"Keeping different systems running from all the companies we've acquired, with different code to maintain and technology to interface with, would drive our cost of ownership through the roof," Treglia says.

It's not just IT costs that are a concern, Robison adds, it's the impact on how business functions. "If you try to wave the magic wand and say we're one company now, but we have two different sets of information systems, you have an unstable platform. You end up with broken processes," he explains. "Until you have data normalized and cleaned up, you have separate companies."

Like other terms, normalized means different things in different areas of IT. Technical database normalization and a discussion of the various normal forms are beyond the scope of this article. Instead, the insurers we spoke to are concerned with the practical reality of how to deal with data stored differently in multiple systems.

"When I use the word normalization, I mean taking data and putting it into a technically usable form that also makes sense for the business. That might imply business normalization more so than IT normalization," says John Lucker, principal at Deloitte Consulting, who leads the firm's advanced quantitative solutions data mining and predictive modeling practice.

There are different areas of data on which companies need to focus post-acquisition, including operational, financial, and analytical data, but one common thread runs throughout. "The meaning is paramount. You have data in particular fields, and you need to know the context in which that data exists," Robison indicates.

Identifying that structure in systems with which a company is not familiar, as often happens after an acquisition, is challenging but critical. "Understanding what data actually means and how complete and accurate it is, is one of the dimensions of data quality," Robison says. "Just like in a manufacturing process, if you have poor-quality raw materials, you end up with poor quality in the process."

Susan Clarke, senior research analyst with Butler Group, also stresses the compliance angle to data consolidation. "Storage is getting cheaper, but the more information you store and the more places you hold it in, the more difficult it is to achieve compliance," she says. "Oftentimes organizations have only a few days in which to produce requested information. And if you've got it all over the place in different systems, it can be difficult to bring the information together in that time."

"There's not a lot of grace period [to compliance], yet you've got the data, and you have to understand what it is," Treglia says. "[Unfortunately], in the early years before Sarbanes-Oxley and Section 404, [system] documentation wasn't something everybody commonly did."

"We also see 'green IT' being on the agenda," Clarke adds. "People want more efficient data centers. There's an awful lot of duplicate information in data centers, and through data deduplication, organizations can reduce their storage [requirements]."

Despite the advantages of consolidation and integration in IT, business issues take priority after an acquisition. "Very often what a company will do in the beginning is focus on merging business structures and cultures, trying to maintain the integrity and the performance of the business to assure agents and customers the change will work out for all. In the meantime, behind the scenes, IT is learning what it needs to do to absorb this new company. So post-merger, you'll often see the continuation of [system] silos," Lucker says.

That strategy cannot sustain itself for long. "At some point, to deliver on the reduction in expense commitments, some rationalization of systems and processes needs to occur," adds Lucker.

Insurers tend to follow a common hierarchy of activity in IT after a merger. "Companies clean up applications first, then their data, then processes," Robison says. "There is a reason people take the application-first approach, but a more careful approach would be to look at processes holistically first and let the decisions about data and applications flow out from that."

Genworth Financial, formed in 2004 as a spin-off from GE Financial Services, targets system and data consolidation from a business-priority perspective. The company is experienced in meeting post-merger challenges.

"Over the years, the number of companies we've integrated to form our company as it exists today is in the teens," explains Mike Shadler, CIO for the retirement and protection segment at Genworth Financial, based in Richmond, Va. Genworth offers life, long-term-care, payment protection, retirement income, and investment products as well as mortgage insurance.

Shadler asserts there are three business phases to a typical acquisition. "First is getting common operating processes, including financials and human resources, which have broader implications across the business. Then we move into sales and product teams, which is all about accelerating growth. Finally, we focus on improving operations from a customer touchpoint perspective," he says.

Data issues run through all these phases. "One issue is the need to consolidate business support systems, such as general ledgers and payroll. Then we need to rationalize our product systems and close off less cost-effective platforms," Shadler notes.

Ideally, the company prefers to convert acquired business to its core administration system, CSC's CyberLife, which contains data on more than two million policies. "We take the opportunity from the consolidation to reinvest the savings into other areas," Shadler says.

That includes creating GENIUS, a new business processing and workflow system the company built to integrate with imaging systems, financial systems, and its policy administration systems via an IBM WebSphere MQ-based messaging hub. "We are able to invest in projects such as GENIUS because of [savings from] consolidation efforts such as CyberLife," Shadler remarks.

There is no magic formula to the conversion process. "With acquisitions, we have very old systems we could be bringing together," he continues. "A lot of times to get through these consolidations we have to do some forensics: what the data is, what calculations exist, what that data links to. That's the biggest challenge of bringing together legacy technologies."

Although Genworth prefers to convert, there needs to be a solid business case first. "Our decisions are based on whether we can enhance the customer's experience or improve productivity. If we can't get at those two things, we don't do it," Shadler explains.

If not, Genworth will retain the acquired system and integrate it with the GENIUS system. "This creates common, standardized processes that are not completely system specific," Shadler says. "We have invested heavily in common workflows and common ways for us to work transactions outside the flow of individual systems."

Cleaning up post-M&A data messes tends to be a bigger problem in insurance than in other financial services sectors, Lucker claims. "Prior to insurance, I worked for a bank that was in the business of acquiring little banks. It got to the point where we could gobble up a bank in a weekend. On Monday morning, Bank X's customers would go to the new bank and their data would be in the new company systems," he says.

"The reason is a checking account is a checking account, a savings account is a savings account, and so on. But insurance is more complex," Lucker contends. "A workers' compensation policy [at one company] might have only the bare minimum of information, just enough to write the policy, vs. one at another company that is more culturally focused on gathering additional data to use in downstream processes."

That results in a typical divergence in data structure and quality between the acquired and acquiring company. Because of the importance of data and the difficulty of dealing with data retained in multiple systems, Lucker advises companies to analyze carefully data quality when performing pre-acquisition due diligence. "Companies should focus less on the functionality of systems being acquired and more on the data being used by those systems," he says.

And in the post-merger push to consolidate systems and gain efficiencies, insurers need to be careful not to discard or damage valuable data assets. "A lot of times when companies look at eliminating a piece of software, they look at how much it costs and how many users it supports," says Robison. "What often gets overlooked is what data it provides."

Consolidation of different platforms invariably leads to a change in the way one company had performed business and in the manner by which users are presented data and information.

"That really becomes apparent in a merger, because you have the same data meaning different things to different people–different definitions of the important entities that are tracked in these different systems. Also, inside each enterprise, there are multiple domains. A customer will be [defined] differently in the marketing department than it will be in the accounting department, so the entity gets used in different ways," Robison notes.

The fundamental place where insurers seem to stumble, Robison reports, is forcing the issue. "A typical mistake I've seen companies make is creating one [enterprise data] definition that ignores the reality there are different domains within the enterprise. Information should be defined differently within those domains and transformed across them," he says.

Therefore, insurers need to manage projects with people who not only can bring both a business and IT perspective on system and data integration but who are skilled enough to resolve organizational quarrels over data definitions.

"What is important is who is the 'data czar'–the person or people paying attention to the architecture and business decisions around what it means to manage data effectively across an enterprise," Lucker says. "That includes things such as creating a holistic information inventory and understanding the type of data you have externally and internally. Looking for people within your organization who have a passion for managing data, leveraging the data's intrinsic value, and creating a culture that recognizes that data is a valuable off-balance-sheet asset are incredibly important."

Having the right resources in place has been essential to post-M&A success at Wilton Re. "If you have just IT folks doing the work, they can go down a path that isn't appropriate from a business perspective because they haven't been exposed to the business," Treglia says. "We have a lot of people with good IT and business knowledge who can raise issues with the appropriate people to get resolutions quickly."

In addition to identifying and promoting people with those skills internally, Wilton Re demands those capabilities from its outsourcers, such as CSC. "Its conversion staff isn't just IT people. CSC understands the business, has seasoned professionals who point out problems that can occur, and has a methodology and understanding of the business that really allows us to streamline the [conversion] process," says Treglia.

ACORD data standards also are important to both internal data integration and external data interchange at Wilton Re, particularly given the number of different insurers with which the reinsurer transacts business. "Being in the traditional life reinsurance business, one of our major challenges is we have a significant amount of disparate info coming in from different sources. We've adopted the ability to use the ACORD XML standards to translate all the data into a common format, so it doesn't matter where the data comes from," reports Treglia.

For the task of conversion from external sources, Wilton Re uses a rules-based transformation solution from Rivers Wave Consulting. "The use of ACORD XML itself has helped in conversions because we're not reliant on positions of data, and it gives us more flexibility for moving data around, but there's no magic to it," Treglia states. "Mapping is mapping."

"We've been able to streamline the [post-acquisition] process by using a common methodology; the same type of understanding of the target, and mapping to that target," Treglia says. "I boil it down to simplicity. Post-merger or post-acquisition, if you have an established methodology, it makes it easier. Instead, some companies look at the separate platforms and then wonder what they're going to do with them all."

Particularly for a company whose business model depends heavily on acquisition, having a repeatable process for system integration and data consolidation is essential. Acquisitions are a key strategy for Meadowbrook Insurance Group, whose most recent purchase was U.S. Specialty Underwriters in April 2007. The company operates through five different insurance companies and also has four agencies, a wholesale brokerage, and a half dozen insurance subsidiaries.

Meadowbrook's first choice is to convert new companies to its INSTEC QuickSolver rating and administration system. The company also looks to convert new MGAs to its ConceptOne management system from Epic-Premier Insurance Solutions. "We determine whether [MGAs] are generating enough data to justify the investment from an IS standpoint," explains Jim Lee, IS manager at Meadowbrook.

Meadowbrook has an established data integration process that relies heavily on a framework created by integration and business process management vendor Adeptia. "We have a basic [integration] shell that is reusable. What changes are business rules around things such as how we handle data, who gets notified in errors, and what data or files are held. If we can reuse that code or framework, it makes our job much easier," Lee says.

Meadowbrook's biggest challenge in dealing with disparate systems, Lee adds, is discovering the meaning of data in newly acquired systems. "The pain part is do we understand our data, what [data] needs to pass, and what are the business and mapping rules around data," he says. "Some of those issues are standard challenges, but we do have some unique rules around data in our systems, and in acquired systems, you don't always have the necessary source data to build a corresponding new record."

That's where Adeptia comes in. "You need to determine programmatically what to put into missing fields. Adeptia is good at building that–it queries to find the business rules to fill the gaps between source and target data, so that data becomes useful information," Lee says.

While leaving systems as is post-merger isn't usually a viable option except for run-off business, there are technical risks to the conversion and normalization process–from broken integration points to performance issues. "There is a usability problem when data is overly technically normalized," Lucker indicates. "Historically, data architects have been focused on technical normalization, which doesn't necessarily make sense for business."

Whether it's dealing with operational, financial, or analytical data post-merger, there also is the danger of paralysis by analysis. Lucker advises companies simply to keep moving forward. "Step one is creating an inventory of your information. Step two is getting your data ready for use, spending more time focusing on meaningful utilization of data and less attention on getting it perfect. You always can do continuous improvement. Step three is finally using the data as a valuable asset and leveraging it as a differentiator," he says.

He illustrates his point with an example from analytics. "If you can start using what you've got and generate some value from analytics, the return can pay for continuous improvement," Lucker points out. "Striving for perfection is a treadmill that never ends. People get tired and frustrated, and it doesn't get done at all vs. it getting done 'good enough.'"