In a war waged against competing insurers, carriers must rely on highly specialized intelligence reports to gauge their bearings and affect strategies accordingly. But before multiple business units can begin advancing in lockstep, IT leaders must first quiet the battles that rage within their organizations. "Over the past ten years, larger organizations have followed the pattern of acquiring business intelligence [BI] solutions at the department level, and they ended up with various tools. [Often], none of them worked together," relates Keith Giles, senior analyst, Forrester Research (Cambridge, Mass.).
Compounding the problem is the reality that, "Carriers across the industry still silo information," asserts Jeff Donaldson, managing director of business consulting firm BearingPoint (McLean, Va.). So it's no surprise that, "One of the more difficult political battles takes place [between business and IT units], when IT needs to standardize BI tools," he says. "Oftentimes, critical applications within certain lines can't be replaced easily, if at all."
But why would a forward-looking IT department get itself in this position in the first place? Forrester's Giles suggests that a fragmented approach led by business leaders is the root of the problem. In fact, many times, internal IT groups were not even made aware of such tools' implementations, he says. Frequently, BI solutions that proved their effectiveness at a departmental level become so popular that user numbers can jump from 10 to 500, relates Giles. "Units would begin to experience crashes and degradation in response times because they weren't designed for large groups. Consequently, business unit heads became overwhelmed and handed their problems off to IT."
To address similar issues, CIBC Insurance Companies (Toronto, $280 billion in assets) recently began a BI consolidation initiative. The insurance carrier "had a basket of different BI technologies," concedes Kal Omran, general manager, MIS and business operations, CIBC. "It was disorganized and everyone was using different tools. We needed to consolidate on a unified platform." Omran decided to replace CIBC's Microsoft (Redmond, Wash.) and Bristol Technology (Danbury, Conn.) BI tools with Cognos' (Ottawa) Series 7 Impromptu for managed reporting, and PowerPlay, an online analytical processing (OLAP) software application used for analyzing large volumes of data.
Currently, the tools, which were implemented in late 2002, are primarily used to monitor sales and new product performance. "CIBC's business development managers are interested in knowing about marketing opportunities and what is not selling well," reports Omran. The performances of five products have been analyzed with the system thus far, he says, and CIBC plans to roll out six more products on the system in the next six months.
To expedite the process of adding insurance products to the system, Omran utilizes a generic data model when assembling CIBC's data warehouses. "We designed them in a way so that we don't need to make any changes to add a new product," Omran explains. This way, "Turnaround time isn't huge unless we are dealing with a whole new product."
Data warehouses, like the ones used by CIBC, hold information for the sole purpose of reporting. Currently, Kentucky Employers Mutual Insurance (Lexington, more than $220.9 million in assets) is working with BI solutions provider Thazar (Frisco, Texas) to develop the insurer's data warehouses. The carrier has recently awarded a contract to Thazar for use of its INSsight Business Intelligence System, relates Mike Booth, IT director, Kentucky Employers Mutual.
In addition to data warehouses like those employed by CIBC and Kentucky Employers Mutual, carriers can enhance the efficiency of their BI processes by utilizing cubes. According to Leo Tucker, director of banking solutions, Cognos, cubes are pre-aggregated measures navigated or filtered by dimensions. Typically, the source of the measures (dollars) and dimensions (customer, time, place) are drawn from data warehouses. These stores are "not intended to hold all of your data, just the key pieces that you report on, on a frequent basis," asserts Tucker. "If you rely on a cube, information can be drawn much more quickly."