With the insurance industry in the financial doldrums over the past decade, there's been little left over to invest in upgrading legacy systems to newer data management. But as the lean years come to an end, some chief information officers are looking to upgrade their core processing systems, policy administration, claims and underwriting.
At least these are the conclusions of analyst Octavio Marenzi, CEO of Celent Communications LLC, the firm that released a survey earlier this year on the insurance industry's IT-spending trends.
"The insurance industry has been under much, much more cost pressure (than the banking industry), and as a result, any new IT development has been harder to justify," said Marenzi. There is some more life to be squeezed out of legacy systems, he said, but not that much more.
Chief information officers in charge of property and casualty, and life and health databases said the No. 1 issue is the technological upgrade of policy administration systems. The survey also found that chief information executives at larger insurance carriers were also more likely to need to upgrade their systems, compared with counterparts at midsize insurance carriers, said Marenzi.
Spending on software and services for policy administration systems on the property/casualty side is expected to reach $232 million in 2006, up from $128 million this year, according to the Celent survey.
While talk of upgrading legacy insurance databases has been on lips of CIOs for years, and technology analysts are fond of citing surveys to demonstrate the millions of dollars being spent on IT, the cautious approach of some insurance CIOs appears to have served them well.
For one, it's still possible to make legacy systems more efficient. There are plenty of mainframe systems that are working just fine, said IBM executives in interviews last year.
In many ways, the IBM executives said, mainframes are still quite attractive. The cost of memory continues to drop, for example, allowing clients to pack yet more data into their systems.
Insurance firms are also expert at the go-slow approach, and are notoriously slow at adopting the latest fads. The so-called "client-server" model, for example, has turned into a bust for large-scale institutions. "I cannot point to one single large institution that successfully moved from a mainframe system onto a client-server system in the late '90s," said Marenzi. "And it's not for want of trying."
One institution, a bank, sank $1 billion into moving from a legacy system to a single Unix-based client-server system, only to have the project ultimately scrapped.
Marenzi, speaking at a sales and marketing convention hosted by the German software vendor SAP, said that 45 percent, "a stunning number," of IT projects at big financial institutions are "total failures" or "have no payoffs whatsoever" for the firm.
In another instance, again from the banking industry, upgrading one system is of little use if the institution can't upgrade all its systems. To illustrate this, Marenzi said he had talked with one banking CIO who was well aware of the advantages of sending monthly deposit account statements to clients over the Web instead of mailing costly paper statements.
But that executive, according to Marenzi, said it would be risky and expensive to go into the legacy billing system and change it to remove clients who don't want paper bills.
"So even though the bank could squeeze more efficiencies out by simply not sending those statements out, they are not able to do it because of existing infrastructure," said Marenzi. "I could name a thousand different processes like that, from opening new accounts, to changing addresses, to issuing new policies, or what have you, that could potentially be made more efficient and operationally better but for the legacy systems where those processes are hardwired."
And then, there are time-frame issues that throw the fast-moving technology industry executives for a loop.
One insurance industry CIO in the midst of upgrading his firm's legacy systems said he was engaged in a 15-year project. "I suppose it's good to think long term, but sometimes one can overdo it," said Marenzi, particularly when the tenure of many CIOs lasts less than four years.
October 1, 2005
Copyright 2005© LRP Publications