Kalido lets Unilever compose Sinfonia
Unilever Latin America has embarked upon an enterprise-scale data warehouse and business intelligence project called Sinfonia that will eventually become one of the largest databases in the world, according to Andy Hayler, founder and chief strategist at Kalido.
At the heart of Sinfonia is the Kalido Dynamic Information Warehouse, a data warehouse solution for preparing, implementing, operating and managing data warehouses. It's designed to work with database solutions from IBM, Oracle and others.
Unilever Latin America manufactures and markets hundreds of consumer packaged goods across 34 companies in 19 Latin American countries. It expects the Sinfonia project to save the firm millions of dollars from economies of scale in the extended supply chain and cross-border business opportunities.
Unilever is able, for example, to test what happens in different markets if it raises the price of a bar of soap. In some markets, the price change will have no impact on sales and, consequently, the company stands to profit. Unilever is also able to review the history of several changes or the impact of a single change over the course of a year.
The project has generated immediate savings by reducing administration costs and streamlining systems management and processes across the region.
In a typical data warehouse, all data is linked and changing one bit of data has an impact on all the data it is linked to, thereby "breaking" the data warehouse, Hayler says. Kalido requires more work up front (entering the data and creating links), but once that's done, subsequent changes won't break the warehouse, he notes.
In addition, a Kalido data warehouse in Latin America can talk to another Kalido data warehouse anywhere in the world, a capability the company calls federation. "You can imagine the benefit of a global company having a data warehouse in one place exchanging information with data warehouses elsewhere," Hayler says.
Michael Alexander is editor-in-chief of Application Development Trends.