Evan Schuman's StorefrontBacktalk

A woman walks into her local Wal-Mart and immediately turns to an area near the front of the store with a bank of screens. She logs in and is shown an array of products from her specific purchase history. But this customer is rushed today, so she tells the system to forget the recommendations and she selects 24 items from her shopping list.

The items that are on the shelf elsewhere in the store? An employee goes to fetch them. The ones that are not yet unpacked in the backroom? Someone in that area grabs those, too. All of the available products are brought up to customer service, where the shopper can pick them up without navigating the aisles. Those items not in-stock at that store? They’ll be shipped to her home. This is where Venky Harinarayan, Wal-Mart’s Senior VP for Global E-Commerce, head of @WalmartLabs and venture capitalist extraordinaire, sees the world’s largest chain headed.

Harinarayan has had his hand in a huge number of Silicon Valley retail ventures. He co-founded Junglee in the early days of E-Commerce, and then sold it to Amazon for more than $250 million. He worked at Amazon and co-invented Mechanical Turk. He’s since sold other companies to Google. In 2005, he co-founded Kosmix, which he sold to Wal-Mart for more than $300 million. He’s also a preferred angel investor of Facebook, which is now preparing a huge IPO. So it’s safe to say this guy’s predictions are worth listening to, very closely.

Wal-Mart has agreed to give Harinarayan’s team four or five Wal-Mart stores as working locations for testing a wide range of experiments. One of the ideas he discussed is the endless shelf referenced above, a place in-store where Wal-Mart can leverage not only purchase history and the preferences of that customer’s demographics, but also the mountains of social media observations associated with that specific customer. Tracking social media is what Kosmix focused on, and it’s what initially attracted Wal-Mart.

The endless-shelf concept goes well beyond letting the customer access—through a password—a customized version of Wal-Mart’s site. If a customer wanted to shop on her desktop device, it’s unlikely she would take the time to drive to Wal-Mart. But if it becomes a way to make the store more convenient, to save that customer the effort of walking into the store in search of products (when she’s pressed for time), this truly enhances the in-store experience.

“Envision this endless shelf, where (customers) can potentially access online within a store. A two-dimensional way to showcase many more products,” Harinarayan said.

It also builds trust. That means recommendations the system might make even more credible and persuasive. Also, this is a painless way to exercise a merged-channel approach. The customer can rattle off everything on her shopping list and say, “Get this stuff for me.” Regardless of where it is—in-aisle, in the back, online—the system finds and makes it available.

Harinarayan also spoke of a tweak on digital signage. Others have discussed the idea of using data about an identified and tracked customer to personalize ads and promotions displayed. But he envisions stores—”within a few months”—with big screens where associates and customers chat back and forth about products. “Where can I find this? Is it any good? Is there a better choice? Does it work?”

“We’re thinking about what a social layer in the stores might look like. If we can let (customers) communicate with one another,” there’s no telling what might happen, Harinarayan said. And, yes, they will be using various filters to prevent obvious naughtiness from filling those big screens. (OK, Wal-Mart shoppers: Start now thinking up creative non-obvious naughtiness.)

The endless-shelf concept would likely have the customer logging in. But is that necessary? And can customers be identified without requiring them to log in each time they shop? Harinarayan says a lot of that could be addressed through mobile tracking. “Mobile is going to be a very important piece of this puzzle,” he said, adding that “finding ways to help people shop in the physical store” is a priority. “For a customer to do this, you have to deliver a custom experience.”

Still, some aspects of customization and customer tracking are risky, in that customers could find them too intrusive or simply creepy. Tracking and customized pricing (where shoppers could use their smartphones to scan products and the pricing could be different for individual customers based on their shopping history or other factors) are two such risky examples.

“The notion of differential pricing is a big, big hoop to jump for many retailers. It is fairly controversial,” Harinarayan said. “Just because it’s possible doesn’t mean a customer will want us to do this.”