To data warehouse, or not to data warehouse? That is a central question in the business intelligence (BI) debate as insurance companies react to the financial problems that have plagued the industry over the past several years. The answer to this question depends on whom an insurer turnsto for advice. Data warehousing and most business intelligence application providers will say you need a warehouse. CRM and ERP vendors will tell you to glean information from operational data. But the real answer seems to be a definitive, It depends.

If a company wants to gain a greater understanding of its customers and products, it needs a history. Thats what data warehousing is predicated onthe ability to standardize on an approach that makes different data transactions comparable over time, says Keith Gile, senior industry analyst at Giga Information Group, Cambridge, Mass. At the same time, business intelligence is the process of making any and all data assets actionable, and more often than not, a companys day-to-day operational systems generate tactical data that can also be interpreted through business intelligence.

You need to ask yourself if you really want a BI tool going in and accessing all of your production system platforms, or would you like to extract data from these systems and output them to separate databases, data marts, or warehouses, says Cal Braunstein, chairman/CEO and executive director of research at the Robert Frances Group, Westport, Conn.