Enterprise Value Award Loan Star

FRAMINGHAM (02/02/2000) - "Need a loan to finish building your new house? Have a seat, have a seat. Now, let's take a look at your credit. Hmm, if you borrow $50,000 at 8 percent, that puts you over your risk cap. Maybe we can reduce your total indebtedness by paying off those three credit card balances you carry. Then we'll recoup the expense in interest by adding a half-point to the house loan. Just checking the math.... No, that won't work. What if we try..."

Consumer loans are what you might call a pens and needles business: lots of uncomfortable moments waiting for approval-much of it watching the salesman (account executive, or AE) calculate and recalculate percentages and principals-followed by the ritual signing of a few reams of documentation.

Household Financial Corp. customers, however, have an easier time of it. They don't have to wait long for loan approvals, and their account execs don't need scrap paper. Household's Vision system integrates all phases of the process from lead to loan and connects to an intelligent underwriting engine that returns verdicts in minutes rather than the industry-standard hours or days.

Household offers auto loans and credit cards, including private label cards for businesses like Best Buy and General Motors, but its biggest business unit in terms of revenue is the consumer finance division, which does equity loans.

This industry is undergoing rapid consolidation. In mid-1998, for instance, Household acquired Beneficial Corp., one of its largest competitors, creating the HFC and Beneficial divisions with 1,400 branch offices dispersed among 46 states. The twin keys to competitiveness in this industry are efficiency and customer intimacy. That means selling and servicing the maximum number of financial products with the lowest operating overhead-all without making the customer feel like a number.

The Vision system has helped Household improve in both of those areas. Sales have increased by more than 10 percent per account exec. AEs spend less time on paperwork and administrative hassles and use the system's lead management functionality to provide customers with better-targeted product offers.

Household totes a five-year return on investment of 129 percent-and because of the system's neural network engine, its benefits are just beginning.

HISTORICALLY SPEAKING Consumer finance companies have historically been reluctant to invest in technology because loan products are complicated and subject to heavy regulatory restrictions. "Perfecting" an equity loan, the industry jargon for finishing all the steps necessary for regulatory compliance, requires the proper signatures on an imposing set of meticulously prepared documents (deed of trust, repayment and security agreements, and so on). The laws governing the industry differ by state, and in some cases even by county. Underwriters-who structure the terms of loan agreements-vary in the scope of deals they can approve. And countless variables play into the loan approval decision-the consumer's income, employment history, credit history, outstanding debts and the like. Capturing all that complexity in a software application is daunting.

But around 1994, Household's head of consumer finance, Bob Elliot (who has since retired), predicted some forthcoming changes in the industry, particularly the consolidation of many lenders into few. Elliot wanted Household to compete by offering faster underwriting decisions than its rivals did. If, for example, a Household customer needed a loan of such scope that the local branch office was not authorized to give final approval, all the papers of the proposed loan were shipped to another location, which meant the customer waited even longer. Clearly Household could get a leg up on the competition if the branch AE could use technology to put that application in front of a central pool of underwriters while the customer was still in the office.

Elliot's thoughts about the changing nature of the industry and his desire for greater efficiency were the starting point for the Vision system. At that time, Ken Harvey, now Household's CIO, was in charge of application development.

Harvey contemplated with some dread the prospect of building such a complicated system. "I thought, 'Gee, we're going to have to write this [system code] 59 times,'" Harvey says-roughly one version of the system per state, plus a few Canadian versions, to meet the local regulatory demands. The better solution, Harvey realized, was to create an object-oriented architecture for the system.

Every functional version shares a common foundation of objects, with custom tweaking done only to the necessary modules.

The challenge was that object-oriented development takes longer and therefore costs more. "Hard-coding a decision tree would have been cheaper and faster, no question," Harvey says. The price tag for the three-year initial implementation was $83 million. "At 2 cents per share [of Household stock], the investment certainly drew attention," says Harvey. That's a whopping technology investment in a paper-based business.

"I give [Chairman and CEO] Bill Aldinger credit. He realized we could leverage our investment much better" with the component-based architecture, Harvey says.

Harvey made the pitch, Aldinger bought in, and work commenced on the Vision system in late 1994. "It took courage to make the investments that were counterintuitive in this industry," agrees Enterprise Value Awards judge Patricia M. Wallington.

Harvey says Vision required three years from the inception of the development work to the full conversion of all functions. Rick Erickson, Household's director of advanced technology, led the development of the software components for the middleware and messaging system and Kathy Mikos, vice president of business development led the development of the system using Sybase's Powerbuilder client/server development toolset. In mid-1996, Harvey's team implemented the system in a handful of branch offices that served as test pilots. Deployment to all branches was completed in January 1997.

SELLING THE SYSTEM Building a system is only the first challenge, of course.

Someone has to use it, and sales departments in general have a history of cool receptions for technology-based selling tools.

That's where Tom Detelich came in. HFC and Beneficial produce the largest revenues of all Household International's operations. HFC's revenues come through Detelich's operations as managing director of sales for the HFC unit.

Detelich, a 23-year industry veteran, arrived at Household when the company acquired Transamerica in 1997, just in time for the rollout of Vision. He was quick to throw his support behind the project, which helped ensure buy-in from the sales force.

Detelich says two key benefits of Vision are immediately obvious. One is the reduced training time for new AEs. The second and more compelling benefit is that the system reduces AEs' administrative overhead. "Our folks spend nearly all of their time servicing customers," says Detelich. More selling time translates into more sales; when the test branches reported higher sales, the minimal resistance among the remaining AEs quickly evaporated. By mid-1999, Vision was handling an average of 11,000 new loan applications per day, amounting to 3 million transactions per day. Now, each time Household develops an update to the system, the company continues the practice of rolling out the changes to a few test branches first.

PAYBACK TIME Again, the name of the game in Household's industry is efficiency and customer intimacy. Household's branches are selling a minimum of 10 percent more loan products than before the system rollout. Branches added to the Household network via acquisition, such as the Beneficial offices, show even better gains-as much as 18 percent. Another key measure of corporate performance in consumer finance is the "managed basis efficiency ratio," which is calculated by this formula: expenses divided by revenue, minus policyholders' benefits. A lower ratio is better, and Household has lowered its efficiency ratio from more than 40 percent prior to rollout to less than 35 percent today. It all adds up to a 40 percent total return on investment, over Vision's first few years of life.

Household continues to tweak the system as regulations change and as AEs generate ideas for new functionality. "Almost all of the impetus for changes comes from the sales force," Detelich notes. Mikos, vice president of business systems, attends all divisional and staff meetings of the sales force to help keep her team abreast of changing needs and new ideas. And because of Vision's component-based architecture, only those objects being changed need an update; Household doesn't have to recompile the entire application. The company can replicate new objects to the branch offices overnight with the push of a button, once it has tested the changes in the pilot branches.

The most interesting thing about the Vision system is that its neural network components now promise benefits that even its developers didn't foresee. This is Vision's lasting value: As it learns, it helps the company make smarter decisions about Household customers.

For instance, say a credit card holder calls, irate about a late fee. He's not a profitable customer for the company: He carries a single card with little or no balance and has spurned Household offers for credit insurance products and equity loans. Why should the Household service rep cancel the late fee? Vision knows why. The system "takes into consideration the potential lifetime value of the customer," Harvey says.

Turns out this customer took out a school loan six years ago and a small auto finance deal for a used car three years ago from another company. His modest income has gone up significantly two years running. Considering these variables, Vision can recognize this late fee as a first offense by a recent college graduate who handles his finances well and may be in the market for significant new loans in the next year. Vision authorizes the service rep to waive the fee. Then the system can prompt the rep with suggestive selling for this now-happy customer-does he know that Household can pay off that old car loan and offer attractive terms on a loan for a newer vehicle?

"If you'd asked me about suggestive selling six months ago, I wouldn't have known anything about it," says Harvey. This feature is expected to create a higher level of value for the Vision system, which is important because Harvey, for one, believes the company is pretty close to offering the best efficiency it can deliver. "The next growth is top-line growth through lead management and suggestive selling," he says.

Taken in sum, the system ties the company more closely to existing and prospective customers. Loan approvals are faster, sales proposals more targeted and customer service more responsive. Cutting out the waiting game and creating more desirable products helps Household forge a customer intimacy that ultimately translates to profits, which in today's stock-market-driven environment is the ultimate in enterprise value.

What other consumer experiences-like awaiting loan approval-beg for IT-enabled reform? Executive Editor Derek Slater can be reached at dslater@cio.com.

ONE VISION

Household's IS team designed seven integrated modules to create a complete package Training All training for Vision is computer-based, and AEs must pass online tests to gain authority to process transactions. Courseware is updated one month before each new release of the Vision software.

Lead management Actions and inquiries by HFC and Beneficial's 3 million current customers yield what the company calls "perfect-timing leads." Say, for example, a customer pays off her car loan and tells the service representative she plans to keep the car another year. Bing, a year later that lead pops up at the local sales branch, prompting the AE to inquire whether the customer needs to finance a new vehicle. If the customer moves cross country, the lead is transferred to the office closest to the new address.

Solicitation Vision handles all contact management activity and creation of loan proposals.

Underwriting AEs can model financial solutions using the customer's current financial information. A neural network host looks at risk, interest rate and other variables in approving loan terms. The company says typical underwriting requests are handled within minutes rather than hours or days, standard in this industry.

Closing The system generates all necessary documents on the local laser printer, including, for example, checks to pay off other lenders in the event of debt consolidation.

Perfected product Every document is bar coded and forwarded to a central loan processing facility. There the documents are scanned for imaging. The system automatically initiates a commission for the right AE, but not until all required documents have been received and scanned. That provides further incentive for the AEs to get all of their ducks in a row.

Service Vision also handles all aspects of customer service, so if a loan customer calls two months after securing a loan, the service rep can view the same information the AE did at the time of the loan process. -D. Slater DEFINING VALUE Household's Vision system helps the company reduce loan approval time, simplify regulatory compliance and cross-sell new products. Deployed in January 1997, the application cost $83 million to build and has resulted in a 10 percent to 18 percent increase in sales per account executive. It also enables Household to assimilate acquisitions efficiently in the rapidly consolidating consumer finance industry.

ANATOMY OF THE SYSTEM Hardware Central processing is performed on an Amdahl Millennium mainframe and two redundant Sun Microsystems E10000 servers. Each local branch office hosts a smaller Sun server running Sybase's relational DBMS, a dual-sided laser printer (which is key in the heavily document-oriented insurance business), and Pentium-class PC workstations for individual account executives.

Software Vision is an object-oriented system built with Powerbuilder Foundation Class Development Tools, Electronic Forms Systems' software for form design and printing, and Asymmetric Icon Author. The messaging layer connecting to MVS environments is a homegrown Cobol system called the Global Transaction Manager.

The neural network engine for underwriting decisions is based on technology from Hector Neilson Corp. and Fair, Isaacs & Co.; Hector Neilson Corp. exclusively handles fraud detection. Strategyware from Fair, Isaacs & Co. is also the basis for the system's suggestive selling component.

Network Vision runs over a single virtual North American IP network provided primarily by AT&T and using Cisco routers and IBM's MQ Series messaging middleware. Network management is handled via Computer Associates' Unicenter TNG environment and monitoring tools from Candle Corp.

Join the newsletter!

Or

Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about AES EnvironmentalAmdahlAT&TCandle IT & T RecruitmentCA TechnologiesCross CountryHolden- General MotorsHousehold FinancialHousehold InternationalIBM AustraliaPowerBuilderSun MicrosystemsSybase AustraliaTransamericaUnicenter

Show Comments
[]