Diet Plan

The first project involved converting client data. Subsequent conversions consisted of PEMCO's personal auto line followed by the homeowners business. What was different about PEMCO's conversion was it did an in-force, or big-bang, conversion, which Shay believes is fairly uncommon for the insurance industry. "Most carriers go through county by county, state by state, or on a rolling renewal basis," he says. "For us, [the conversions] were pretty good size." PEMCO's auto line comprises about 187,000 policies, and the homeowners business has approximately 150,000 policies.

PEMCO had a great deal of clean-up work with its data, Shay relates, but as the project proceeded, it got easier because the carrier learned from earlier mistakes. "We learned from lessons as we went along," he says. "Each [conversion] got easier, and the results were better subsequently."

It took PEMCO 19 months to convert its personal auto line of business and another 12 months to convert the homeowners line. Other system-related work was required to get to where the carrier could begin to convert the data. With the first conversion, Shay explains, the insurer needed to get foundational pieces in place. "Each subsequent line of business really ought to take you less time, not only because you become more proficient at it, but you begin to put some key foundation pieces in place," he says. For example, a billing system might be part of the first piece of the project. There are some differences among subsequent lines of business, but the carrier won't be installing a new billing system with each line. "That ripples across the various components of the system," he points out.

When carriers begin large policy or claims administration replacement strategies, data conversion usually is the last step in the planning process, according to Forte. Companies eventually discover there is a great deal of dirty data that needs to be cleansed, and it is going to cost them more money than planned. Some data conversion tools Forte has seen cost $1 per policy. "If you have millions of policies in force and a few million policies sitting around for historical purposes, maybe that $10 million claims replacement project is going to turn into a $20 million project," says Forte. "That scares away a lot of insurers."

PEMCO tackled most of the data conversion internally, although the carrier did use some outside help. Shay maintains a key factor in the success of the conversion was the creation of a specially trained group of employees who were charged with handling both pre- and post-conversion issues. "Because these were employees of the organization, they were in a position to know a lot about our data and our business," he says.

PEMCO didn't consider contracting with an external firm to handle the bulk of the work primarily because of the cost, explains Shay, adding other companies might go through similar evaluations and reach different conclusions.

One of the fears PEMCO had with using an external data conversion team was the carrier believed an outside firm would not take ownership of the project the way internal staff would because the internal employees were going to live with the results of the conversion. PEMCO did augment the team with external contractors, but in Shay's view, that was not as positive as the internal experience because some of the contractors simply were punching the time clock. "We had some cleanup we discovered subsequently, and you can trace it back to a great extent to some of the augmentation resources we had," he says.

The data conversion took longer than PEMCO anticipated, acknowledges Shay, who warns those about to undertake such a project they need to be ready for the long haul. One thing PEMCO learned from the first conversion to the last was to go through dress rehearsals. For the personal auto conversion, the carrier had time for just three such rehearsals. "The purpose of the rehearsal is to determine whether there are some problems that need to be addressed," he says. When PEMCO approached the homeowners implementation, because of some of the cleanup work that had to be performed after the auto conversion, the carrier conducted eight dress rehearsals. "By the time we did our full [homeowners] conversion, it went almost flawlessly because we had found and fixed just about every problem that could be identified in each of those rehearsals," says Shay. "By the time we pulled the trigger, [the data] was very clean."

How clean? After converting 180,000 personal auto policies, Shay estimates about 500 policies had errors due to bad data. When PEMCO did the homeowners conversion of about 150,000 policies, there were just 20 policy errors. "It was a substantial improvement," says Shay.

Forte also contends buy-in from the business side is essential. "You can't have a bunch of IT people mapping data from one system to another because in general [the IT people] are not using, accessing, or inputting that data on a daily basis," he says. "A lot of projects fail because business people aren't brought into the process."

Carriers also face the problem of change management. Companies have business users who have worked on one system for many years, and when an insurer introduces a new system to them, the insurer can't expect those workers to start quoting and underwriting applications immediately. "[Conversion] is painful, and it's demoralizing from a workflow standpoint in the day-to-day operations," says Forte.

At PEMCO, it was critical the business side recognized the data conversion as a business problem that needed to be solved, asserts Shay, rather than an IT problem. This attitude helped the project succeed, he claims. "We had a lot of support from the business community," he says. "My guess is we would not have succeeded as well if this had been the IT department's problem." In fact, the conversion team was run out of PEMCO's underwriting department. "I think there was solid business ownership from the very beginning," says Shay.

In his experience, Forte indicates, carriers are guilty of a failure to document their actions. "Insurance companies have done data conversion projects before–whether on their own or with vendors or system integrators–but nobody understands why certain fields were put in there and why certain attributes were put into certain texts," he says. When the conversion team tries to reconcile those issues, team members scratch their heads, he says. Without documentation, he points out, data conversion is an arduous process. "There is so much customer historical data insurers keep on policyholders–whether in force or not," he says.

Potentially making matters worse, insurance companies never go through just one data conversion project, Forte continues. "It's going to be multiple times," he says. The first time insurers go through a conversion project, they need to make sure the staff documents the entire process. This way the company can understand why something was mapped in a certain way, and the staff can answer questions with clarity if the system has been documented. "It helps you for the next time you have a project because you can learn from your mistakes," he says. "And you will make mistakes. The next time you go through one of these projects, it should go a lot smoother."

Carriers also need to focus their efforts, advises Forte. "You can't take a policy or claims admin system and just turn on the switch," he says. "You've got to figure out what policies you want placed into the new system first. [Carriers] usually start in the renewal process to determine their threshold of pain."

O'Connell sees the concept of platform consolidation becoming a hot topic for insurers and a driver of data conversion. "Most of the companies in [the insurance] space today have multiple solutions," he says. "That's very costly, not only from an IT perspective but also from an operational perspective because different systems drive different procedures and policies."

There are many ways of doing the same transaction, but using multiple systems is an inefficient way to do business. As companies continue to feel the pressure to reduce costs, become more efficient, and improve service, they recognize the problem with multiple systems is at the core of this issue. "People finally are realizing they have to tackle this issue, and that's going to drive some kind of consolidation or conversion effort," he says. "Ultimately our belief is companies are recognizing they need to get consolidated to be competitive."

O'Connell works primarily with life insurance companies and states the problems life insurers deal with don't exist in the banking or investment fields. "We've seen as high as 20-plus systems within a [life] company," he says. "That can lead to 20 different ways to doing the same type of transaction or process."

O'Connell estimates roughly 30 percent to 40 percent of Accenture's clients are looking for outside help with their data conversions. Many companies have started the process on their own only to realize it is not as simple as they imagined. The main effort in conversion projects involves the data-cleansing process, he explains. After having worked on data conversion projects in the insurance, manufacturing, and utility industries, O'Connell has found the data in insurance typically is dirtier than with other industries. Companies have realized there is no magic tool to clean the data, either. "There are tools to help that process, but ultimately it's a lot of detailed work going through record by record and getting the operations people involved to understand the data," he says.

The reason insurance data is dirtier than that of other industries, O'Connell comments, is because insurers have to keep data around for such a long time. With life insurance policies, the data can be around for 40 or 50 years. Customers' life profiles change in that time–they get married or divorced, or their health changes. That's not the case with other businesses.

In addition, the frequency of contact between customers and carriers is not as high as it is with other industries. "You don't call your insurance company every day," he says. "If you move tomorrow, the first person you are going to call is your bank. You may not even call your insurer until you get the bill. There is less interaction."

Another factor is life insurers have developed some complex features in their products, observes O'Connell. "That's required almost a bastardization of the systems to support that complexity," he says. Business users may use a particular field on a system for something totally different than it was designed for. Changing a particular field is easier than changing an entire system. "You add those things up, and I think it has presented a more complicated issue for insurance than for other industries," he says.

From a purely operational standpoint, fewer systems to support is the biggest benefit for the IT department. Consolidation of environment has benefits from both a cost and a resource perspective. "You reduce the number of skills you need in-house, and you can focus your organization on building one set of skills from a technology perspective, whether that's Java, .NET, or whatever architecture platform a company has selected," says O'Connell.

Second, carriers can be faster on their feet. "The time it takes to implement a new product should be reduced to the extent you eliminate the older legacy systems that are not as well architected," notes O'Connell.

The reason carriers have been hesitant to undertake data projects is because they have seen other carriers try and fail, suggests O'Connell. The expense of the project also made it hard to justify. "Organizations got to the point where it was easier to add another system instead of convert," he says, but the industry has crossed the tipping point on that scale today. "For most companies, the incremental cost of implementing another system is so great, it's changed the value proposition," he contends.

The utilization of offshore resources has improved the cost proposition for this kind of work. Still, O'Connell believes the real work remains the data cleanup, and that job requires knowledge of the data, which requires company input.

In the area of mergers and acquisitions, United Trust Group has been a busy company. UTG has been formed mostly through M&A. COO Jim Rousey claims the company has been through 30 data conversion projects.

"What we like to do is take a team and do a study," he says. This generally takes about one week on site as the UTG team meets with staff from the interested company and examines procedures and systems, he remarks.

"We then begin to look at it in comparison with how our system works," says Rousey. "We start seeing what we call the poison pills." The on-site team determines whether UTG will need to do modifications or work-arounds. "Modify is a bad word in our company because generally it means big dollars to be spent to make our system work the way their systems do," he adds. UTG looks for work-arounds that are economical so the business data can be converted into one system.

That study will answer most of the carrier's questions, such as how much time it will take to get the project done and what the cost will be, reports Rousey. Should the two sides reach an agreement, out of that study emerges a project plan, Rousey explains, which will be the guiding document with time lines to make sure the business gets converted.

UTG labels its merger projects, continues Rousey. If it is a green project, the team feels it's going to hit all the targets; if it's a yellow project, it has some danger; and if it's a red project, UTG likely is going to hold up on proceeding.

Rousey describes his work environment as that of a model office. Co-workers work and talk with each other in almost a think-tank situation. "That helps when we hit a problem," he says. "We're limited with our staff, so we look at companies such as Universal Conversion Technologies to help us if we have more work than we possibly can do."

When Forte was involved in a merger as an underwriter, he recalls there literally were two desktops on his desk for the policies from the two companies. "It took a long time to merge those systems and data together into one unified system," he says. "I was a young guy at the time, and I was a bit appalled by that scenario, but that really was status quo for most companies."

Examining the data and systems of a potential merger or acquisition target is an important early step in the process, advises Rousey. The UTG team finds out as much information as it can on the system, and then when the two sides conduct due diligence, UTG spends time speaking to the target company's IT staff. "If you go into an acquisition and you don't figure out [potential problems] ahead of time, you're throwing darts at a dart board," he says. "You have to know you can bring in [the new system], put it on your administration system, and know what it is going to cost you–not just the data costs but the people costs. Fortunately, we've had a good team that's been able to do that over the years. When you look at the number of policies we administer and the amounts of money we are spending, we are very efficient."

If the systems won't mesh, Rousey says it can be a deal breaker. An expensive data or systems conversion basically drives up the price of the deal. "It better be a deal breaker for you," he says. "If not, you are going to spend a lot more money than you expected, and you are not going to be profitable."

Many companies that do acquisitions operate multiple systems. "It's probably already bitten them, and they just don't feel [the pain] yet," says Rousey. "It's going to bite them, because the systems get older and things are changing. If the system isn't one that is fairly supported and upgraded, you begin to lose some of the knowledge base of your system. You have risk you may not even be aware of."

A lot can be learned about an organization by just walking in and watching it, Rousey asserts. "You learn to listen with your eyes as much as your ears," he says. If UTG finds a company where everything is up-to-date–policies, procedures, and such–Rousey believes the company probably has kept its data in pretty good shape. "If you go in and things are scattered, and you ask for something and they have to go looking for it, you probably are going to have some problems," he says. "That's one of the indicators we see."

Dysfunctional data used to be deemed a necessary evil, according to Forte, and insurers put off conversion projects as long as possible, opting not to replace decades-old legacy claims and policy systems. Eventually, the constraints on IT budgets catch up with them, though. "If you have insurers spending 85 percent of their IT budget maintaining systems, you can bet the CFO is not going to continue writing checks for that," he concludes.