Roughly 4% of web visitors click and buy a suggested product.
Two years ago ComputerGeeks.com decided its homegrown recommendation engine wasn’t working.
“It didn’t make intelligent recommendations,” says Michael Buonomo, Internet strategist for Computer Geeks, No. 203 in the Internet Retailer Top 500 Guide. “It was human-managed and there was no real methodology behind it from a data or strategic thinking standpoint.”
As a result, few consumers clicked and bought the products the site suggested.
To offer more enticing recommendations, the retailer turned to a merchandise recommendation engine from Certona Corp. that generates product suggestions based on a shopper’s past purchase history and behavioral patterns.
The retailer placed those recommendations on its product detail pages, category pages and on its checkout pages.
“How often do you buy a pack of gum when you’re checking out at the grocery store?” asks Buonomo. “It goes along with human nature to buy something that entices you when you’re already making a purchase. This takes that concept to the next level because it goes well with the item that a consumer is purchasing.”
After two years with Certona’s recommendations roughly 4% of web visitors click and buy one of its suggested products. Moreover, a back-end tool that tracks consumer behavior to attribute sales has shown that the site’s recommendations have paid for the software 14 times over, says Buonomo.
That success led the retailer about six months ago to add personalized recommendations to its transactional e-mails that confirm the details of a consumer’s recent purchase. However, the retailer has yet to see much traction from those e-mails.