How Home Depot is Building the Future of Retail with AI, Google, VR

By Pamela Bump |  January 10, 2018

While Amazon has hit many retailers like a merciless wrecking ball, $95 billion Home Depot has so far been unscathed. E-commerce, after all, isn’t yet a convenient option for the contractor or homeowner picking up some lumber, a box of nails, and a few gallons of paint to take care of the day’s project. In its most recent quarterly results, Home Depot’s same-store sales leapt almost eight percent, and the average customer’s purchase was up five percent.

Home Depot VP of Online Prat Vemana

But the company is exploring new technologies like artificial intelligence and voice-driven search, as well as running design sprints and building minimum viable products, with an eye toward creating a more “interconnected” shopping experience between stores and the online world.

From the company’s Atlanta headquarters, Prat Vemana, VP of Online, oversees the Interconnected Experience Team, which designs and deploys digital tools aimed at making in-store and online shopping seamless for customers, no matter how complex the project.

“It’s a little bit rare in the retail industry, but we’re blessed to be in that customer segment that loves to shop in both channels,” Vemana said in a recent interview with InnoLead.

Home Depot’s online platforms, including its mobile-optimized website and dedicated app, hosted more than 1.8 billion visitors overall last year. And while foot traffic in the company’s 2,200 retail locations continues to grow, last year’s Black Friday shopping period saw its mobile traffic surpass desktop traffic for the first time.

“Forty-five percent of the orders that are placed online are picked up in store,” Vemana says, adding that roughly 15 percent of the traffic to the mobile site is initiated when customers are already standing in a Home Depot store.

Vemana previously served as Vice President of Global E-Commerce, Product and Analytics at Staples Inc. There, he helped oversee the creation of a network of software development labs in Cambridge, Seattle, and Silicon Valley.

Sprinting Towards Innovation

Over the last four years, Home Depot’s app and mobile site have rolled out a number of new features catering to in-store shoppers — from locating a specific product to figuring out which tool is necessary for a project. According to Vemana, new app and site features are often the products of customer input and short design sprints.

“Every week, my product and design teams are in people’s homes or [at] customer job sites, where we are bringing in a lot of real-time insights from the customers,” he says.

“We have a three-day or three-week design sprint, depending on the size of the opportunity, where we go through different techniques of how to get from an idea to an MVP,” he adds.

With the three-day sprint, the first day is spent analyzing and synthesizing customer and stakeholder needs, while day two focuses on creating a prototype or design.

If the concept being prototyped requires minimal engineering or coding, it goes into store testing for customer input during day three.

“Already this year, we’ve done a few hundred tests,” Vemana says.

For the more complex concepts, he adds, “we take [them] through lab testing, which involves usability testing” to gather quick feedback.

Vemana says the most successful sprints are done in collaboration with staff from other areas of the company. Recently, he’s been working with store-based team members to determine if Google Home could be useful to help customers find their way around store locations in the Atlanta area. In this scenario, a customer could ask a Google Home-enabled device where to find a product.

“If you’re using the Google Home, what does it mean to actually talk and ask directions about where the product is? I can’t determine that without a store innovation team actually being part of the conversation,” Vemana explains. “That’s an example of where our team, a store ops team, and data sciences team are working together to understand the problems.”

Before Google Home can be brought into a store, Vemana says Home Depot teams would use a design sprint to ask questions like, “Do customers feel comfortable asking Google where a product is in public?” They also might use this time to discuss logistics, such where signs should be placed to alert customers to the new feature.

Every week, my product and design teams are in people’s homes or [at] customer job sites, where we are bringing in a lot of real-time insights from the customers. – Prat Vemana

In-Store Navigation and Augmented Reality Experiments

The Google Home possibility expands on a previous Home Depot app and mobile site feature launched four years ago. The “In-Store Locate” tool allows a customer to get aisle directions based on a specific product or category.

More recently, the app added GPS navigation to offer customers in select locations turn-by-turn directions to the area they’re looking for. If a customer is hunting for more than one product, this feature also shows them the quickest shopping route around the store. Customers don’t need to download the app; they can also text the product they’re looking for to an SMS number shown on posters around the store. “It’s one of the most popular features in our mobile properties,” Vemana says.

Home Depot has recently begun to roll out several offerings that use augmented reality and virtual reality to help customers visualize what a product will look like in a particular environment. Three years ago, the company launched a tool that allowed users to place faucets and ceiling fans into images of their home.

“We’ve been expanding the assortment [of products that can be placed into photos] into decor categories, like furniture and lighting,” Vemana says. And more sophisticated AR and VR experiences are on the way. A key first step is creating three-dimensional digital models of thousands of products, with accurate dimensions.

Using augmented reality to envision the end state of a renovation project.

Thinking About Three Buckets

As is the case with many other innovation teams, Vemana says that it’s a continual challenge to sift through good ideas to find those with the highest potential. “The number one challenge of driving an online team is picking the areas to focus on,” Vemana explains.

So when deciding if the team should proceed with a concept that emerges from a sprint, Vemana and his colleagues ask themselves if it “fills three buckets:” customer behavior, business proposition, and tech feasibility.

“If all three [bucket values] are known to have value, then it’s a no-brainer. We know [the product’s] value is clear,” he says.

One of the team’s more complex endeavors has been implementing two-hour delivery. The service is currently being tested for its “bucket value” and logistics in some Seattle and Las Vegas locations.

“People are ready for [the delivery service],” Vemana says. “We already have been delivering from store. We know that delivering from store increases the value,” he explained, adding, “That [testing] is more about, ‘Can we roll it out? What does [the feature] look like? Which markets make sense?'”

He added that the another “tricky” element of developing an idea is timing. His team may often weigh whether they want to launch a new technology at the earlier end of its “hype cycle,” or at its peak level of hype.

As an example, Vemana pointed to augmented reality, virtual reality, and machine learning technologies. While these tools have attracted plenty of buzz in recent years, Vemana explains that there have not been many use cases for them until recently.

“Technology is available and the customers are ready for it,” Vemana says. But it’s important to ask: “Is there a business benefit for it? Is it something we are ready for?”

Below, several slides from a presentation Vemana delivered in 2017, “Actual Intelligence Based on A.I.” The complete presentation can be found in our Resource Center.