When a beer drinker cracks open a cold one, it's unlikely that business intelligence is top of mind. But for Labatt Brewing Company -- a subsidiary of the world's largest brewing company by volume -- BI offers its executives a comprehensive view of operations, which is essential for making timely decisions in the fast-moving consumer packaged goods (CPG) industry.

In 2004, Labatt already had a full suite of BI tools from Cognos, including ReportNet for advanced Web-based reporting. However, the company wanted a technology that could draw information from its transactional systems and allow users to analyze the data in ReportNet and other applications -- while simultaneously resolving any inconsistencies between data sources.

In August of that year, Labatt deployed data warehouse software from Kalido that sits on the company's Oracle enterprise data warehouse. Operational data resides in 11 UNIX transactional systems, which includes everything from information on distribution partners to customer segmentation data.

"The idea was to bring together targets from Cognos planning and actuals from our operational systems and report on all of it," says Jonathan Starkey, manager of data and application architecture at Labatt.

Labatt had two options at its disposal. "We could have installed an ERP system," says Starkey. This would have meant a bolt-on approach to the Oracle data warehouse by installing various add-ons to handle BI needs. The company considered tools from Cognos and Business Objects, but decided both were too analytics-based. "We needed more of a focus on supply chain, operations, and sales and marketing. The tool had to be more flexible," says Starkey.

The second option, which Labatt ultimately selected, was to employ a data warehouse application, says Starkey. "When your business is changing quickly, and your manufacturing footprint is being realized, it's quicker than installing an ERP system. The warehouse approach accommodates the changes that a business undergoes. And there is a minimum of data engineering," he says.

Traditional data warehouses are custom built by IT or a systems integrator, and there are limitations to that design approach. For starters, they are susceptible to maintenance when the underlying systems change, and it can take months for any changes to reflect in the overall business structure. According to Starkey, data warehousing software such as that from Kalido is allowing Labatt to reflect business change in a matter of hours.

Users at Labatt work with 48 dimensions and 93 transaction data sets. That work, according to Starkey, is made easier by using what's known as "effective dating" in the Kalido software. As far as databases go, normal referential integrity rules are structured to be point-in-time. As the dimensions of a company like Labatt's change often -- sales territories both are examples -- this limits users' ability to go back and analyze data after a change in structure has occurred. The difference with using a data warehouse application is that it manages such rules over time, says Starkey. Users can take data on everything from supply chain issues to brand offerings and make comparisons that will affect future decision-making.