How to track web-influenced sales at the store

Anyone analyzing web retailing quickly understands that the web is not just an online buying medium. As important to retailers is the influence that the Internet exerts in creating sales at the store that are the product of online research. How to measure the web’s impact of those sales, however, is a difficult proposition and an inexact science. “Multi-channel retailing is more than just three channels (store, catalog and web),” David Towers, vice president of e-commerce operations for J. Crew Group Inc., told his audience at last week’s Shop.org Annual Summit in New York. “It’s three squared, because you have three channels for browsing and three for buying, and there are all these different iterations of customers you serve. How am I able to measure who’s doing what and where, and how can I hit that behavior?” Knowing the answers to such questions, said Towers, can determine “when I invest and when I de-invest and, for example, not send a catalog to someone. This is a matter of optimizing costs across the organization and of proving the worth of all of the channels so that you optimize all of them for all your customers.”

J. Crew understands the synergy among the store, the catalog and the web as well as any retailer. Though it started as an apparel cataloger, it branched into mall stores that now vastly exceed catalog volume, and its 7-year-old web site this year will generate $140 million in sales, for the first time exceeding what the chain’s catalog generates in sales over the phone. But measuring who is doing what across all sales channels, conceded Towers, “is one of our biggest challenges,” one that is made more difficult by regulations and legislation protecting privacy. Retailers, for example, could identify store shoppers and web shoppers by capturing credit card information at the store and matching that with credit card data captured in web transactions, something J. Crew did at one time.” Towers asked his audience whether they used this technique, and when no one raised a hand, he added, “Good, because that practice is now illegal.”

Technology abounds for measuring shopping behavior at web sites and catalog call centers, Towers said. The real problem is identifying the origins of a store sale—whether it came from a new or repeat customer or from a loyal catalog customer or from a first-time web visitor. How J. Crew takes on this challenge has more to do with creative merchandising than scientific measurements. “We are looking at coupons on the web to maximize shopping the store,” Towers said. “We also want to be able to create highly targeted and personalized e-mail messages to get web shoppers to come to the store. We want to be able to use personalization technology so that we don’t give a broad message but a compelling personalized message to visit the store. We want to say to the customer, get up from your Dell, go to the store in Austin near you, here are the directions, park in lot B, walk into the mall just past the Orange Julius and into our store. There you will find a sweater in stock that is a perfect fit with the jeans you just bought online. That’s the type of personalization we are trying to achieve.”

In essence, Towers told his audience that trying to drive web shoppers to the store is likely to be more fruitful than trying to measure precisely how much store volume is influenced by the web. “We are trying to measure the sale impact of the web on the store,” he said. “But we approach the problem by saying here’s what we can do as action items on the web to drive more people to the store. We aren’t saying, see, we are driving a lot of people to the store.”