The Case for an Incremental Approach

As a lifelong football fan, I have long rued the ascendancy of the short pass. Ever since Joe Montana implemented the West Coast passing scheme to such glorious effect in 1980s, teams all across the league have adopted its principles—opting for shorter, safer completions at the expense of a potential big gain down the field. While no doubt effective, this risk-adverse philosophy also can be dreadfully dull to watch.

While SOA implementations are not a spectator sport, a variant of the question bedeviling football coaches—whether to play it safe or go for the end zone—faces insurance company CIOs moving toward a service-oriented architecture. At Insurance Networking News’ first annual Web Services and SOA summit, held in New York on Jan. 23, 2008, a rough consensus emerged that when implementing SOA, a stepped approach is often preferable to grander ambitions.

Ron Schmelzer, managing partner at Baltimore, Md.-based SOA research and advisory group ZapThink LLC, said an insurance carrier would be well advised to implement SOA in a limited area first, and gauge its efficiency before rolling it out enterprisewide. “How do you eat an elephant?” he asked. “One bite at a time.”

One primary benefit of the incremental approach, Schmelzer noted, was a reduction in risk and project time. For a three- to six-month project, the cost of a troubled or failed implementation is low, while if a three- to five-year plan should go astray, the implications can be catastrophic. What’s more, by undertaking a long-term plan, IT is making the dangerous assumption that a carrier’s underlying business needs will not change markedly, he said.

The case for an incremental approach was clearly illustrated in the remarks of Gary Plotkin, VP and CIO for The Hartford’s P&C operations. As CIO of a 198-year-old company burdened with a thicket of legacy systems, some dating back 47 years, few people could benefit more from a rapid, enterprisewide SOA implementation than Plotkin. Yet, Plotkin said a rapid move to SOA isn’t always advisable or even feasible. To buttress this point, he unveiled a chart that showed (in finite, excruciating detail) the Byzantine interrelations between the 475 business applications, 212 functions with 200 separate software or infrastructure solutions that the Hartford had running when Plotkin came on board as CIO.

“This is a very complex environment,” Plotkin said, adding that he showed the chart to the Hartford’s board of directors to underscore the enormity of the challenge that the company faced. “Something that took 47 years to create does not get resolved in two. I wish it did, but we need to attack this problem piece by piece.”

Thus, Plotkin has wisely committed to methodically, yet inexorably, simplify the company’s IT infrastructure with SOA, quite content to eat the elephant one bite at a time while heading for the end zone.

(c) 2008 American Banker and SourceMedia, Inc. All Rights Reserved.

http://www.insurancenetworking.com http://www.sourcemedia.com

For reprint and licensing requests for this article, click here.
Analytics Data and information management Policy adminstration
MORE FROM DIGITAL INSURANCE