Smart and cheap: How to improve business intelligence on a tight budget

Cognos 8 BI suite to identify and develop high-value products that couldn't be easily commoditized by its low-cost competitors.

It started by purchasing transactional data from retailers in the toy, fashion and apparel markets, adding that data to its existing data warehouse, and analyzing buying trends. Creativity also uses Belmont, Mass.-based Smart Software Inc.'s SmartForecast forecasting software.

All that analysis has led to more "design-oriented, fashion-oriented" products, such as a line of paper dolls based on the television show Project Runway.

The strategy appears to be working. Creativity's fashion-based offerings and its other unique designs have become the dominant portion of its business, accounting for more than 50% of its sales and representing an even greater percentage of its margins, Mulholland reports.

4. Centralize Business Intelligence

To help identify the right areas to focus on, Mulholland started an analytical "center for excellence," a group that includes representatives of different parts of Creativity, from sales to operations. "You're trying to elevate the IQ of everybody in your company in terms of knowing the key business metrics and measuring them accurately and in a timely way across all areas of the business," he says.

Moving toward that goal, Creativity developed common tool sets and profitability models for its sales and finance groups. Reports are pushed to the desktops and viewed in dashboard applications. From there, Mulholland says, users "can go in and do further analysis."

IBM has been promoting such centers among its Cognos customers as a way to create a standardized set of models using existing business intelligence tools. A set of BI dashboards developed for one department, for example, can be extended to other units. That way, new groups don't have to reinvent the wheel and can get up and running more quickly.

5. Put More Data in Your Warehouse

The current economic downturn is a great time for organizations to review what they're tracking and to add more data from business operations into their data warehouses to find additional savings. Just be very selective about what you add, experts advise.

Anne Milley, director of technology product marketing at SAS Institute, suggests adding data from call centers, Web logs or other sources. The question companies have to ask in these times, she says, is, "What do I have that I can get into the data warehouse at a relatively low cost?"

As sales slowed at Creativity during the downturn, Mulholland and the center for excellence team changed their focus from keeping up with growth to cutting costs. One project involved providing a feedback loop linking the back-end ERP system and the CubiScan system that's used for shipping.

CubiScan is a laser-based scanning and weight-measurement system from Quantronix Inc. that's designed to ensure that goods are properly packaged to meet customer specifications. (If they're not, the penalty fees can be "considerable," Mulholland says.) While the ERP system issued packaging instructions with the orders, the stand-alone CubiScan system wasn't returning data on whether shipments were actually packaged properly -- and many were not. "There was no feedback loop," Mulholland says.

The IT team used the Cognos ETL (extract, transform and load) tool to bring the CubiScan data into its data warehouse and then built exception reports for shipments where specifications for package dimensions hadn't been met. Mulholland expects the project to pay for itself in three to five months.

6. Make Better Use of Data You Already Have

In some cases, doing "more with less" may simply be a matter of taking