual breadwinner, she’s kept on the phone punching digits because the husband is perceived as the decision maker in the gold bracket.” Adds Smith: “That’s why it’s incumbent to household the entire family — you don’t want to distance one member. I’ve seen five-member families each barraged with the same marketing material in the mail on the same day. Talk about a ‘customer experience.’ “
The Pared-Down Approach
Fireman’s Fund Insurance in Novato, California, evaded many of these concerns in its customer segmentation strategy. The solution: Keep it focused on sales and marketing, and not on service. The insurer uses data warehousing software provided by IBM Global Services and data analytical tools from Cary, North Carolina-based SAS to identify the best customers in each of its markets. “Knowing who our best customers are on a niche basis, and then householding them, lets us know how best to service their insurance needs,” says Michael DeVoe, Fireman’s Fund senior director of customer research and strategies.
Adds DeVoe: “We discern the number of insurance policies in a particular household and then, using internal data and external demographic data, we create a ‘life-stage segmentation.’ As these customers age or move into larger homes, we know the products best-suited to cross-sell them, or the loss-control advice they may need.”
Segmentation strategies give insurers an opportunity “to do something with all this customer information we have,” continues DeVoe. “Historically, it’s been very cumbersome to get our hands around this data because, like other large financial services companies, we tend to be structured along silos and have legacy systems that are not integrated. We’ve got three Ph.D.’s here now who are highly analytical people who do nothing but integrate all this data using the various tools. It’s a minor investment for something that is generating millions in profit, simply by tailoring marketing to customers individually.”
Or as the Bard wrote in another context, “Suit the word to the action, and the action to the word.”
Russ Banham is a contributing editor at CFO.com.