Golden Copy: Tread Lightly On Data Warehouses

As part of a panel discussion on data quality and governance at the North American Buy-Side Technology Summit in New York last week, an industry tendency to continue working with legacy data warehouses arose as a matter of concern.

Generally, firms have not yet figured out what to do with legacy data systems, or what they should do, said Paul Zajac, managing director of information technology at Credit Suisse Asset Management, speaking after the panel session. Tara Castleberry, vice president of enterprise data management, program delivery, AIG, put it more bluntly, saying legacy systems can't just be thrown into a lake.

Castleberry, during the session itself, identified some questions a data manager should ask before taking action about any legacy system—including whether IT needs to be "micromanaged" by the business, and whether IT should be able to try something but fail. New technologies, such as graphing of databases to produce topological pictures representing data, can help firms see where opportunities to enrich data exist, she said.

Zajac had another relevant question—how are firms dealing with volume and aggregation across all their businesses?

If these remarks are an accurate barometer of industry thinking, firms are becoming more aware that they have to ask themselves what data system and management they would use to replace a legacy system, and how their internal IT culture handles data management challenges.

One could still debate whether a stronger push is needed—or should be applied—to remove outdated legacy systems. If a firm wants to pursue that possibility for its data warehouses—and therefore its data management—it's becoming clear that cannot be done rashly or lightly.