J. Crew uses data warehouse and data-mining app to analyze sales trends from all operations.
Buying a new sweater, but not sure what shirt goes with it? Clothing retailer J. Crew Group Inc. is using a data warehouse and data-mining software to help online shoppers make those decisions.
The New York company began using DigiMine Inc.’s Enterprise Analytics data-mining software in mid-September to analyze sales data from its Web site, retail stores, and catalog sales operation. The data is collected in a 500-Gbyte data warehouse, running on a Microsoft SQL Server database that J. Crew and DigiMine spent nearly six months developing. The companies wouldn’t disclose the system’s price.
“The key is building the data warehouse,” says Jayson Kim, J. Crew senior director of Web marketing applications. Most companies, even those with retail and catalog sales channels, only collect and analyze clickstream data from their Web sites. J. Crew is combining clickstream data with product-sales data from the legacy systems that process transactions from its catalog operation and retail stores. “That gives us a more holistic view of our customers. Unless you capture all that data in one repository, you’re never going to do really effective analysis,” Kim says.
J.Crew is using the data warehouse and DigiMine tool to identify product-sale affinities–to determine what J. Crew clothes, shoes, and accessories customers frequently purchase together. That information is fed to applications running the Web site, which uses Art Technology Group Inc.’s ART Dynamo software. When online shoppers click on an item, the Web site recommends complementary products that the customer might be interested in buying.
The system appears to be increasing online sales, Kim says, but it’s too soon to conduct a financial analysis to quantify gains. Looking ahead, Kim says the data warehouse will help J. Crew analyze sales trends, build customer profiles, and generate product recommendations for E-mail marketing campaigns.