When insurers examine business intelligence systems, industry observers say, data warehousing capabilities sit at the forefront of the effort."When it came to data warehousing, most insurers saw it as massive projects driven by IT and not business," says Matthew Josefowicz, senior analyst at Celent Communications Inc., Boston.
"The objective in creating a central repository was not taken from a business perspective. This produced challenges of how one would use a data warehouse, how users would access it, and how they would be able to make correlations of the data," he says.
Presently, a business case is made first and the development of the data warehouse is carried out more conservatively, adds Josefowicz. "You don't stuff every last data source into a warehouse. It's more selectively carried out, with a long-term view in mind," he notes.
Similar to other automation initiatives where insurers failed at initial implementations, carriers have learned their lessons about how to best parlay their data into an asset rather than a liability. Better focus is part of it. Technology advancement hasn't hurt either. Deploying data warehouse solutions linked to the Web in conjunction with software, insurers can now more effectively convert reams of disparate customer policy, claims and billing data into insightful revelations.
But make no mistake, developing an internal data warehousing remains a significant undertaking, rife with obstacles.
When Chubb launched a business intelligence solution in 2001 known internally as Pinpoint, "the development of a data warehouse accounted for 65% to 70% of the overall effort," explains Jeff Hoffman, vice president of customer and market intelligence, at Warren, N.J.-based Chubb Group of Insurance Cos.
Developed by Cary, N.C.-based SAS, Chubb invested about a year to build the architecture and construct the platforms needed to optimize its data.
"One of the keys was implementation of an extract transformation and load software tool that could convert raw data into usable information," says Hoffman. "We realized a few years ago we needed to focus on producing clean, consolidated data, to have a system in place where you're not leaving it to chance."
