Ripe For The Picking?

With the drive among carriers to increase productivity and reduce costs, policy administration systems vendors are scrambling to offer products that deliver better return on investment, but insurance IT executives still see room for improvement in those products.

A sampling of such executives, when asked to comment on their likes, dislikes, and wants regarding policy administration systems, raised several issues. While they agree on many points, there are other unique concerns dictated by the size and goals of the individual companies.

Ernie Pearson, director of systems development for SECURA Insurance, praises vendors for creating products that are compatible with industry guidelines or open-based solutions around ACORD and XML standards. "It's important because it gives us more of a common platform for exchanging information from a vendor system to a system we might develop to an agency management system to third-party systems," he states.

Another positive with today's policy administration systems is the ability for a carrier to configure the software, he notes. Such flexibility enables carriers to change rates or other data or specify how documents will print. "You don't have to go back to the vendor," Pearson explains. "Carriers can take ownership of the changes and configuration. More vendors are heading in that direction."

On the other hand, Pearson maintains, many policy administration systems are not well developed. "They're immature," he contends. "For example, during the initial stages of a request-for-information process, vendors claim they can handle 'all commercial lines,' but they show you personal auto. It's easy to crank something up quickly and have something to show, but beneath the surface, policy admin systems are not as fully developed as they need to be.

"I think [vendors] sometimes misrepresent or 'oversell' their products," he continues. "We participated in some demos where they failed. They can engage you with whiz-bang bells and whistles and technology, and they convince you they can build it. Depending on your appetite for risk, some of them probably can."

Another challenge to carriers trying to pick a policy administration system vendor is the possibility the vendor will merge or be acquired, says Pearson. "You don't know whether the horse you bet on will be operating as an independent entity in 18 months." If not, he asks, "what happens to the application and the support? Now, you may have to switch to a different product that could impose cost and time-frame issues on you."

All this, says Pearson, makes the purchase of policy administration software a little tricky. "Policy admin is a core feature, but with a lot of changes, it's kind of risky," he warns.

In terms of which product improvements he would want, Pearson indicates more "customer configurability" would head the list. Overall, however, he believes vendors need to offer "more comprehensive solutions and better integration with a business rules engine (or at least a solution designed with the ability to integrate)."

He summarizes his company's experience: "We're pretty mainstream America; we don't write any unique lines, and we're in fairly innocuous states. Yet it seems like in our discussions with vendors, we find things we expect to be in systems are not there."

One example of such an omission is "proper handling of out-of-sequence transactions," Pearson adds. "That's something everyone would need. It's not unusual; it's a fairly typical occurrence in every company. Some vendors say they have it, but when you look more closely, they don't have it or it doesn't work properly."

"We [operate on] a large scale, so the number-one factor for me is scalability," says Andrew MacDonald, CIO and vice president at The Hartford. A policy administration system "must scale to millions of policies and millions of customers." He also emphasizes the need for what he calls full-function policy administration. "If the software is created well, it [covers the] full life cycle of the policy process," he explains. "Many vendors out there can support the full life cycle."

Like Pearson, MacDonald applauds the flexibility offered by many policy administration products in terms of configuration. Such flexibility, he suggests, allows carriers to maintain a competitive advantage–depending on how they use it.

One potential stumbling block for vendors, MacDonald mentions, is the complexity of the U.S. insurance market, where vendors will need "proven experience in multistate, multiline writing." He remarks some new players to the U.S. market haven't been able to deliver, even though they have had success overseas. "There are some good products that really aren't proven in the U.S.," he says.

According to MacDonald, the market as a whole is "rationalized; there aren't as many players as there used to be, so there aren't as many options in terms of providers." The number of vendors to choose from is shrinking as there are more acquisitions in this space, he adds. "We end up getting less competitive choice, which is troubling for the future."

When asked what needs to be done to make policy administration systems better, MacDonald repeats his call for an "engine to be able to execute the full life cycle of policy." The challenge there, he points out, is how carriers implement such a system in the current traditional business environment.

"Because there are not that many vendors, there is a lack of sophistication in the industry," he asserts. "There is not a focus on rapid, intelligent implementation practice. There needs to be a balance between quick implementation and understanding the effect of what I'm doing. Vendors don't entirely understand the full impact to a company, especially a large company such as The Hartford. The challenge is how to be smarter about implementation of policy administration systems."

MacDonald also calls for a more sophisticated data conversion process, similar to what happens in the banking industry. "When a bank acquires another bank, over weeks or months it can acquire and rapidly convert data to a single, fully integrated system," he notes. Insurance data, however, is not as structured as banking data, making data conversion the most difficult part of acquiring another company.

"It holds the industry back in a lot of ways," says MacDonald of the lack of standards. "Adherence to standards such as ACORD is key, but ACORD goes only so far. It's very good for what it does, but not a lot is modeled in the ACORD standard."

Ideally, MacDonald adds, "we could have policy administration systems coming out of the box such that they are fully integrated into the agent market space with agency management systems." If agency management systems and insurance policy admin systems could work together out of the box, "you wouldn't have to worry about integration," he points out. "That would be a huge advantage for everyone, and it would drive a lot of cost out of the industry."

Another feature MacDonald would like to see is the ability to manage different brands within one's policy administration system. "Branding of policy products will be as big in the U.S. as it is around the world," he predicts. Consequently, policy administration systems will need to have the ability to support true branding in partnerships. "We don't see that built into [policy administration systems] today. It's happening heavily in Europe right now–like buying insurance in a supermarket," he comments.

Michael Fergang, vice president and CIO at Grange Insurance, in Columbus, Ohio, likes the thin client (less maintenance and patching) approach of today's policy administration systems. Such a structure, he contends, provides "better integration with billing and claims. I think all companies struggle with that. Being able to view information as policy-centric vs. customer-centric is a real strength."

When asked about how systems could be improved, Fergang says he would like to see more sophistication around rating engines. "The tools must be more complex in the current environment," he states. "Many players sell just rules engines and rating engines; the key is to be able to incorporate them."

Fergang also maintains even though today's systems employ good tools such as XML and service-oriented architecture, "we still see 'proprietariness' in some of these applications. Vendors will embed rules that don't work for everybody. It's a challenge; they try to serve the masses, but every company has its own needs.

"I'm not sure vendors have a strategy beyond the then-current release," he continues, adding, "There seems to be a more positive attitude about policy administration systems overall."

Challenges for implementing a new policy administration system arise when two or more administrative systems already exist in the same enterprise, he says. "Think about the number of interfaces you have to build to do your core business. Now, you want to replace them with one system, but what about all those interfaces?"

He emphasizes that for "large companies with millions of policyholders, it's not a trivial matter." Carriers must manage the conversion while keeping the business going, "and the business is not slowing. I can't envision having five different admin systems."

On the whole, Fergang believes policy administration systems need to do "a better job of reporting and [allow for] more sophisticated integration with other systems," along with warehousing and integration with business intelligence.

Dan Mets, director of IT applications for Nationwide, observes when it comes to policy administration systems, "a lot of the vendors are building in a lot of flexibility and functionality around agency relationship and customer relationship. There are shared components between what an agent would use and what a customer would use on the Internet."

On the other hand, he comments, "I haven't seen a package that has been proven on a large scale–for companies the size of Nationwide. Our volumes are so high, it takes us out of the off-the-shelf market." The integration of a policy administration package in a high-volume environment may be a much larger task than the acquisition itself, with the purchase being only about 10 percent of cost, he adds.

"There are two sides to integration:data conversion and system interfaces to other core applications," he explains. "We own multiple companies that all came with their own systems. Keeping them all in synch is very complex." Developing data interfaces for such companies (rather than integrating the systems) "is the quickest but not the cleanest way, and it tends to be fragile, highly complex, time-consuming, and costly," he indicates.

Integrating such systems, on the other hand, is much more time-consuming, asserts Mets. "Interfacing is the incremental feed of data back and forth on an ongoing basis; integration is merging data from two companies into a common application."

One significant challenge for policy administration system vendors, he says, is the difficulty of keeping up with technology advances. "Technology changes rapidly, and insurance companies develop standards as to what they want," he states. "Technology is changing faster than vendors' ability to recoup investment from it. The rate of change will outpace a lot of vendors. Smaller [vendors] are at a disadvantage, and if their technology is exceptionally good, they get swallowed up by the larger vendors."

The biggest problem, continues Mets, is "a lot of vendors understand product, but they have little appreciation for complexity of implementation–which is the majority of the cost by far." He points out larger vendors are likely to have people on staff who understand complex implementations. However, "some newer vendors that are less installed wouldn't have an appreciation of complexity of implementation with a larger carrier.

"Business demands change faster than vendors' ability to integrate changes," Mets says. "That's another issue the vendors will have to find a way to address. Long-term technology planning is difficult, because business needs change so fast in a large carrier. Your chances are better when you do it a piece at a time.

"An approach or a tool set to make integration easier would be the most valuable tool on the market," Mets concludes. "It would accelerate integration work and the conversion process and address data quality across different applications."