Less than a year ago, e-Loan and Mortgage.com could boast blazing initial public offerings and soaring stock prices; indeed, the shares of the latter more than doubled in value in the two months following its IPO. Dominant off-line mortgage companies such as Countrywide saw their stock prices plunge. But in the months since then, the World Wide Web hasn’t remade the way the mortgage business is conducted. For example, Mortgage.com announced in February that it was leaving the consumer market to focus on its business-to-business (B2B) counterpart. What happened?

During the early days of Web-based providers, companies like e-Loan and Mortgage.com imagined that consumers would go on-line, input how much they wanted to put down and to finance, and choose the type of financing they preferred—fixed rate, variable rate, or some other. As the applicants waited on-line, mortgage companies would check their credit. If everything was fine, consumers would have a mortgage commitment within moments of hitting "enter." The new channel would squeeze out costs and intermediaries, thus creating more value for consumers and profits for new entrants.

But this vision didn’t reflect a sophisticated understanding of either consumers or the way the contemporary mortgage market truly works. A careful look at the steps and players in the process shows why the complex, seemingly dysfunctional business of getting and creating a mortgage will remain a messy mix of on- and off-line, direct and intermediated transactions (Exhibit 1). A number of business models have the best chance of prevailing amid the clutter.

Stumbling blocks

In many ways, the mortgage market is suited to the Internet. It depends on timely information, since changes in pricing often have a large impact on borrowers. Its products, in their minds, are commodities. And the process of obtaining a mortgage is expensive and burdensome. Mortgages, however, are obtained only at seven- to ten-year intervals, and people generally want the assurance of human help, since a home is usually the biggest and most important purchase in their life.

The first business-to-consumer (B2C) Internet players (and their investors) soon realized that building a trusted consumer brand would be very expensive. Even generating an initial interaction with a customer proved problematic. In the third quarter of 1999, as a result, e-Loan’s marketing cost per loan came to $4,000 a unit; Countrywide, by comparison, paid only about $1,000 in total origination costs. Moreover, on-line lenders were especially hurt by the infrequency of mortgage originations, since, unlike their off-line counterparts, they did not have any "natural" hedges, such as mortgage servicing. The pure mortgage B2C business model therefore failed quickly. The pure consumer mortgage market was too narrow to support large Web-based enterprises and valuations as well as too cyclical because of its dependency on the movement of borrowing rates.

Furthermore, the mortgage process is really too complicated to take on-line easily. In the United States, the process for actually getting funds into borrowers’ hands varies by state: in some places, for example, you need a lawyer; in others, you don’t. And closing the loan requires gathering information from—and coordinating the activities of—more than 40 different parties, from the termite inspector to the title insurance company to the human-resources employee who can confirm the applicant’s place of employment. Transforming all these paper and phone interactions into a seamless process conducted entirely on the Web is no easy feat.

The consumer’s obvious desire to deal with real people when purchasing a mortgage (Exhibit 2), the infrequency of purchase, and the complexity of the process are not the only reasons for the relatively sparse prospects now facing Web players in the consumer mortgage business. Another problem is the fact that the business really isn’t quite as inefficient as it might seem. By now, the very well-established presence of Freddie Mac (the Federal Home Loan Mortgage Corporation) and Fannie Mae (the Federal National Mortgage Association) has made spreads much narrower than they were at one time.

Twenty years ago, those two government entities created a secondary market in housing loans—a market that has been chipping away at the inherently local nature of the mortgage business. It has been local in nature because the value of a given house depends on its location. Although Web players hope eventually to standardize the appraisal process, a lender based in one city now has little idea of the market value of a house in another, even when the lender knows its exact size. In short, local real-estate values, the character of the neighborhood, the quality of the fixtures in a house, its proximity to good schools and shopping, and myriad other considerations influence the purchase price and determine the kind of security the lender requires. Those factors help explain why, for most of the 20th century, local savings and loan (S&L) associations and savings banks, which also had the benefit of special tax treatment, did most of the mortgage lending in the United States. Such institutions, with their detailed knowledge of the seller, the property, and the buyer, had all of the information they needed at their disposal.

When Freddie Mac and Fannie Mae were created, 50 percent of all mortgages were originated and held by S&Ls; today, the figure is less than 20 percent. The existence of a secondary market has allowed many other players, including Internet-based ones, to enter the mortgage market. Internet players reasoned that if they could manage to acquire customers, they could rely on Fannie Mae and Freddie Mac to get their loans to the capital markets. The capital that the Internet players would otherwise have needed to support a full lending operation could instead be focused on brand building and customer acquisition.

But the national model is not yet a reality. The two organizations, it is true, do play some role—either as holders or as packagers and sellers to investors such as pension funds—in 40 percent of all mortgages outstanding and in a much greater share of new loans. But for mortgages to be securitized and then sold in bulk, like must be packaged with like. The culling process thus requires a great deal of specific knowledge of such things as regional tax policies, local crime rates, and the condition of local school systems. Traditionally, local and regional intermediaries pulled together most of this detailed information, a process that early B2C players found too difficult and costly to replicate.

There are other challenges. The mortgage market may be huge—$1.3 trillion and growing by 6 to 8 percent a year since the US Federal Reserve started keeping these statistics in 1953. But the rate of mortgage originations, being dependent on interest rates and economic conditions, is very volatile in the short term. If interest rates rise, people put off purchasing homes or refinancing—as they did in the fourth quarter of 1999, when mortgage applications in the United States plummeted by 26 percent.

Fundamentally, any mortgage business, old or new, is competing for a slice of the profits generated over the life of the loan, including up-front origination fees, the interest spread between the borrowers and the loan holders, and the servicing fees. Given the highly competitive nature of the mortgage business, many of these margins are already quite narrow. For example, the interest differential on a $175,000, 30-year, fixed-rate mortgage loan is $3,500. That is down from an average, 20 years ago, of about $7,000 for a mortgage of the same size and duration. In all likelihood, the transparency provided by the Internet will narrow the differential further.

One kind of mortgage may remain lucrative: so-called jumbo loans of $240,000 or more, which is well above the average in all but a few areas of the United States. Government regulations exclude Fannie Mae and Freddie Mac from these markets.

Four workable strategies

Despite these difficulties, and the fact that fewer than 2 percent of all mortgages were generated on the Internet last year, many market analysts predict that on-line volumes will grow to 15 to 20 percent of the total market by 2005. (These figures don’t reflect the much larger share of consumers and intermediaries using the Internet for a portion of the process.) And there will be opportunities arising from the way the intermediaries that will control the remaining 80 to 85 percent of the market interact and from the way the settlement process evolves.

So though new Web-based players have less opportunity than some of them, or the stock market, once expected, there is still room to bring the Web’s advantages to bear on the mortgage industry. Already, 70 percent of all realtors use the Internet to post listings. Brokers want tools such as reverse auctions and search capabilities for locating loans with the most favorable terms, as well as software that can generate terms on the basis of a borrower’s qualifications. Thus armed, the brokers, or a new set of real-estate and consumer finance agents, would be able to deliver more value to their customers. The four most promising models for doing so follow.

1. Owning the mortgage product

Although more than 60 percent of new mortgages were securitized and sold to public investors through the federal lending agencies last year, the remainder, worth $3.6 trillion, remains on the balance sheets of the banks, the S&Ls, and other financial institutions. As always, whoever can economically underwrite a loan that has the features consumers want will dominate the market. But in the future, doing this will be much more difficult because the emergence of better tools for comparison shopping at both the consumer and the intermediary levels will eventually create a market with only a handful of very large contenders.

To profit by holding on to a mortgage, a lender will need top-notch pricing and underwriting skills, outstanding product innovation capabilities, and access to deep, cheap, and flexible funding sources. Because of Fannie Mae’s and Freddie Mac’s role as government-sponsored entities, they will retain their advantage over non-agency lenders, except in the jumbo and subprime mortgage markets, in which they don’t lend at all.

2. Providing the tools

Mortgage.com and e-Loan originally intended to serve the consumer market by seamlessly connecting it to mortgage providers. Although the prohibitive cost of acquiring customers vitiated that model, it soon became apparent that the technology the two companies had been developing to realize it was equally applicable to B2B commerce. That technology could therefore be turned into a product and sold to other industry players, such as mortgage brokers and small banks.

Even the big players have recognized the opportunity in B2B tools. In March, for example, Freddie Mac, GMAC (the financing arm of General Motors), Microsoft, and several banks (Bank of America, Chase Manhattan, and Wells Fargo) announced a joint venture that will provide an Internet platform for real-estate agents and other intermediaries in the mortgage process. Embedded in the platform will be Freddie Mac’s on-line loan-underwriting software. Other opportunities of this sort include auctions and exchanges where lenders bid for brokers’ business; systems for collecting, updating, and standardizing data on local crime rates, public schools, and recent housing sales; and tools, such as IndyMac’s LoanWorks service, that help borrowers and intermediaries find alternatives to the origi-nal loan choice.

The transaction frequency of the B2B model leads to lower costs per transaction, greater comfort on the users’ part, and wider choice

Because upward of $800 billion in mortgages still involves human intermediaries, supporting them with Web-based tools can generate real value. The B2B model is clearly more robust than the pure B2C one because the higher transaction frequency of the former leads to lower costs per transaction, greater comfort on the users’ part, and a wider range of choices (thanks to the larger number of suppliers the model attracts).

As the mortgage evaluation process becomes more uniform, attacks by specialists at particular points in the business system will create economies of scale. Eventually, off-the-shelf solutions will replace the functions (including on-line documentation, certification, and appraisal) that have been developed by integrated firms and specialists alike. Pure plays of this kind are just starting to be launched, but incumbents such as Countrywide and LendingTree have already developed impressive applications in-house that could be built into entirely new businesses and eventually spun out.

3. Owning the consumer’s balance sheet

An alternative to the mortgage specialist routes is the more ambitious goal of integrating the consumer’s entire balance sheet, of which mortgage finance is one of the largest single components. Thanks to the advent of the Internet, this kind of cross-selling is more feasible than it was in the past. The Net’s ability to aggregate the financial information of individuals makes it possible to fine-tune terms and pricing—using assets other than homes as collateral, scheduling payments to match consumers’ income flows, and so forth. The aggregation of all of a consumer’s financial information also permits its instantaneous transfer to a loan application, with no need for further verification.

The companies in the best position are not traditional mortgage specialists but rather the likes of Charles Schwab, Intuit, and major banks—even companies in other industries with access to a lot of information about customers. It isn’t hard to imagine the day when on-line agents with equivalent access to consumer information could provide real-time advice, such as the suggestion that it was time to refinance.

One of the most formidable barriers to the dream of limitless cross-selling has been the preference of most people for best-of-breed solutions and the widest possible array of choices. Players should therefore not only allow consumers to comparison shop but also direct them to the product that best meets their needs, even if it comes from another company. For many businesses, this will require a big mental shift. Presenting alternatives obviously poses a risk of losing the loan itself, but such companies can keep customers by providing better guidance than the customers themselves could obtain on the Internet.

4. Controlling the home ownership experience

A fourth workable strategy would be to bring together the mortgage and all of the other products and services needed to buy and maintain a home. Mortgages represent only 50 percent of the revenue generated by home purchases (Exhibit 3) and less than 29 percent of the revenue generated in the first two years of residency. Winners will need to offer a broad array of products as well as access to a number of high-quality loan providers. "Event-driven" sites have shown that they can transform a single transaction into a multifaceted purchase experience: Autobytel, for example, has expanded from car sales to car financing, and The Home Depot has started offering home equity loans. Other offerings, by Homestore.com and Move.com, for example, include liaison and coordination with moving companies and help with renovation plans.

Such businesses could expect to earn advertising, referral, and subscription fees in addition to some part of the mortgage origination fee. But many mortgage industry players are unlikely to adopt this strategy, because they lack marketing sophistication and the skills to bring all these disparate pieces together.

The on-line mortgage market is evolving rapidly and therefore demands that both incumbents and attackers define themselves strategically. To do so, incumbents need to ask themselves a number of questions: Which markets are best for us? How can we compete most effectively with the new breed of competitor? How can we make the benefits of integration tangible to consumers? Should the company simply extend its product set, or should it offer the maximum range of choice? Clearly, those companies that succeed will have blended the best of both the virtual and the traditional markets.

Attackers, unhindered by inherited infrastructures and procedures, should assess whether they can muster the financial resources to compete against more established players. If they can, they must decide whether the route to creating a distinctive and sustainable position is to focus on a few products and services or to assemble an integrated suite of them. With margins coming out of the business, the smallest and least efficient companies will need to capture the value remaining in their customer relationships and balance sheets if they are to survive.