Don’t Trust Your Instincts? Wal-Mart Uses Algorithms to Find Gifts People Want.
Wal-Mart has launched its first Facebook application that helps people buy better gifts for their friends.
The application, called Shopycat, makes product recommendations based on the items people have liked or talked about in their news feed.
It is not particularly flashy. The logo is a picture of a cat sticking its head out of a shopping bag. The tagline reads “the right gift every time.” But the app’s rudimentary design fits nicely with the experimental nature of social commerce. Retailers are just starting to figure out ways to leverage social networks. Wal-Mart is no exception.
Wal-Mart promises that while it looks simple, the technology running behind the scenes is more complex than you might imagine.
The product was built by @WalmartLabs, which was created after Wal-Mart acquired Kosmix, a Bay Area start-up that was tracking social networks to determine people’s interests.
Venky Harinarayan, founder of Kosmix and co-founder of @WalmartLabs, said developing the application was surprisingly difficult. First, they had to find the relevant information on a person’s Facebook page. Then, they had to find products that best matched those interests.
For example, if someone likes Lady Gaga, the most obvious product to recommend is her albums. But a fan would likely already own those. A better gift is something more special — a collector’s item or a limited edition. That’s a more complex problem.
Since gifting is a practice humans naturally struggle with, maybe algorithms can do a better job.
After using Shopycat, Harinarayan learned his wife was a fan of “Game of Thrones,” the TV series on HBO. She has posted several times on Facebook about the show, but he hadn’t noticed. “Facebook is so transient and things flow by. Here’s a way to aggregate it all and put it in one place,” he said.
The technology also sometimes fails. When Harinarayan viewed gift ideas for me, it recommended a number of Sony products. Months earlier, I liked Sony’s fan page in order to gather information for a story — not because of any deep admiration I had for the company’s products. An algorithm would have a hard time knowing that.
“I think it’s a good first pass,” Harinarayan admits. “But there’s areas we should get better at. Right now, it is our goal to give one good gift idea per person. If we do that, then we are successful.”
The application is available to Wal-Mart’s 10 million fans on Facebook if they give permission to install it.
Harinarayan said Shopycat is the mega-retailer’s first foray into using social and promises that more is coming soon, including social experiences in the store and on its Web site.