Real-Time Sales
By Chris Costanzo
By tracking transaction patterns, real-time marketing improves sales guidance. But will reps use the tools properly?
In an ideal world, banks would be represented by the very best salespeople, like the smooth personalities who sell Jaguars and BMWs. It doesn't make financial sense to deploy a highly compensated sales force to sell mass-market products, however, so financial institutions largely rely on branch and call center reps, who are often encumbered with service and administrative duties.
How, then, do banks enhance customer responsiveness without busting their budgets? Technology may yet provide the answer. Even though full-blown customer relationship management systems have disappointed, a more focused and pragmatic application, loosely termed "real-time marketing," is beginning to prove itself on the front lines.
Real-time marketing systems analyze transactions to identify so-called trigger events — signals that customers may be poised to increase or decrease their business with the bank. Once such alerts pop up, the systems go to work formulating responses that the reps can use, drawing on a blend of demographics, historical transaction data and other information. As customers accept and decline offers, the systems continually refine their prescriptions.
The potential payoff is huge. Gartner Inc., the Stamford, Conn.-based research firm, says campaigns that incorporate trigger events into offers made on a monthly basis generate response rates ranging from 4% to 5%, which rises to between 16% and 50% when the system can respond to customer activity triggers every day. This compares with paltry response rates of between 2.3% and 3.3% for traditional telemarketing campaigns.
However, such success is not guaranteed. While generating more relevant and timely sales leads definitely helps, getting the human sales force to take full advantage remains a challenge. Reps may not be properly motivated or trained. Necessary support systems — for prioritizing leads, fulfilling sales, or creating sales scripts — may be lacking. "Some banks are getting great results, some are not," says Gartner research director Kimberly L. Collins. "A lot depends on execution."
Given the setbacks that banks have encountered with CRM systems, which were supposed to electronically supplement virtually the entire customer experience but often failed to pay off, institutions will need to make sure they address the full range of factors influencing the success of real-time marketing. Fortunately, the tighter focus of these systems makes them more manageable, and early results are encouraging.
Rapid Response
In its purest form, real-time processing means responding to events nearly instantaneously, as they occur. The most prevalent applications in financial services have been defensive in nature, such as detecting fraud attempts and hedging other forms of risk.
During the height of dot-com mania, it was thought that electronic real-time response capabilities would be a great match with online banking, but the costs and complexity were way out of line with the revenue potential.
Then marketers noticed how information systems could track customer transactions in a variety of channels, including telephone voice response units, the Internet, branches and automated teller machines. The industry as a whole also affirmed that human interaction — not electronic interfaces — would continue as the anchor of customer relationships.
Putting all of this together, institutions began blending fresh transaction data and customer analytics into tools that customer representatives could use to influence situations where business might be imminently won or lost. "Time is of the essence when you've got a relevant offer," says Ann Christensen, a former senior executive at FleetBoston Financial Corp. who is now a consultant at Christensen, Williams & Associates of Andover, Mass.
Typically, real-time marketing systems filter transaction behavior, flag special situations and develop recommendations. Some institutions are capable of funneling such intelligence to reps in the midst of conversations. Customers who dial in to the call center, for example, might be identified as good prospects for a particular product.
In this scenario, reps would gauge the course of the conversation and the customer's receptivity, and maybe even enter into the analytical system additional information about the customer's financial situation gleaned from the current phone call. The real-time system would then spit out a product recommendation that, because it is based on the most up-to-date information available, has a much higher chance of resonating with the customer.
That's the ideal, but in most cases, there is at least an overnight lag in getting the data to the field. To combat customer attrition, for example, a bank would flag actions that indicate a customer might pull an account. Overnight, it would assemble all this information into the form of retention leads for reps to follow up on the next day. In this approach, real-time information is part of the equation for developing leads that get acted upon later.
This strategy is an improvement over the traditional practice of making predictions of customer behavior based on what similar customers already have done. But technically, it is not real time. "How 'real time' do you have to be?" asks Collins. "The correct term may be 'right time.' It's still quicker than what banks are doing today."
Most banks are still getting used to the concept. In a 2002 study, Gartner found that less than one-quarter of financial services companies (banks, brokerages, credit card and insurance) were able to monitor customer transactions for events that will trigger marketing contacts. Within this group, half of the institutions (one-eighth of the total) responded to customer triggers on a monthly basis, and only 8% (one-fiftieth of the total) did so in true real time. One-third of the total respondents said they have no plans to use event triggering soon.
Retention Play
While selling obviously is a big focus of real-time marketing, customer retention also is a prominent application. Fifth Third Bancorp of Cincinnati, for example, knew that for every 10 accounts it opened, nine were closing somewhere within the established customer base.
The $84 billion-asset bank had tried a number of approaches to stem attrition, says John Zugschwert, a vice president and the manager of marketing information systems. One initiative sought to identify customers who were likely to leave, perhaps because they only had one product with the bank.
While this approach helped to identify the right customers to contact, it did not make recommendations on the types of messages or offers that reps could use to dissuade potential defectors, nor did it offer guidance on when to contact clients. "The problem was that the account might be flagged as being erosive, but you didn't know when," Zugschwert says.
Fifth Third also tried what it calls the "insulation strategy." This involved identifying the most valuable customers and making them feel so special through regular calls and other personal service that they would never want to leave. The problem with this approach was that it treated all the high-value customers the same, without identifying those on the verge of leaving. "It worked, but we didn't get dramatic results," Zugschwert says.
In October 2001, Fifth Third began piloting a system that analyzes each customer's transaction patterns every day. The system, provided by San Antonio-based Harte-Hanks Inc., incorporates more than 90 business rules, which compare daily transactions with the customer's history and pinpoint telling changes in behavior. When the system signals that an account might be closed, it develops appropriate retention leads. But Fifth Third realized it needed to do more than just generate leads. "The key is execution," Zugschwert says. "This is sales process execution."
Toward that end, Fifth Third internally developed a system to deliver the leads to reps in the branches and call centers. The system prioritizes the leads based on either the profitability of the customer or the profit potential of the sales opportunity. The branches get the hottest leads — no more than 10 per branch per day. "We generate leads every day and we expect customers to be called on the day they're delivered," Zugschwert says.
To ensure follow-through, Fifth Third's system tracks whether follow-up calls were made and logs customer responses. Every week it reports results to senior management and every month it examines its effectiveness in expanding relationships.
The ability to respond to customer events that happened only the day before has helped. Within the first two months of the pilot program, Fifth Third had achieved full payback on its system costs. By the end of the six-month pilot, it had achieved a 400% return on investment, according to Zugschwert. Six months after it officially rolled out the system, Fifth Third had cut new-account attrition by half and reduced overall household attrition by nearly a third.
Transaction Tripwires
As bank marketers know, the insidious corollary to outright defection is diminishment, where customers retain accounts but curtail their usage and balances. This is the problem that First Tennessee National Corp. of Memphis will try to address using real-time marketing, says Suzanne Copeland, a vice president in the bank's creative solutions group.
Previously, the $25 billion-asset First Tennessee had built models to help it predict which customers were likely to move business away from the bank. These customers were targeted in direct mail and telemarketing campaigns, as well as with special incentives.
Echoing Zugschwert at Fifth Third, Copeland says models that predict certain customer behaviors based on static information do help, but that blending fresh transaction data into the mix makes the models more accurate and actionable. "It's more about reacting the moment an action takes place, versus using predictive tools that show it might take place," she says.
In a new approach that was rolled out in November, First Tennessee generates leads based on just-completed customer transactions and distributes those leads through an in-house system. Reps receiving the leads will be expected to follow up on them within 24 to 48 hours, Copeland says. The company has set a goal for the new system of reducing diminishment-related revenue shortfalls by 5% annually, beginning in year one.
Fifth Third and First Tennessee are among a handful of banks that are either using or planning to use real-time transaction information to hone next-day sales pitches. Salespeople control these campaigns by deciding exactly when and how to contact customers. The next level of real-time marketing is to initiate pitches on the fly, as customers randomly call or visit the bank. Even fewer banks are prepared to do this.
One that is trying is ING Direct USA, the direct banking arm of Amsterdam-based ING Group. The Wilmington, Del.-based bank, with $12 billion of assets, has been working to beef up inbound marketing — or selling to customers when they initiate the contact — since its inception three years ago.
The effort is especially important to ING Direct, since its policy is to restrict marketing to only those customers who specifically request this activity. "We get the customer's permission," says David Lewis, the chief marketing and information technology officer. Most other institutions market to customers unless they ask them not to, or "opt out."
So far, only 13% of ING Direct's 1.5 million customers have opted to receive outbound offers, to which the bank has responded with two mass e-mails. The rest of the time, it engages in subtler marketing efforts, such as posting banner ads on its Web site or querying dial-in customers on their interest in a new product.
Real-time marketing is seen as a way to avoid irritating customers with offers they don't need. "The key is relevance to the consumer," Lewis says. "That's where the value of real-time technology comes into play."
The system, which at ING Direct is an amalgamation of a number of software packages from different vendors, tracks every single customer contact, right up to the most recent one. ING Direct does not analyze these transactions on an individual basis, but rather in terms of groups of customers who behave in similar ways. It then develops offers for each group of similar customers.
In the call center at ING Direct, the system has helped reduce the number of sales pitches that fail, an important metric for the bank. "We wanted to decrease ineffective cross-sales as much as increase sales," Lewis says. "There's as much to be saved from not annoying the customer."
In addition, the unit has a rule that a live voice must answer 80% of calls within 20 seconds. So it uses its system to identify the most likely sales opportunities so reps can concentrate on those and not waste time on others.
ING Direct's system, which Lewis describes as "cost-effective," has paid off. The bank expected to reduce ineffective sales pitches by 10% and instead reduced them by 40%, Lewis says. At the same time, it boosted the percentage of inbound calls resulting in sales to 10%, up from 2% to 3%, he says.
Motivating Agents
Edinburgh, Scotland-based Halifax Bank, a subsidiary of $589-billion asset HBOS Group, likewise uses real-time technology to reduce the chance of bothering customers with offers they don't need or want. Halifax installed a system built by E.piphany Inc. of San Mateo, Calif., in its call center at the end of 2000 and has since expanded it to tellers throughout its network of 750 branches in the U.K. "Customers appreciate the fact that we're not asking them about things they already have," says Ruth Southern, Halifax's head of CRM development.
With the E.piphany system, call center agents receive relevant marketing messages as they are handling calls. The messages are based on customers' most recent interactions and are continuously adjusted as those behaviors change. They pop up on only about 20% of the calls and do not interfere with ongoing transactions, Southern says. Depending on how the conversation is going, agents can choose to make the pitch or pass it up.
Customers respond to about 10% of every 1,000 leads that the agents present, according to Southern. After these leads are transferred to the appropriate sales agents, about half the customers actually buy a product. Previously, Halifax had not passed any lead information from the call center to sales agents, so this volume is significant. All in all, Halifax recouped its investment in six months, Southern says.
At the same time, Southern cautions that "technology is only about 1% of the equation" when it comes to real-time marketing. Much more critical is motivating agents and salespeople to follow up on the leads. Southern recommends publicizing early successes as a way to overcome the resistance of call center agents to making pitches. Having salespeople notify and thank agents when one of their leads pans out into a sale also helps. "A lot is just about shaping and building confidence," Southern says.
Lewis of ING Direct agrees that sales agents must buy in to the process. "If the sales associates are not motivated, then the technology has no hope of working," he says.
This speaks to the larger issue of front-line execution, which looms as a make-or-break factor in most retail banking initiatives. If real-time marketing is to avoid the fate of CRM, then plans to install the technology must go hand-in-hand with plans for how to use it.
Ms. Costanzo is a freelance writer based in Brooklyn, N.Y.