How do you get people to buy more stuff on e-commerce Web sites? That’s the $1 million question, literally. RichRelevance and Overstock.com are offering the $1 million RecLab Prize — aimed at academics and other geeks — for the team that builds the most powerful online recommendation engine.
The idea is to create advanced algorithms which help online retailers more accurately show products that might be of interest to the individual shopper. Obviously, there’s big money to be made in personalized product recommendations, a concept that Amazon.com helped pioneer. In 2009, Netflix offered a similar $1 million prize related to movie recommendations.
RichRelevance Chief Scientist Darren Vengroff, a former principal engineer at Amazon.com, tells GeekWire that researchers competing for the prize will be able run their recommendation algorithms against live shopper traffic for the first time.
“We want to advance the state of the art by enabling really smart people in academia to work on some very interesting real-world problems,” said Vengroff, who works out of the company’s Seattle office.
In a lot of cases, Vengroff said that researchers work on the best approximations to the real world that they can find. However, in many cases, those situations miss key elements of context and feedback.
“The RecLab approach finally solves this problem by bringing their code to the data rather than releasing the data into the wild,” he says. “When we advance the state of the art in this way, Overstock benefits, RichRelevance benefits, and the contest winners obviously benefit as well.”
The prize will be awarded to the team that builds an algorithm that delivers a 10 percent or greater lift over existing product recommendations on Overstock.com. Nominations close on December 1, with the winners to be announced next March.
Overstock.com has worked with RichRelevance since 2009, using the technology to more accurately predict what people may want to buy.
