The North Face sees A.I. as a perfect fit (+ video)

26.01.2016
The North Face wants its customers to get the perfect jacket for whatever they're doing - skiing in Vermont, ice skating in New York City or just trying to stay warm on the way home from work.

So how do they give online customers that extra assistance and know-how Artificial Intelligence, or A.I.

"The issue is that online shopping over the past two decades has been about a grid of products on a white background," said Cal Bouchard, senior director of e-commerce at The North Face. "That's how customers find their products. We've developed onsite search and navigation, but we still made the consumer do the work. We wanted to take the conversation you might have with an associate in a store and see if we could put that as a service online: 'Here's what I need. Here's what I want.'

"We want to make it much more personal and much more intuitive," she added. "Consumers are going to get smarter and smarter and say, 'Really This is all you got' They want some help. They want some advice. You can go to Google and do all the work or you can give them advice. Consumers are going to demand something new from the online shopping industry and I think this is it."

The North Face is a 48-year-old, U.S.-based retailer that focuses on outdoor apparel and products like jackets, boots and backpacks. Based in Alameda, Calif., the company has about 60 retail and outlet stores in the U.S., as well as locations in other countries, including Canada and the United Kingdom.

Bouchard said the company decided to use A.I. when trying to develop a way to make it easier for customers -- particularly new ones -- to find just the right jacket out of the 350 or so the retailer has available. It can be a daunting job to figure out what buyers need in terms of weight, protection and style.

A.I. comes the closest to giving a buyer an experience that feels like a human sales associate is helping them choose a jacket.

"To sort through all those jackets is hard and they're all designed for different activities and environments," Bouchard told Computerworld in an interview at the National Retail Federation's Big Show conference in New York last week. "A.I. is most powerful when there are more choices than you can handle.... It's still a computer, but it's the most human interaction possible."

The North Face teamed up with IBM to use its Watson natural-language, machine learning system as well as software builder Fluid; it created the Expert Personal Shopper (XPS) software that uses the Watson technology.

The A.I.-powered online shopping assistant, which has been 12 months in the making, is live on The North Face site. Once in the jackets and vests section, click on Shop with IBM Watson.

The A.I. system asks questions like, "Where and when will you be using this jacket What type of precipitation do you expect" And "What kind of activities will you use this jacket for"

The system then assesses the criteria, sorts through the selection of jackets and offers up its best picks.

"It's at least not just grabbing pieces of data about you and putting them in a program," said Bouchard. "It's coming out and asking you questions and it takes in your answers."

So far, it's been working well.

The system went live in mid-November with the first part of a three-phase rollout. At first, a small group of customers was invited to try it out. Then, "elite" customers were allowed to test it. Finally, in mid-December it was pushed out live to everyone.

According to Bouchard, 75% of the customers who have tried it said they would use it again.

"When we read the comments, we found the people who needed the most help found it the most helpful," she added. "They might not talk jacket technology, like powder skirts and Gor-Tex. People who don't know that lingo found it the most useful."

Customers who know a lot about jackets and what they need found the A.I. system less helpful than newbies and were often quick to discount other options presented.

With that in mind, Bouchard's team is going to continue working on the system to make it more appealing to a larger pool of customers.

"I think A.I. is a game changer," said Bouchard. "Where we are right now is the first step. This didn't change our business in December but it will change the game. More people will engage our products. We're not as interested in the fast sell. We want to engage people in our brand and get them stoked on it. It's going to change online shopping over the next five years."

Jeff Kagan, an independent industry analyst, said The North Face is cutting edge when it comes to figuring out how to use artificial intelligence in retail.

"A.I. is one of the most exciting trends in retail and it's only going to be getting better," he said. "We're still in the first inning of this new game. Cutting-edge brands will use this as a competitive advantage until the rest of the marketplace catches up to them."

Rob Enderle, an analyst with the Enderle Group, said the technology will allow retailers to give their customers a seemingly more personal shopping experience. "It will more effectively allow for highly customized user experience at massive scale," he said.

Setting up the system wasn't easy, though. The smart system needed teaching before it could be used.

"When we first started working with the IBM Watson technology, I thought I was getting the Jeopardy champion right out of the box," she explained. "It didn't know everything. You have to teach it about your consumer and you products. That was a lot of work. I underestimated that."

It took a year for Bouchard's team to input information about the company's product catalogue and weigh the importance of different aspects of a jacket, such as whether it's more important to keep someone warm or dry.

"You have to test it a lot," said Bouchard. "A.I. is not a decision tree where you know if the consumer answers yes to this and no to that, the answer is this. A.I. is much more fluid. You have to test it a lot to find the holes in the logic. You have to spend a lot more time testing than with other software.

"...We tested for a long time," she said. "They'd come out with a release on Tuesday and we'd test it on Thursday and they'd work on it over the weekend and we'd test it the next Tuesday."

Right now, The North Face is working on a new design that will roll out in April and is designed to be more interactive and more mobile friendly.

Before the holiday shopping season later this year, Bouchard expects the system to have gained a lot more content and product information. With that added data, the system won't just recommend a jacket but will be able to offer up more info about the technology behind the jacket, as well as links to related information, articles or blogs.

"Considering how much we've done over the last 12 months, we're so much smarter now," said Bouchard. "I think we can get a lot more done over the next 12 months.... I think we're ahead of the game here."

(www.computerworld.com)

Sharon Gaudin