Many a carrier has embarked on ambitious projects only to have overlooked a critical issue: Will their policy administration system be able to handle the big leap forward? While hindsight is 20-20, and no insurer can pinpoint exactly how much it will expand in two to three years, much less in 10 or 15, carriers do have many strategies at their disposal to avoid the resultant problems. Those lacking the necessary foresight face grave headaches indeed. Some are obligated to jettison their long-time policy admin system because of its failure to scale. Replacement projects can take 18 months or longer for insurers, which are forced to embrace band-aid solutions in the interim, according to Jeff Goldberg, senior analyst in the insurance practice of Boston-based Celent.
Carriers that find their modern systems floundering "suddenly are dealing with weeks of turnaround, when the whole reason to spend millions of dollars and 18 months on an implementation was to avoid that kind of issue," he says. Some systems that simply cannot handle the policy load either slow down or crash altogether, triggering the carrier to filter business manually and push clamoring customers onto its Web site.
"Bringing a system into production that is not ready from a scalability standpoint is nothing short of disastrous," explains Phil Ehlen, CTO of the P&C division for Falls Church, Va.-based Computer Sciences Corp. (CSC). "I typically see the failure occurring before one gets into production, and what that results in is lost money and lost time."
Those costs "massively multiply" when the problem is discovered after production begins. Top officials who make such ill-fated decisions also are destined to pay lofty consequences, he adds. At the same time, "an inability to quickly react to market changes may translate into loss of market opportunity and brand tarnishing," says Marc Dutton, managing director of FJA-US, Inc., a New York-based insurance technology provider of life and health policy admin systems.
Further, the higher operating costs resulting from scalability failures may erode the firm's competitiveness and market share, he adds.
Scalability woes also lead to bitter acrimony between the insurer and its vendor, which can spill into the courts. However, insurers are increasingly showing greater prescience. Over the past couple of years they have become more aware of the need to build scalability assurances into their contract with the vendor, notes Patrick Hatfield, a partner in the law firm of Locke Lord Bissell and Liddell. Insurers also have awakened to the reality that while they could take scalability somewhat for granted in the past-when older Cobol-based mainframe policy admin systems built for long-term use reigned-today, with modern systems, they can not.
With carriers feverishly embracing modern policy admin systems in recent years, issues of scalability inevitably jump to the fore. Though a modern distributed services-oriented architecture system offers rich benefits, installing a new policy admin system always is fraught with risk, explains Goldberg.
PREVENTING CALAMITY
Consequently, longevity is one of the characteristics that insurers such as Indiana Farm Bureau Insurance focus on when choosing a policy admin system vendor. "We are making 10- to 15-year decisions here at a minimum," says Greg Clancy, SVP and CIO of the Indianapolis-based provider of farm, auto, life, home and business insurance. The firm chose Exceed, a policy admin system from CSC, for its sizable automobile and home portfolios partly due to its proven scalability and because "CSC has such a big footprint in the insurance industry that they will be able to migrate the functionality to different technology footprints as time goes on," says Clancy. "They have the scale as a company to survive for the long term."
Such proven scalability-the vendor's largest client has 2,750 simultaneous users in production, its largest number of transactions per hour on a single production instance is 45,000, and 97% of its clients' measured screens have less than a two-second response time - has motivated the insurer to quickly get all of its P&C business billed out of Exceed Billing.
Experts consulted for this article generally urge insurers to guarantee that they have policy admin systems capable of scaling for at least 10 years.
"If you are going to spend millions of dollars and, almost more importantly, the effort and the derailment to your business for two years to get something up and running, it needs to work for more than four or five years," says Celent's Goldberg, adding that bringing a lawsuit against a suddenly defunct vendor is hardly a pleasant task.
For Allianz Life Insurance Company of North America, such a system needs to scale for over 15 years, according to Todd Carlson, capability manager, for the Minneapolis, Minn.-headquartered provider of fixed and variable annuities and universal life insurance. "All I've got to do is add this type of software and move it to this type of iSeries box and it should be fine," says Carlson.
Allianz selected the ID3 policy admin system, which has demonstrated the ability to complete 16,400 transactions per hour and to support 2 million policies, from Fiserv Insurance Solutions, Cedar Rapids, Iowa, for its fixed annuities business.
Comprehensively testing the scalability of a policy admin system is particularly vital when it comes from a young vendor with few insurance clients, though this should be considered de rigueur in general. Three parties should be included in the testing: the insurer, hardware provider, and software provider, according to Lukas Hoffmann, a developer of Policy Management, a scalable system from the Walldorf, Germany-based supplier SAP. The customer defines the products and its calculations, the hardware partner optimizes the systems and the software vendor identifies avenues for improvements in standard components, as well as in custom enhancements.
Ideally, the vendor will be able to illustrate to the insurer that its system had scaled for a carrier with a similar profile, and the insurer will fully engage itself in systems testing.
Best practice insurers will have their own scalability performance lab, explains Goldberg. They will have built a place in their data center where they can test their own applications before going into production. In assessing a new policy admin system-or any new vendor acquisition-they will install it in their scalability environment and will write their own text scripts, ensuring that the system will work.
For instance, Allianz conducted performance tests at an IBM lab to see how many policies Fiserv could handle.
"We try to have a full staging or regression environment that can kind of mimic our production environment, and try to have as much automated test cases [as possible] because, at some point, it becomes an issue of cost to have that type of server sitting around just for testing," explains Carlson. Allianz has found that Fiserv is able to process most of its transactions in less than one second.
Indiana Farm Bureau Insurance credits much of its successful implementation to its no-shortcuts approach that called for ample testing designed to release bottlenecks in its policy admin system. This included seven months of QA testing, two months of performance testing and sending its officials to various CSC user groups, says Clancy.
PREDICTING COMPANY GROWTH
When an insurer's policy admin system is unable to stomach its growth, it is often partly to blame. The insurer must equip its supplier with the information it needs to scale effectively. Instead, carriers frequently assume they can just provide vendors with statistics such as the number of their users.
Instead, vendors need to know the carrier's peak transaction load, explains Andy Yohn, CTO of Duck Creek Technologies, a Bolivar, Mo.-based technology firm that provides policy admin systems to carriers including SUA Insurance Co. and Harleysville Insurance. He says a vendor also needs to know what technologies the carrier will integrate into its policy admin system; such a system can be slowed by a carrier's stodgy, outdated infrastructure.
While vendors often struggle to extract the information they need from insurers, the latter has dramatically improved upon past practice, when it could be weeks or even months before it gave the vendor the data it needed to feed its scalability model, says CSC's Ehlen. Still, vendors continue to frequently face the hurdle of getting the right people engaged in scalability discussions. "It isn't always that the carrier doesn't have the information," explains Ehlen. "It is making sure that the people who have that information are engaged and accessible or even known to the team that has helped to implement the software." Creating a performance testing environment that adequately mimics an insurer's production environment also can be a tall task for vendors.
When policy admin system implementations go awry, scalability concerns may only represent one element. But if scalability troubles aren't nipped in the bud very early, they can wreak havoc on the whole enterprise.
"It is all about processing the business, getting the right answer, helping the carrier to be successful and getting their policies sold, processed and serviced," Yohn says. "That is still the No. 1 goal. The scalability improvement is just: We want to make sure that that is not a hindrance."
Daniel Joelson is a freelance business writer based in Alexandria, Va.
(c) 2008 Insurance Networking News and SourceMedia, Inc. All Rights Reserved.
