It's that time of year: The malls are full of poinsettias, a jolly Tim Allen is on the big screen, and whether by hunch or by data crunch, retailers are counting on customer knowledge to bring good cheer to the bottom line. For many companies, up to half of all sales come in the last two months of the year. Forrester Research reports that online holiday season retail sales in the United States will hit $27 billion, up 23 percent from 2005. Will investments in data mining, Web analytics and customer data integration tools prove their mettle?
Even as retailers invest in advanced technology, the business itself grows more complex. Traditional channels must make room for the online channel; sales and marketing dollars must be deployed across search engines and portals, not to mention all the usual places, such as catalog mailings. Retailers have also consolidated, which means that enterprise business goals must be met, even as lines of business push hard just to meet their own revenue and customer service objectives. 数据挖掘研究院
"Each line of business is different," says Aaron Cano, vice president of customer knowledge at 1-800-Flowers, which has focused this year on integrating brands and building its "information delivery framework," or enterprise data warehouse. "People often don't realize that our company now has a diverse gift product line that includes Plow & Hearth, The Popcorn Factory, HearthSong and Cheryl & Co. The business model for cookies isn't the same as the one for garden merchandise, which makes defining a customer and establishing a single version of the truth challenging." 数据挖掘研究院
The company wants its call centers and online sales engines to know that a first-time caller to 1-800-Flowers is a frequent, highly valued customer with HearthSong--and to treat them appropriately. "The biggest challenges are to improve the data flow and to build out a business rules engine that can do a lot of this in real time," Cano explains. Having spent nearly 15 years in banking, he is familiar with modeling, forecasting and the back-office application of rules engines. "But none of that meets what we're looking for in retail, where we need to know how to differentiate offers in real time, while the customer is on the phone or online." 数据挖掘研究院
Already a user of SAS Enterprise Miner, 1-800-Flowers is working with SAS to develop a customized retail solution that packages intelligence and rules. The company's SAS investment reduces the data processing involved in preparing for enterprise analytics, but 1-800-Flowers has no intention of outsourcing data processing for mining and predictive modeling to third parties. "If an external provider builds the models, I learn nothing," Cano points out. "You learn from the data, not from the model, which is just the end game. By working with the data to develop the models, we learn so much about our customers and what's driving their purchases."
1-800-Flowers gathers data coming and going: that is, from both buyers and recipients of purchases. But transactions only tell so much. As search engines mature and make better use of taxonomies (see our cover story), customer clicks will offer even more insight. "We'll be better able to let personal customer interest drive where we position products on the Web site," Cano says. 数据挖掘研究院
Across the retail world, Santa's data-savvy elves sound readier than ever for business. Here's to a holiday season in which your wishes come true--and get delivered on time. 数据挖掘研究院
