advertisement
advertisement

Failure Is Not an Option

How can financial-services companies and banks improve the odds that big technology bets will pay off? At the Wall Street Solutions Center, they try before they buy.

Up on the 14th floor of New York’s historic U.S. Realty Building, hard by Wall Street in lower Manhattan, programmers and testers work at bright-orange tables in rooms with sleek gray walls. But this is no dotcom incubator or some new-age brainstorming center. It is a place where serious executives from serious companies come to avoid making serious technology blunders.

advertisement

Making a major IT purchase has always been fraught with risk, especially in the high-anxiety world of finance. Spend millions of dollars and disrupt operations for months, only to find that the new system was never actually tested in a company like yours. Merge with another bank, and then wrestle with the fact that it uses completely incompatible systems. Trust a trendy consulting firm, and then watch it go out of business mid-project.

Unveiled in January by Plural Inc., the Wall Street Solutions Center (WSSC) was designed to help companies make big technology bets — without betting the future of the company in the process. With two separate 100-MB local area networks, a storage-area network that can hold up to one terabyte of data, and 4 data centers with 21 Compaq servers in the cold room next door, the state-of-the-art laboratory is capable of simulating a system with some 6,000 transactions going on at once. “We lose weeks, if not months, by jerry-rigging tests internally,” says James Oliverio, CTO for investment banking at Credit Suisse First Boston, which is using the WSSC for several projects. “There are always the ‘gotchas.’ This center lets us kick the tires while not taking up our space.”

Plural is an e-business consulting firm (annual revenue: nearly $90 million) with a real history (it’s been around for 11 years), big-name clients (including 33 of the country’s leading investment banks), and high-powered strategic partners in Microsoft and Compaq. Plural makes no bones about its preference for Microsoft software and Compaq hardware (and Microsoft recently took a minority stake in the firm), but CTO David Osborne insists that the center can handle virtually any kind of simulation necessary.

The lab took shape back in 1998, when Plural was hired by the NASDAQ to build its Surveillance Delivery Real-Time System. The mission-critical Windows NT-based system monitors billions of trades every day and alerts NASDAQ officials to potential securities violations when any trades or quotes deviate from the norm. To convince NASDAQ that the system could handle a 4-billion-trade day, Plural designed a prototype that processed 800 messages per second for eight days.

The success of the project convinced Plural that it had a winning model. Indeed, the WSSC is now taking reservations from a range of companies that want to send their tech crews essentially to camp out in the center with their sleeping bags, Sony Playstations, and Cheetos and put the programs through their paces (aided by full-time Plural consultants). “We may have to spray down the lab,” jokes Osborne. If clients are satisfied after the initial trial, which costs some $20,000 per week for a one- to four-week project (the original outlay is applied to the purchase), Plural is betting that they’ll commit to buying an entire system or integration project. “Everyone is looking at costs and scrutinizing everything now — much more so than when things were great,” says Plural president and CEO Neil Isford, a 20-year IBM veteran who was, most recently, vice president of e-business at IBM Global Services. “When you see all of these things working together and how much money you can save, it gets your attention.”

It got the attention of the Federal Home Loan Bank of Chicago, which is considering using the WSSC to help with its conversion from Windows NT to Windows 2000. The bank is also building its own internal lab, says Martin Grill, the bank’s VP of communications and systems manager, “but the value comes from Plural’s personnel having developed real expertise in this area. Trying to develop expertise internally is a crapshoot. This gives us a way to try before we buy.”

advertisement

It may also give Microsoft a way to meet a tough business challenge — persuading Wall Street’s technology leaders to choose its offerings over those of dominant players such as Sun Microsystems. On Wall Street, Microsoft’s market share hasn’t even cracked the double digits, concedes Bill Hartnett, Microsoft’s global director of financial services: “There is still that perception that Microsoft isn’t really an enterprise player. The intent of the center is to provide all three companies with a place to demonstrate their capabilities.” Plural also intends to open a lab in Chicago to focus on Microsoft’s .Net applications. “Failure is not an option,” Hartnett declares. “We want to throw Sun Microsystems out of Wall Street.”

Learn more about Plural on the Web (www.plural.com).

Sidebar: Five Ways to Avoid Disaster

Plural president and CEO Neil Isford has seen his share of technology-enabled successes — and more than a few meltdowns. Here’s his advice for avoiding disaster.

1. Be honest about the risks. Before you commit, get all the facts on the table. Go to a consultancy with tire tracks, one that has done this before. Make sure that they outline all costs — including any potential hidden costs — up front.

2. Test, test, test. Find an environment in which to test the technology, internally or externally. Then do it until you can’t stand it anymore. It’s too big a decision not to. Then go through a pilot phase that really looks and feels like your own environment. The “people” aspect of implementing a new system is often more challenging than the technology itself.

3. Love your project team. Once you’ve made the “go” decision, make sure that the people around you not only can deliver the new system but also will tell you when there’s a problem. If there are just a bunch of junior people in place, you’re looking for trouble.

advertisement

4. Give yourself wiggle room. There are always speed bumps. You may be committed to launching an application by a certain date, with X weeks of test time built in. But if there’s a problem with the software, the schedule will start to slip, and test time will start to shrink. If you’re not disciplined, you could go live before you’re ready — which could be a disaster.

5. Keep talking. Undertaking a massive IT project requires careful communication, every step of the way. So define your milestones — not only so that you can track your progress but also so that you’ll be ready for the inevitable glitches.

Contact Neil Isford by email (isfordn@plural.com).