How analytics backs up the human cross-sell decisions at Timberland

Human knowledge works best when it comes to offering cross-sells online, but it works even better when decisions are verified by analytics, says Tim Diaz, director of e-commerce for The Timberland Co. “We take a non-traditional approach to determining what people will buy online,” Diaz says. “A lot of people spend lots of money on clickstream analysis to let a machine tell them what their customers are buying. We hired people who worked in the store who tell us, ‘I noticed when someone bought these boots they also bought socks.’”

Timberland checks the merchandisers’ knowledge, however, by running sales and other actions through an analytics program from Omniture Inc. “We have tools to back up the cross-sell decisions,” Diaz says. “We analyze clickstreams and item affinity as a 2nd way to check on management decisions.” Timberland also analyzes the previous-page and next-page reports from the Omniture system to see where customers came from and where they went after viewing a product.

A drawback to using human cross-sell knowledge, Diaz says, is the difficulty keeping up with cross-selling. “We were surprised at how daunting a task it can be to manage online where every item needs copy and imagery,” he says. He says the effort has been worth it, with a steady increase in average order value and average items per transaction. “If we get a 10% increase in units per transaction, that’s very good,” he says.