Many data warehousing projects never got off the ground, but insurers and vendors learned some valuable lessons from those failures. The most important one, asserts Josefowicz, is to have a more strategic approach to the situation. Plan the project better, and understand what you are putting into the project, what youre hoping to gain from it, and what kind of implications [the warehouse] is going to have for the operating environment, he says.

Data warehouses and data mining spawned an era of business intelligence (BI), and Josefowicz believes the different types of data usage created confusion for technology professionals and business managers alike. That is one of the reasons Celent coined the phrase data mastery to get away from the various categories. Data warehouse implies setting up a big database, dumping everything into it, and then running reports off it, he says. What we think is important is not the specific architecture that is used but the technology strategy of being able to integrate and analyze the data. Whether that is done by porting in the data warehouse and running reports or whether its done by integrating various data sources so they can feed into either analytics applications or operational knowledge-worker applications is another element. I dont think BI necessarily implies data warehouse. Data warehouse is one architectural structure that is a potential data mastery strategy.
Data quality is another important issue for insurers, particularly on the property/casualty side, where more complex risks are dealt with. Its not just policy data, Josefowicz points out. Its also operational data and financial data.

Data mastery initiatives along with policy administration and document management make up the three areas where insurers are seeing their highest return on their investments, Josefowicz maintains. Weve seen increased interest across the board, but it continues to be careful spending, he says. The crazy years of the late 1990s are gone, and theyre not coming back. I think thats a good thing because that kind of craziness can lead to multimillion-dollar projects that dont deliver any business value. Those kinds of proj-ects arent good for the vendors, either. You might sell a couple of them, but you get a really bad reputation and youre not going to sell any more [of those systems].
The proper way to system work is through a more focused approach toward technology investment, advises Josefowicz. What is going on now is yielding more business value, and thats better for the insurers and also for the vendors, he says. Its less initial easy money, but longer term its better strategy for the vendors.

ROBERT REGIS HYLE