The growing number of influential Weibo commentators are increasingly opening their own online shops or promoting products.
Beer and diaper customers are in the same shopper demographic: who knew? Technology-based recommendations uncover such associations and drive online buying, Be Free says.
Technology developer Be Free Inc. says that technology will never replace merchandisers online – but as the recent experience of one of its clients shows, technology certainly can help. Using Be Free’s BSelect technology, a product recommendation engine, a large multi-channel apparel retailer, that Be Free will not identify, saw the volume of total sales coming from clicked-on product recommendations zoom to more than 30% from its earlier 7.5% rate within one month of implementation. Average revenue by day from clicked-on product recommendations rose to $656 from $129, for sales that totaled almost $16,000 higher per month just from recommended products. Previously, product recommendations had been based on merchandisers’ recommendations.
Be Free has been known historically for its affiliate marketing network, which still represents the bulk of its revenues, but in 2000 it released BSelect after acquiring another company that had built the core technology, based on Computer Learning Theory initially developed by Harvard University scientist L.G. Valiant. The results for the apparel retailer were from a new version of the original product, released about six weeks ago. The software improves the hit rate on recommended products by using a shopper’s historical purchasing behavior, gathered via a cookie, crossed with trend data to come up with more personalized product recommendations.
The data generated can be used to launch the personalized recommendations on a page, or it can be used by merchandisers to select and group products for targeted online promotions. Be Free CTO Sam Gerace tells Internet Retailer the product is able to generate relevant and personalized recommendations through analysis that uncovers trends and associations between customers who buy various products that might not be immediately apparent, but which can inform merchandising.
For example, “People who buy diapers also buy beer,” says Gerace. “The correlation is astounding. Arguably, when I go down to my local beer distributor they should have a display of Pampers. Intuitively, it makes sense. It’s an age group of young parents and that’s the kind of consumption that interests them, but it’s not an association a merchandiser would necessarily come up with on their own.”
Another example, Gerace says, might be merchandisers accustomed to grouping complementary cosmetic products with lipstick. "It often turns out to be the opposite of what you’d expect. Women aren`t looking for a whole array of matching makeup when they are shopping for lipstick; they`re looking to refill their lipstick drawer. So what you ought to be recommending if you’re featuring a light shade of lipstick is a medium and dark shade that goes with it,” he says.
The product delivers intelligence for clients from data drawn exclusively from individual merchants. Gerace says the company has the capacity to pool data from different merchants, but so far, BSelect clients have not authorized any pooling of data, though they`ve expressed more interest in the subject recently. A possibility that’s closer in is that small groups of non-competing companies pursuing the same customer demographic might agree to pool data to find associations that can inform e-mail merchandising and marketing, Gerace says.