SOA: TD Banknorth Is Banking on It

By Mel Duvall  |  Posted 2006-10-02

A number of technologies have promised to reshape and revolutionize the financial services industry over the last five years. But John Petrey, chief information officer of TD Banknorth, a Portland, Maine-headquartered bank with 600 branches, didn't need a revolution. He needed an answer to an old, familiar problem—integrating and getting more value out of existing systems.

And he's far from alone. While a number of promising technologies, such as banking cards with smart chips or cash cards with radio frequency identification (RFID) transceivers, have fizzled, one technology appears to be making a dramatic impact in the financial services sector: service-oriented architecture (SOA). Banking giants such as Bank of America and Wachovia, as well as other financial services leaders like Merrill Lynch and Charles Schwab, are finding that SOA is not only solving many of the challenges they face today, but is a technology they can bank on in the future.

View the PDF -- Turn off pop-up blockers!

There are varying definitions for SOA in the industry, but the overriding premise is that SOA is a computing architecture that allows an enterprise to make its applications and computing resources, such as databases, available as "services" that can be called upon when necessary. SOA leverages standard mechanisms, such as eXtensible Markup Language (XML), to simplify the process of exchanging data.

In an industry like financial services, it means a bank can take an existing legacy application, such as a program for compiling client account statements, and make that application available as a service that can be accessed by other applications. If the bank wants to make that "client account statement" service available in an online banking application, for example, it taps into the existing service rather than create a new application. Enterprises can create entire libraries of such services, which can then be used like electronic building blocks to speed application development.

The industry has particularly taken to SOA because it has a huge investment in battle-hardened legacy applications, and because a flurry of mergers and acquisitions has created major integration headaches.

"When you invest in any given technology, you are essentially placing a bet," says Banknorth's Petrey. "You're betting that the technology will be a survivor and that it will deliver a payoff to the business. My conclusion [after working with SOA for three years] is that SOA is absolutely a technology worth betting on."

NEXT PAGE: Branching Out

Branching Out

The same can't be said for many other technology bets being placed when Baseline was launched five years ago.

Much of the thinking at the time was based on the premise that banking, retail banking in particular, was about to go through a significant consolidation phase. Banks would be merging—which, to a large extent, they have—and some banks would shutter branches to save costs. Banks would instead make increasing use of new technologies such as more advanced automated teller machines, telephone banking systems and Internet banking applications to steer customers away from expensive branches and their employees.

Meanwhile, competitive forces prompted banks to change strategies. Instead of closing branches, the nation's big banks went on a building spree. In the year 2000, there were about 60,000 bank branches nationwide, according to the Federal Deposit Insurance Corp. Today, there are about 86,000 branches, a 43% increase.

What went wrong with the scenario? "Banks underestimated the importance of the branch," says Mark Greene, general manager of IBM's global banking unit.

The bank as it existed a decade ago has changed. It is no longer a place where customers primarily cash checks, pay bills and update balance books. Instead, customers are much more likely to walk in seeking advice on how to invest, establish lines of credit, buy home, life and auto insurance, or sign up for a credit card with loyalty rewards.

"What's interesting is all of these other channels—ATMs, online banking, telephone banking—did not replace the branch. They supplemented the branch," says Jim Adamczyk, a senior executive with consulting firm Accenture.

Increasingly, technology such as service-oriented architecture makes it possible for a single employee—a teller or banking associate—to access and sell the wider range of products. SOA, for example, can be used to reach into legacy applications that store customer bank accounts, mortgage information or credit card data and present the information in a single window.

TD Banknorth embarked on an SOA program out of necessity. In early 2003, Petrey received word from a key technology supplier, Fidelity Investment Services, that Fidelity would no longer provide support for a middleware platform it had assumed in its purchase of Alltel Information Services. The middleware had been used by Banknorth as a key integration technology between a new Internet banking application and its back-end systems. "It forced us to reevaluate our options; it was an integral piece of our infrastructure," Petrey says.

At the same time, Banknorth was pursuing an aggressive acquisition program, one that continues to this day. Since 1987, it has completed close to 26 bank acquisitions, and grown from about 285 branches in 2000 to approximately 600 branches today. In March 2005, TD Financial Group of Toronto acquired a 51% controlling interest in the company.

As a result, according to Petrey, Banknorth had big integration challenges while it was trying to introduce new services.

The company retained a Boston consulting firm, Riverton, to help investigate its architecture. Banknorth decided that SOA had evolved so it could produce benefits beyond traditional point-to-point middleware integration methods. Chief among the benefits: the power to reuse services, such as the ability to look up customer accounts, in new applications, as well as monitor business processes.

After a proof-of-concept pilot with several vendors, Banknorth selected webMethods' SOA platform in mid-2004. "We decided to take a think-big, start-small approach," Petrey says. In other words, Banknorth visualized where it wanted to go with SOA over the long term, but started with a small project. The case called for a Web service to expedite a customer address change. The service would allow a call center agent or branch employee to initiate an address change, and automatically have that change flow through to a customer's various accounts or products with the bank.

Next up: A small-business loan origination service and an internal application to handle project requests. Then, Banknorth bit into its first mission-critical application in the fall of 2005—its online banking system.

NEXT PAGE: A Project Without Hiccups

'No Hiccups'

The new SOA-based online banking system went live in May with no problems, Petrey says. "There were no hiccups—none," he adds. The success was due in part to the fact that the SOA team cut its teeth on straightforward initiatives, but it was also a factor of the project's high visibility.

Still, the integration team learned some lessons. One challenge: figuring out the right modularity, or how basic to build each of the services. You can create a service, for example, that allows you to look up customer bank accounts. Or you can create a service with more functionality that will allow you to look up all accounts, such as mortgages, insurance products and credit cards. If you want to reuse that service in other applications, you may only need one part of that service.

"What you're after is creating services with a high reuse factor," Petrey says. "It took some time to figure out the right level of granularity."

Petrey says it's tough to calculate a return on investment for the bank's SOA strategy because the project is in its early days, when a large chunk of costs are incurred. However, in a recent evaluation of an SOA-based system to produce online customer account statements for a large financial institution, Jim Crew, former head of database operations for Merrill Lynch, pegged initial development costs at about $467,000. Payback over a 10-year period was estimated at close to $11 million, with most savings coming from reduced mainframe usage, which is typically charged back to the business unit. Crew is now an executive at SOA Software in Santa Monica, Calif.

Petrey reckons that application development time has been cut in half due to the ability to reuse Web services. That has been a major benefit as the bank looks to absorb its latest acquisitions, including the purchase earlier this year of Hudson United Bancorp and its 204 branches.

"It puts us in a position to integrate the applications we pick up from our acquisitions faster," Petrey says, "and with less impact on the customer."

At A Glance: TD Banknorth

Headquarters: 2 Portland Square, Portland, ME 04112
Phone: (207) 761-8500
Business: Largest New England-based financial services company. Majority-owned subsidiary of TD Financial Group.
Chief Information Officer: John Petrey
Financials in 2005: $1.4 billion in revenue; $274 million in profit.

NEXT PAGE: A Check On 5 Banking Technologies

A Check On 5 Banking Technologies

Bank branches were supposed to go the way of the dodo bird, most industry analysts said five years ago. Technology, in the form of online banking, more advanced automated teller machines (ATMs) and sophisticated telephone voice response systems, would make a visit to the branch unnecessary. On the contrary, banks opened thousands of new storefront-type branches nationwide in the hopes of securing more profitable lines of business such as mortgages, credit cards and insurance products.

The resurgence of branch banking wasn't the only surprise. There are five technologies that analysts thought would bring major change to financial services firms, but which, for the most part, have yet to make much of an impact.

  • Smart cards. Instead of carrying around credit, loyalty-shopping and bank cards, technology in the form of a single card with an embedded computer chip would allow consumers to consolidate this data onto one smart card. Use of such cards remains a rarity. Why? While there have been

technology obstacles, the greatest reason is perhaps turf wars, says Mark Greene, general manager of IBM's global banking unit. Each bank, each credit card company and each store wants to own the customer experience.

  • RFID-based banking cards. With an RFID bank or credit card, consumers would approach an RFID-enabled vending machine or cash register and key in a passcode to pay for their items—no more card swiping. For now, however, RFID technology is still cutting its teeth in areas such as supply chains.

  • Mobile commerce. Forget having to visit a branch; consumers would do all of their banking by cell phone. Banks have initiated telephone-based voice response systems. Yet, many consumers still prefer to stop at an ATM or visit a branch.

  • Internet banking. Access to banking and financial services over the Internet is available.

But consumers concerned about online security have limited their use of these services. Jim Adamczyk, a senior executive with consulting firm Accenture, says banks have introduced more secure banking features over the last year.

ING Group, for example, lets customers choose pictures, such as a basket of oranges, to personalize their banking home pages, making it much harder for fraud artists to copy.

  • Improved customer relationship management and business intelligence. This technology is making the biggest difference to date. By integrating back-end banking systems, such as savings and checking accounts, with credit card accounts, mortgages and other products, banks are able to get a complete view of their customers. In so doing, they can serve clients without having to direct them to multiple employees, and can better predict what products customers might be likely to need next.—M.D.

NEXT PAGE: Financial Services Then and Now

Then and Now

Financial Services: SOA
5 Years Ago Today
Smart cards with embedded computer chips are touted as replacements for magnetic-strip bank and credit cards. Smart cards remain a rarity.
Need is established for more secure Internet banking and financial services applications. Banks getting more innovative with secure sign-on procedures, but concerns remain. Fragmented view of the customer prevails.
Second-generation customer relationship management systems and service-oriented architecture provide view of all customer accounts, and the smarts to know what to sell the client next. Banks consider shuttering branches and moving clients toward ATMs and Internet and telephone banking. Banks have instead gone on building sprees, realizing branches are valuable in selling high-profit products.