The telephones at Capital One Financial Corp. ring more than one million times a week. People call to ask about their MasterCard balance, or whether a recent payment was received, or why their interest rate has jumped. And more than 1 million times a week, here's what happens -- before a caller hears the first ring:
The instant the last digit is punched, high-speed computers swing into action. Loaded with background information on one in seven U.S. households and with exhaustive data about how the company's millions of customers behave, the computers identify who is calling and predict the reason for the call. After reviewing 50 options for whom to notify, the computers pick the best option for each situation. The computers also pull and pass along about two dozen pieces of information about the person who is calling. They even predict what the caller might want to buy -- even though he or she isn't calling to buy anything -- and then they prepare the customer-service rep to sell that item, once the original reason for the call has been addressed.
All of these steps -- the incoming call, the data review, the analysis, the routing, and the recommending -- happen in just 100 milliseconds. That's one-tenth of a second, or one-eighth of the time that elapses between beats of a human heart.
Make no mistake: Capital One has some wicked-fast computers. And its growth rate has been nearly as fast. The company took shape in 1994, when it began as a spin-off of Signet Banking Corp. It now ranks among the 10 largest issuers of credit cards in the United States, with 16.7 million customers (it added 5 million new accounts last year alone) and total balances of $17.4 billion. Its stock chart looks more like that of an Internet startup than like that of a bank. Stock in the company, which was first offered at $16 per share, has traded as high as $140. The company, with 11,000 employees, has a market value of more than $7.8 billion.
Capital One's cofounders -- a pair of buddies with no previous hands-on experience in the banking industry -- had rocked the credit-card world while they were at Signet, where they had invented the "teaser rate" card and the "balance transfer" option. Those two innovations sucked millions of customers away from established companies, and the two men brought them over to their own company. The real secret of Capital One's success, though, has been its commitment to endless innovation. Lots of companies claim that they compete on knowledge. Capital One has enough information on consumers to fill the hard drives of more than 200,000 personal computers. It uses that information much as a physicist uses a particle accelerator: Cap One analysts and product managers come up with an idea for a product, bounce the data a bit, test it, tweak it, and launch it as fast as possible. In other words, they use the scientific method to design credit cards.
"Credit cards aren't banking -- they're information," declares Rich Fairbank, 48, chairman and CEO, whose father is a physicist. "When we started this company, we saw two revolutionary opportunities: We could use scientific methodology to help us make decisions, and we could use information technology to help us provide mass customization."
Says Nigel Morris, 40, Cap One's president and Fairbank's alter ego: "Very few companies have the ability to test and learn." But those twin capabilities -- testing and learning -- form the foundation on which Capital One is built. The company tests every product offering, every procedural change, every job applicant. It keeps records on every customer interaction and on every card purchase, and, with the patience and persistence of a good scientist, it runs experiment after experiment. "For every action we've taken," says Jim Donehey, 50, Capital One's chief information officer, "we know what the reaction has been. If we sent out a blue envelope and a white envelope, we know which envelope went to which customer -- and we've recorded what the reaction was in each case."
For example, Capital One started selling things (insurance, long-distance service, buying-club memberships) to customers who called -- after tests showed that people prefer to buy things when they call Capital One, rather than when Capital One calls them. Today half of all new Cap One customers buy something (other than a credit card) within their first year as a customer.
Last year, the company performed 28,000 experiments -- 28,000 tests of new products, new advertising approaches, new markets, new business models. As a result, it can deliver the right product, at the right price, to the right customer, at the right time. It offers 6,000 kinds of credit cards, each with slightly different terms, requirements, and benefits, and each requiring a slightly different monthly statement. Some credit-card holders have a no-fee Mercedes-Benz affinity card with a credit line of $20,000. Others pay $29 a year for a card with just $200 worth of credit. Some have a credit card with a Canadian moose on it. Others have a card with a map of Japan and an image of Mt. Fuji on it. One reason why Cap One has attracted millions of customers is that it's able to present itself a little differently to each of those customers.
"What we've done," says Fairbank, "is to create an innovation machine." That machine has just begun to kick into gear. "This is not just a credit-card strategy," Fairbank insists. "This is a marketing revolution that can be applied to many businesses."
For Fairbank and Morris, the credit-card business has been a grand experiment in using information technology to figure out what people want to buy and how Capital One can sell it to them. Says Jory Berson, 28, vice president of marketing and analysis at Cap One: "I want us to become the place where people go to find anything they want to buy."
"We can answer your question before you ask it."
Joe from Florida is a tough call -- a confused older man, the holder of a MasterCard with a $100 credit limit. He informs his phone rep, Tim Gorman, 29, that he knows what happened the last time he called Capital One: "I've got it written on the wall right next to the phone," he says.
Nancy from North Carolina is easier. A moment after her voice materializes in Gorman's earpiece, information about her credit card pops up on his computer. Nancy is a desirable customer: She has a $6,500 credit limit on her Visa, she's been a Capital One customer for four years, she pays no annual fee, and she hasn't paid late in the past 12 months.
Her opening line: "I want to close my account."
Gorman is used to that greeting. He works as a Capital One "retention specialist," and all of the calls routed to him are from people who say they want to close their account.
Gorman knows better. Most of them just want a better deal.
"I just got a card with a 9.9% interest rate," says Nancy, "and that's a lot lower than my current rate."
Nancy's current rate is 16.9%. Gorman's computer displays three counteroffers that he can make in an effort to keep her business. Those offers have different interest rates -- starting with 12.9% and ramping down to 9.9%.
"Well, ma'am, I could lower your rate to 12.9% . . ."
Nancy grabs the first deal that Gorman offers, and Gorman scores a "save." Is he concerned that Nancy didn't get the best deal available? Not really. "Who's to say she really was offered a 9.9% card?" he says. "Maybe that was just something that she saw or something that came in the mail." Nancy is happy, Capital One is happy, and Gorman has earned incentive pay for keeping her -- plus a little extra for getting her to take a deal that preserves much of Capital One's revenue.
Superficially, Gorman's exchange with Nancy sounds like what happens at credit-card companies everywhere. It's standard these days for account information to pop up when a call comes in (computers identify the account by linking it to a phone number), and many companies bargain over interest rates and fees. But at Capital One, that's where the interaction begins, not where it ends. The history of "intelligent call routing" (a term that Capital One coined) -- with its tiers of computer-enabled decision making -- is a case study in how a culture of innovation can take small problems or modest ideas and turn them into breakthroughs.
In this case, the problem had to do with the phone bill. Two years ago, calls from Capital One customers were taking too long to handle, and an analysis showed that customers were not to blame. Callers simply weren't getting to the right place at the right moment: People with a lost card or a fraud problem ended up reaching an ordinary rep; people who just wanted to know their balance stayed on hold to talk to a live rep; people unhappy with their interest rate called the "lost card" number on the back of their card and had to be transferred to customer service.
"All that time -- to take a call, to bridge the call to the right person -- that pisses off the customer," says Greg Gannon, 44, a technology manager who was brought in to look at the problem. "You wait for an agent, you wait for a transfer, you wait again for an agent."
Plus, says Jim Donehey, "All that time, we're paying for the call." Companies pay by the second, and at a place like Capital One, which takes 2.5 million live calls a month, even one extra second per call adds up to real money. "The real question was, How can we minimize all that?" says Donehey.
The company tried lots of options. For example, it discovered that some people called much more often than the average of five times a year. "We sent out a letter at one point that said, in effect, 'Please don't call so much,' " says Donehey. That test yielded this result: If you want people to call you, send them a letter telling them not to. "That was what I call 'a Homer Simpson moment.' "
The ultimate solution -- suggested by the technology folks -- had less to do with an animated character than with artificial intelligence: Why not predict the reason for each call and then send that call to the agent who is best able to handle it? In fact, a company like Capital One, with 3,000 people answering calls, knows plenty about why customers reach for the phone: Ninety percent of all calls fall into 1 of 10 categories. Raise your customers' interest rate, and they call. Put an automatic fraud stop on their card, and they call. Send out a new card that needs to be activated, and they call. Some people call once a month to find out their balance; some people call three times a month.
It took months to analyze these calling patterns, to figure out which events in a customer's credit-card life would predict a call. Decision-tree software had to be written, equipment had to be bought, phone switches and computers had to be taught to talk to each other. Along the way, people working on the project arrived at some powerful insights. For example, customers who call each month to check their balance or to see whether their payment has arrived could be identified and routed to an automated system that answers the phone this way: "The amount now due on your account is $364.27. If you have a billing question, press 1. . . . " Or: "Your last payment was received on February 9. If you need to speak with a customer-service representative, press 1. . . . " The genius of this system, says Donehey, is that "we can answer your question before you ask it." A phone call that might have taken 20 or 30 seconds, or even a minute, now lasts just 10 seconds. Everyone wins.
A routine problem in search of a quick solution led to a new way of doing business. Today customers get where they are going immediately, and they get the information that they need quickly. Customer-service reps handle those calls that need to be handled by people, and they don't waste any time passing calls to colleagues. For example, customers who have received prototype products are automatically routed to staffers who know about those products.
This calling system has become a competitive advantage for Capital One, making possible both lower costs and better service, and the company is understandably reluctant to share specific performance data about it. But a few round figures are available: The system went into effect last June, and, says Donehey, "Within a few months, it was right 40% of the time." Gannon says the prediction-success rate is now at 60% to 70%.
And the system just keeps getting smarter. Do you routinely call from your boyfriend's phone -- the number for which is not on file at Capital One? Eventually the computer will figure out that his number should be associated with your account. If you call to close your account, you'll encounter a subtle measure of what Cap One thinks of your business -- because the system will do a real-time analysis of your value as a customer. People worth keeping are routed to a live agent like Tim Gorman. People whom Capital One is not unhappy to be rid of are routed to a voice-response unit and allowed to close their account using their Touch-Tone phone. Meanwhile, more capabilities are in development. For example: The system will learn what language each customer prefers to do business in and then route calls accordingly.
"This started as such a simple problem," says Donehey, "but IT enabled us to go back to the business side with a solution that went beyond solving that problem. The flow of information between IT and the business side and the customer is constant. A lot of the stuff that really makes our technology work has nothing to do with technology -- it has to do with attitude."
"Every interaction is a selling opportunity."
Larry from Pennsylvania is on the line, baffled by his new Capital One "Silver" Visa, with its $1,000 limit and its $59 annual fee. The card comes with coupons offering discounts on hotels and rental cars, but Larry can't figure out how they work.
Barbara Brannon, 57, a veteran customer-service rep and one of Cap One's telephone stars, talks to Larry as if she has nothing else to do. She concedes that the Silver Visa carries a fee that's $20 higher than that of Larry's former card, a regular Visa. But, she says, "you can get that back with a single discount at a single hotel."
Once Brannon has helped Larry understand his new card, she says, "Sir, do you do any catalog shopping?"
"Are you interested in any kind of interior-design shopping?"
"Catalogs?" Larry asks.
"It's like the Silver Visa," says Brannon. "It's a service that we offer -- Capital One Interiors. It's a catalog with a lot of coupons. If you order it, you'll pay for it by the second or third time you use it."
"Thanks just the same," says a suddenly lucid Larry. "I get enough junk mail already."
Brannon is a star -- not just because she understands how to handle routine credit-card confusion, but because she is an unerring saleswoman as well. After all, why not try to sell a guy who is confused about coupons more coupons?
On a typical day, Brannon sells something to roughly 15% of the customers who are routed to her. Using an analysis of each customer's buying habits and demographics, her computer terminal automatically tells her which products to offer. Whether Brannon makes an offer on a particular call is up to her. With customers who are particularly upset or who are being transferred to another department, she doesn't. On this winter evening, though, she's scoring a sale with one out of every three callers. "I don't shove it down their throats," says Brannon. "I throw it out if it's something that I think they might like."
A woman from Miami, who has called to find out if she has to pay her $39 annual fee every year, cheerfully signs up to pay $59.95 a year for Privacy Guard, a service that helps you check your credit report. "That's called a sale!" says Brannon. "That's called fattening your paycheck. Give me more calls!"
Esther from Connecticut, who had called to check her current interest rate, signs up to pay $59.99 for Capital One Edge, a program in which you receive a catalog of discounted merchandise as often as 12 times a year.
Carolyn from Illinois, whose credit history has won her a Visa with a $200 limit and a $59 annual fee, calls to find out how to get a cash advance. After answering that question, Brannon warns Carolyn that her Visa has less than $100 of credit available. Then she offers Privacy Guard to Carolyn. The cost: $59.95. "If you want to get your credit report, this is the way to do it," Brannon says. "This is what you need." (Actually, if your Visa has a $200 limit and a $59 annual fee, you don't really need a credit report.) Says Carolyn: "Yes, I would like to try that."
No one at Capital One can recall whose idea it was to sell things to customers when they call the company, rather than the other way around. Most credit-card companies, including Capital One, have long tried to "cross-sell" their customers -- often by using inserts in monthly statements to tout everything from calculators to cruises.
The idea to add a new twist came, naturally, through data analysis. "It happened three or four years ago," says Nigel Morris. "We were doing outbound telemarketing, calling at dinner time, and we looked at the statistics on who bought things, who told us to get lost, and who bought things that they never used. We thought, 'This isn't working that well. What if we talk to people when they call us?' "
For Morris, a transplanted Brit who vibrates with energy and who has a mischievously contrarian streak, it seemed like a natural: "Gosh, if you call me and I'm trying to sell you something, then I'm going to treat you very nicely. That's going to promote better service. But people said, 'Service people are not salespeople. The systems don't exist to do this -- to service people and sell them at the same time. And anyway, even if you could do it,' they said, 'you would never sell enough to make the longer phone calls worthwhile.' People said it couldn't be done. That's all the motivation we needed."
In the case of inbound cross-selling, the first test product was easy and sweet. New customers must call an automated line to activate their card. "We thought, 'That's the perfect time to sell them something,' " says Jory Berson, who is in charge of cross-sell opportunities at Capital One. "The first thing we sold was a balance transfer: 'Now that you've got our card, is there any debt that you want to transfer to us?' " While customers waited on the line for a computer to verify and activate their card, they could sign up to move balances from other cards onto their new account.
Customers loved it. Or, at least, they bought it.
The response was so enthusiastic, relative to Cap One's effort, that Berson began looking for other products that Capital One could sell in this way. "I don't have a big desire to sell you groceries, because once I have to ship stuff, I can't save you any money," he says. "But auto insurance -- I can sell that cheaper. Mortgages, long-distance service, auto loans, cellular service -- same thing."
Berson sat down with his partners -- Eric Nelson, 30, Cap One's IT director, and Marge Connelly, 37, who runs all of the company's call centers. At their first meeting, in the wake of the balance-transfer experiment, "the brainstorming was less about 'Is this a good idea?' than about 'How do we do it as quickly as possible?' We weren't trying to pretend that this is rocket science," says Berson.
Indeed, the biggest hurdles were human. "At first blush," says Connelly, "folks look at the idea of selling to someone who's calling, and they say, 'There's a conflict between servicing and selling.' Some associates said, 'I don't feel good about making this offer to customers.' We were listening, trying to appreciate that feeling."
The solution, says Connelly, was "to elevate this beyond the immediate transaction level. If I'm a phone associate, my mission is to meet my customer's needs. And people have a pretty wide range of needs. If I've got this great product, it might save a customer some money, or it might create convenience. If I'm committed to service, there's no way I'm not going to consider offering that product."
Within three months of that first brainstorming session, live reps started fielding calls during which they both serviced and sold. Today Cap One does almost every cross-selling effort in partnership with another company. It makes a deal to market a product: The Hartford provides the insurance, MCI provides the long-distance service, Damark International provides the catalog clubs.
"The creativity comes in how we do that," says Rich Fairbank. "We do three things that other companies in the industry don't do. First, every interaction is a selling opportunity -- although I don't want to give the impression that we can't wait to get through your problem before we try to sell you something. Second, it's done in an information-based way: What's sold and how it's sold are decided statistically, according to information that we have. And third, we are totally committed to selling products that are of very high value. I say this with passion, because it's easy for companies to trap themselves into selling low-value things."
(For his part, Jory Berson says that he "feels better" about the auto-insurance offering, the mortgage deal, and the long-distance service than he does about Privacy Guard or about the catalog memberships. But even the latter, he says, "are presented fairly, at a fair price -- usually a lower price than other issuers offer.")
What ultimately kicked Cap One's cross-selling into high gear was Greg Gannon's intelligent call-routing system. It was a perfect fit: The same computer that predicts why you're calling can also predict what product you might want to buy. In fact, where your call gets sent often has as much to do with what Cap One wants to sell you as it does with how you will get your question answered.
Customers who are judged most likely to buy are sent by the computer to Cap One's most skilled sellers -- to Barbara Brannon's group. Every one of the three dozen non-credit-card products that Capital One offers has a statistical model behind it that outlines which kinds of customers will find it most appealing, and under what circumstances. A specific "product offer" comes to a customer-service rep in the same data burst as a caller's account information. So the decision about what to offer is made the instant someone calls. "It's got to be real-time," says Berson. "If normally I'd sell you long distance, but you're calling because you lost your card, I'd be a fool not to sell you credit-card registration."
Capital One doesn't reveal how much of its revenue or profit comes from non-credit-card sources. "It's a huge part of our business," says Berson. "We now make more than 1 million sales in a year through customer-service marketing -- not including inserts, or telemarketing, or automated calls." Those sales occur during the 30 million live calls that Capital One handles each year. In 1998, for the first time, half of all new Cap One customers bought another product from the company within 12 months of signing up for their credit card. "That's amazing penetration," says Berson, "and it leads to tremendous profitability."
"A cell-phone is just a credit card with an antenna."
When is a credit-card company not a credit-card company? That's the question now facing Capital One. If the company can serve its customers faster and smarter than its competitors can, if it can sell products unrelated to credit cards, then why continue to think of itself as a credit-card company at all?
According to cofounders Fairbank and Morris, Capital One's real capital resides in the hundreds of terabytes of data that the company has collected on the behavior of current and potential customers, and in the vast testing experience of its analysts, who know how to figure out what really matters to (and about) those customers. The company's knowledge base positions Capital One to tackle all kinds of other information-rich businesses.
Already, a Capital One subsidiary, America One, is selling cell-phone service in much the same way that the mother company offers credit cards -- by offering a slightly different deal to each type of customer. (Quips Morris: "A cell-phone is just a credit card with an antenna.") Cap One doesn't confuse its customers with choices. Instead, it tries to figure out what its customers will want. "Our secret," says Fairbank, "is that we are trying to deliver very simple solutions to the customer."
Fairbank claims that within two or three years, most people will know Capital One for reasons other than its credit-card business. "Fifty percent of what we're marketing now did not exist at this company six months ago," he says. "And 95% of what we're marketing today didn't exist two years ago. I'm proud of that fact -- until I reflect on its implications. It means that 50% of what we'll be selling six months from now doesn't exist yet."
Says Jory Berson: "We're becoming a company that people can turn to for just about anything. When you get ready to buy a car, the first thing I want to go through your mind is, What kind of deal can Capital One get me on an auto loan or on auto insurance?"
What about the car itself? "Oh, sure," he says. "A car-buying service is on my list."
Sidebar: What's the Hard Part? Innovation Never Ends
Capital One scored some early victories in the credit-card game by changing the rules. Its founders invented the "teaser rate" -- the banking equivalent of a cheap mortgage that gets adjusted up after a certain period. The teaser attracted millions of customers, and eventually competitors began to mimic the idea. The result was a merry-go-round of so-called teaser hoppers. People would move their balance to a low-introductory-rate card; as soon as the rate expired, they would switch to another card. Over time, Cap One was being eaten alive because of customer attrition.
The lesson: Companies that thrive on change can never rest. Competitors are quick to copy good ideas. The way to compete on innovation is to keep innovating.
That's the hard part. And that's why Capital One is so obsessed with making its own innovations obsolete. Consider the teaser rate. "Customers have gotten smarter about this game," says Dan Friedman, 29, a credit-card product manager. "But customers don't want to play it." What do they want? People want a simple card with a good rate, says Friedman. So, for two and a half years, Friedman's team offered prototype no-fee, fixed-rate cards to a wide spectrum of Americans. The cards had slightly different rates, slightly different credit limits, slightly different terms: "We were doing at least a dozen tests each quarter."
The team absorbed data about which rates appealed to which kinds of customers, about how customers used the new cards, and about which set of factors produced the most profitable card. "We found a rate at which we got good loyalty and low attrition," says Friedman. "That's what we wanted." There was just one problem: The card wasn't profitable.
"That is where smart people come into play," says Friedman. Two of Friedman's analysts concluded that they needed to attract customers who were better credit risks -- in other words, people who defaulted less often. "They suggested different underwriting standards. As soon as we had that hypothesis, we could test it with data that we already had."
The result of all this experimentation: Early in 1998, Capital One began offering a permanent, fixed-rate card to a select group of customers. The card has a fixed annual rate of 9.9% and no annual fee. The company that changed the game with the teaser rate is changing it again. Today, in fact, Cap One doesn't even offer a teaser-rate card to new customers.
Sidebar: Testing, Testing
At Capital One, the essence of innovation is experimentation. Every idea for a new product starts with a test, or a series of tests. If a test succeeds, the company ramps up fast.
Each test offer is handled by a small group of reps, whose opinions about the products are also taken into account. Cap One tracks not only whether people buy something but also whether they use it. The company has shelved products that customers didn't buy or didn't use. Selling software didn't work. Neither did selling mutual funds (although plans are under way to launch another mutual-funds sales program). Behind every test is a three-person team: someone from marketing, someone from operations, someone from IT. There are no silos at Capital One. Each business unit has a "business information officer" who is responsible for making technology work in that unit.
Testing doesn't apply only to products. Cap One hired more than 100 people a week in 1998, but it doesn't conduct anything resembling routine job interviews. Instead, applicants take tests, sit through guided interviews, and do role-playing exercises. In the case of customer-service reps, supervisors never even meet candidates; they meet only the people who get hired. "We found out that interviews by line managers didn't predict job performance at all," says Dennis Liberson, 43, the company's HR chief. That discovery was made, of course, through follow-up testing -- through comparing actual performance of employees with what their supervisors had predicted when they were first hired.
This year, 1,000 of Capital One's 11,000 employees will sit down to take tests. HR has developed 10 new pre-employment tests, and it wants to see how good they are. It will administer those tests to selected current employees, who will in turn have their job performance evaluated. Test results will then be anonymously correlated with how well those employees do their work. "We think these tests will predict certain competencies," says Liberson. "But before we start using them, we want to make sure."
In other words, the testers have a new test that they need to test. The company's testing culture is so deeply intricate, so wound around itself, that a senior executive recently pointed to an M.C. Escher print in his office -- a print of people climbing or descending a series of interlocking staircases -- and laughed: "That's our organizational chart."
Capital One sometimes thinks so hard about what it does -- and tests its ideas so thoroughly -- that the whole system can make an outsider's head hurt. But most people at Capital One will tell you that testing does not give them a headache: Rather, it opens their minds to new ideas. "Testing," declares Wendy Alexander, 29, international product manager, "liberates you from conventional thinking."
Charles Fishman (firstname.lastname@example.org), a Fast Company contributing editor, is based in Raleigh, North Carolina. He scored 40 out of 40 on Capital One's management-screening test. Even so, the company rejected his application for a no-fee, 9.9% MasterCard. Visit Capital One Financial Corp. on the Web (www.capitalone.com).