Inside Walmart's Social Shopping Skunkworks

The online social world can put consumer actions in context, which makes the targeting of retail offers eerily more precise. No wonder the world’s biggest retailer also wants to be the world’s biggest social-shopping company. Here’s how Walmart's @WalmartLabs division is doing it.

Retailers have long relied on transaction data to get a sense of their customers. Buy Fage Greek yogurt at the supermarket and you may receive a coupon for competitor Chobani's yogurt with your receipt. Buy a Batman DVD and you may get advertisements for other superhero films.

But what if didn't buy that DVD for yourself, as the retailer's messages assumed? What if it was a last-minute gift bought on your way to a friend’s birthday party? That context was not captured by the transaction data.

“Transaction data often works, but it often fails, in large part because it’s so narrowly focused,” said Robert Yau, the director of project management for @WalmartLabs, the retailer's social strategy division. “With the growth of social media, we have more data points we can use to better target recommendations.”

Yau was working for Kosmix, an advertising-supported Internet directory, when Walmart bought the firm last year and renamed it @WalmartLabs. (He will be the keynote speaker at SocialDiscovery.org's Social Discovery Conference, which will take place August 6-7 in San Francisco.) Already, his team has developed Shopycat, a gift-finder app built on top of Facebook. 

How Interested is the Consumer?

The key to delivering effective retail messages, Yau says, is identifying a customer's interests. Social media provides context that makes it possible to do that. 

Yau breaks customer interests into long- and short-term. A Facebook profile is particularly useful for identifying long-term interests. For example, Yau's own Facebook profile says he is an avid runner - and has since he created the it - signaling a long-term interest. Counterintitively, he views long-term interests as a negative indicator for sending retailing messages related to those interests.

“People who have those long-term interests are very knowledgeable and know what they like,” he said. “It’s very hard to make a recommendation to me about running shoes.”

Short-term interests are more promising, since potential customers haven't yet developed biases such as brand loyalties around them. Retailers can divine shoppers' short-term interests by following newsfeeds (in aggregate) rather than profiles. On a given day, a West Coast heat wave might prompt a migration to the beach in Los Angeles and a cloud of barbecues in San Francisco. As fall approaches, feeds might point to renewed strong interest in the television show Glee.

Contextual data like this can guide programming everything from Walmart.com to individual stores.

The Promise of Pinterest

Yau is also watching for ways to take advantage of social discovery. Sites like Pinterest help users discover new interests and products by presenting collections that have been curated by other users.

Social discovery is “huge for us,” Yau says. However, he views it a long-term opportunity because will take time to a way to find a way generate consistent returns from Pinterest data. “There’s a lot of expecation, but no one has really nailed it down yet.”

The Privacy Issue

Yau is quick to point out that the company collects social data on individual customers only if they opt in through the Shopycat app. But with a billion shoppers, Walmart has access to loads of social data that can help pinpoint regional trends and stay ahead of what shoppers want.

“There are a lot of misconceptions around social shopping and social discovery,” Yau said. “The way I look at it, we go to the store with an idea of something we need to purchase, and the retailer wants to try and upsell us. If you’re going to try to upsell me, you need to get it right, and social data helps us better focus those offers. It gives us the ability to figure out that Batman DVD was for a friend and not something you have a personal interest in.”

ReadWriteWeb’s Dave Copeland will be speaking at the SocialDiscovery.org’s next Social Discovery Conference on Aug. 6-7 at the Fairmont San Francisco. ReadWriteWeb readers can get a discount of more than $1,000 if they register by August 3.