No Big-Bang Projects Here
At Columbus, OH-based Nationwide Insurance ($24.5 billion in assets), its data warehouse has high-level business sponsorship. ""There is a strong partnership and this is not a technology initiative,"" according to KellyCannon, vice president, information technology, Nationwide. ""One thing that can derail a data warehouse is if it becomes a pet project of IT. The business has invested in the data warehouse because they see good value in it.""
Nationwide's data warehouse initiative centers around the creation of a customer information file (CIF), a central file that contains all of the information about a customer. However, creating the CIF has not been a big-bang project, according to Toby Cook, director of business intelligence solutions at Nationwide. ""We are prioritizing the work,"" he says. ""We have a lot of legacy and we look at the most important systems first."" And although the CIF will have a customer focus, it is being built from a product focus. ""The toughest challenge is the maturation of the data warehouse, and unless you take the huge and overwhelming approach, you have to build from a product perspective,"" Cook explains. ""First you win with auto, then you bring in homeowners. We are essentially backing into the customer view one piece at a time.""
Nationwide uses Validity's (Westboro, MA) Integrity product to check and verify names and addresses that are stored in the insurer's Teradata data warehouse. Currently, the data warehouse occupies five terabytes of disk space, although Nationwide has nine terabytes available. Nationwide started the project in 1999 and is gradually adding more information to the CIF. ""We had a lot of different views of the customer,"" says Cannon. ""If I am a Nationwide client, I just want one view. That is the biggest step for us.""
Nationwide's seemingly slow progress is not uncommon, according to Teradata's Helms. ""There is no big bang,"" he says. ""It is not an easy process to map out data elements from a company. There may be 50 elements and many of them overlap. A company has to look at all of the opportunities and that may take two to three years,"" to complete a data warehouse project.
Although American National hasn't needed three years to complete its data warehousing project, it is taking time. The carrier needs to consolidate data for its CRM applications, including call centers and Internet applications to deliver information to agents. ""About three years ago, we came to the conclusion that we had to combine our data needs,"" says Morand. Most of the data resides in IBM mainframes, he says, and the company wanted to standardize on DB2 for performance and reliability. ""The data we have is voluminous."" American National had help from IBM Global Services for its initial implementation, and used Neon Systems, Inc.'s (Sugar Land, TX) Shadow Direct, an application platform to mainframe integration software. Shadow is used to provide a unified view of customer information by enabling access to the data from distributed systems.
However, showing interim results is vital to keep and gain support from business leaders. At American National, the company is moving quickly to put information in the hands of its independent agent sales force to help satisfy the ROI needs of management. ""The independent agents are free to sell"" other carriers' products, Morand says. ""We are trying to put technology and information in front of them so they will favor our products. The data warehouse is very exciting and is a strategic asset.""
IT leaders at Nationwide, also realizing that the data warehouse is a long-term investment, use ""success stories"" from the project to justify continued investment, says Cannon. ""We recently brought together data on auto policy premiums and linked it with losses,"" he says. ""That is the first time that has been available to underwriters."" Previously, he admits, underwriters could request to access the data but it could take as long as a week. ""Now the analysis can be done in a day. We can better understand the data and do a better job at pricing.
""We have put forth a lot of effort to develop quantifiable returns,"" according to Cannon. ""There are high capital costs with a data warehouse. Often it is hard to be specific about what you were going to get out of it. But we realize that sometimes we gain more understanding"" of the returns after the implementation.
ROI Tips for Data Warehouse Projects
In an age of short-term ROI, long-term data warehouse projects appear to be on the outs. But even large data warehouses can show near-term ROI. Here's how:
Since the industry is emphasizing cost reduction, focus on providing data to ""reduce inquiries from a call center by providing complete information,"" or ""combine mailings for customers with more than one product to save on mailing costs,"" says Ron Barker, senior principal and insurance practice leader at Knightsbridge (Chicago), a data warehousing solutions provider.
Once executives who control financing realize that the project can produce quantifiable savings, ""you begin to gain some traction,"" points out Mike Helms, insurance industry consultant, Teradata (Dayton, OH). ""Once you can show a return, executives will commit to the next round of funding."" Also, use an incremental funding method that requires that any returns from the project be reinvested in the next phase. ""The returns will help fund the next cycle,"" Helms adds. ""The more data you bring in, the more can be leveraged.""
And when the data warehouse is a fan (e.g., CEO) favorite, move towards softer benefits, Barker says. ""Softer benefits, such as customer retention and cross selling, are harder"" to justify, he says. ""Churn is expensive. If you have 15 percent churn and you don't know why, you are in bad shape. If you know why""and you'll be more likely to with a data warehouse""you can cut it off at the crossroads.""
Greg MacSweeney is editorial director of InformationWeek Financial Services, whose brands include Wall Street & Technology, Bank Systems & Technology, Advanced Trading, and Insurance & Technology. View Full Bio