JD.com and Alibaba create indexes to identify Chinese shoppers’ spending trends, which help retailers gain insight.
It's certainly more refined than looking into a crystal ball and guessing the future, but more Web retailers are relying on sophisticated personalization software integrated with the right databases to predict what items shoppers will impulse buy.
Personalization software uses sophisticated mathematical models linked to an Internet retailer's inventory, merchandising, and transaction history databases to predict what products a shopper may buy spur-of-the-moment. Some personalization applications ask the shopper to fill out a series of electronic forms and create a customer profile.
The next time the shopper visits the site and clicks on a particular merchandise category, the retailer's "frequent customer program" calls up the profile and begins offering them products they're likely to buy such as a new CD from a favorite artist.
But new product recommendation engine technology using collaborative filtering goes far beyond just individual profiling. This advanced software considers what item a shopper is looking at, searches through integrated databases for profiles of other customers who've bought that product, identifies what else those people purchased and then uses that information to make other merchandise selections to the original customer.
For example, it seems highly unlikely that a customer shopping online for a $25 blue oxford shirt will also wind up buying an $85 cigar humidor, but that's precisely what's happening on sites such as Skymall.com. "The software analyses hundreds of similar customers and their buying patterns and convinces the buyer to say to himself 'Hey, how did they know I wanted one of those," says Steve Larsen, vice president, Net Perceptions, Eden Prairie, Minn. "The technology creates a wish list of items that tells them they are in the right spot at the right time to buy exactly what's being offered.