TOOLS OF THE TRADE
According to Joseph Prunty, president of customer profitability specialist Core-PROFIT Solutions (West Chester, PA), a variety of tools and approaches-data marts, data warehouses, business intelligence (BI) tools and data-mining solutions-can be worked together to supply intelligence. ""What really matters is that the solution puts you in a direction that makes sense to you as the organization you are, and want to be,"" he says. ""It might be line-of-business based, product based, fee based or customer based. The one thing it can't be is everything to all people."" This is not to say that companies should not aspire to the goal of enterprise-wide business intelligence, Prunty explains, but rather, that to get the right answers, you first have to know how to ask the right questions.
This was a point insufficiently understood during the industry-wide data warehousing push, according to Judy Johnson, vice president, insurance information strategies, META Group (Stamford, CT). ""Data warehousing was a big disaster, for the most part, in insurance companies because everyone got on the bandwagon with very inexact requirements,"" contends Johnson. ""What we learned from the experience was, first, how you put the data together determines whether you can use it after you've put it together, and, second, that not everybody can mine and analyze data with the same tools.""
To the extent that the industry remains largely line-of-business and product-line segmented, true, enterprisewide customer intelligence will remain out of reach, Johnson says. In the meantime, business intelligence and data mining are likely to yield success through more focused applications, she adds.
Headway is being made in fraud detection in the healthcare and P&C arenas, says Johnson, partly because of its huge cost, and because it presents a more manageable focus. In those arenas, she says, ""they have some very specific needs for some very specific patterns of information, and they are actually capturing the information they need.""
At the special investigations unit (SIU) of Prudential Property & Casualty Insurance Co. (Holmdel, NJ, $1.6 billion in direct written premium), the people who need good intelligence are the fraud investigators. Tom Mulvey, the division's national SIU director, oversees the use of Alta Analytics' NetMap for Claims ""visual-link"" analysis tool to detect patterns associated with fraud. ""We dump all of our claim data into a repository, and it is then available to do a multitude of cross-reference regimens using NetMap,"" Mulvey explains.
Legitimate claims tend to be singular events, while fraud is most often a repeat offense. Therefore, the key to fraud detection is tracing relationships among claims, Mulvey relates. In the past this was difficult because investigators would search by name only. Since aliases are one of the chief weapons of fraud perpetrators, this presented a serious obstacle. ""This tool is good at looking through that camouflage,"" Mulvey says. ""It will immediately connect us to other cases through whatever data field the software finds a relationship: telephone, address, fax number, tax ID number, VIN number or whatever.""
NetMap's visual readout resembles the kind of charts-seen in countless crime films-used during investigations to trace relationships among otherwise disparate entities. By mining data to trace those relationships, the software actually substitutes for that laborious charting process, Mulvey says.
""A large-scale case involving a multitude of claims would typically involve countless man-hours trying to manually trace all the relationships to see how a scam works,"" Mulvey says. ""NetMap does in a couple of hours what might have taken a couple of weeks.""
Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio