Skip to content

Machine learning will revolutionize ecommerce — end to end

Join our mailing list

Signifyd regularly publishes free reports packed with business insights, commerce trends and data from our massive Commerce Network. We’ll only email when we have something meaningful to share, no more than once per week. And of course you can unsubscribe any time.

It’s hard to talk to any retailer for more than four or five minutes without the phrase “customer experience” coming up.

And why not? Building great experiences for customers is not only how you get them to buy what you’re selling, it’s how you get them to come back again and again and buy some more. It’s why so many businesses pitching their products at this week are promoting tools that provide personalization, better search, merchandising insights and new ways of shopping online.

It’s a great conversation to have — the one about how to turn to technology to do better marketing and merchandising. But it doesn’t go nearly far enough.

In fact, the curious thing about the customer-experience conversation happening at the annual gathering of ecommerce professionals is that it tends to stop at the point that the customer hits the “buy” button.

Think about all that happens after a customer converts on the web — payments, fraud detection in ecommerce, picking and packing, delivery, post-purchase customer support. And then think about what a lousy customer experience it is when any one of those steps breaks down.

Bowen Smith, who was manning his company’s booth Wednesday, says retailers spend all kinds of time and money trying to perfect the front-end of the customer experience — marketing and merchandising. But that’s just a start.

“It all gets down to getting the right item in the right box, processing the return as quickly as possible, having adequate inventory levels,” says Smith, a vice president with Promotion Fulfillment Center, a third-party logistics company.

None of which is to say that the discussion about the back-end of the customer experience and the technology tools to improve it was completely absent from 2017. In the middle of the exhibit hall, Dave Barrowman, vice president, innovation for Skava, stood on a stage Wednesday morning and talked about how the discussion of machine learning in ecommerce had been too limited.

“In retail, a lot of it has been around product recommendations,” said Barrowman, who was a Netscape employee in its heyday and earned his retail chops at the Gap. “I think one of the themes of this is going to be so much more than product recommendations and search optimization. We’re going to see machine learning impact retail from end-to-end and across channels.”

He pointed to no less an ecommerce expert than Jeff Bezos, who more than a year ago talked to Amazon shareholders about the importance of not limiting machine learning innovation to marketing and merchandising and instead setting it loose on back-end operations.

“At Amazon, we’ve been engaged in the practical application of machine learning for many years now. Some of this work is highly visible: our autonomous Prime Air delivery drones; the Amazon Go convenience store that uses machine vision to eliminate checkout lines; and Alexa, our cloud-based AI assistant,” Bezos wrote in his 2016 letter to shareholders.

“But much of what we do with machine learning happens beneath the surface. Machine learning drives our algorithms for demand forecasting, product search ranking, product and deals recommendations, merchandising placements, fraud detection, translations, and much more. Though less visible, much of the impact of machine learning will be of this type – quietly but meaningfully improving core operations.”

Barrowman’s vision includes expanded machine learning technology on commerce sites and in brick and mortar stores. But it also counts on a number of machine-learning driven initiatives to improve back-end operations — including work on reducing returns and improving operations in customer call centers to benefit retailers and their customers. He’s hardly talking science fiction and some of what he talked about is being done, just not to the extent that it soon will be.

When it comes to returns, Barrowman said, data can be used to predict when an item is likely to be returned as soon as it’s ordered. With that knowledge, a merchant could recommend that a shopper go to a physical store to try the items on. Or it could offer a discount for final sales. Or it could send a return-prone shopper to product comparison pages, so he or she could get a better idea of what features and characteristics he or she actually wanted.

Call centers will also continue to deploy machine learning to answer consumers’ questions and direct them to information they desire. The machines will team with humans to increase customer satisfaction and therefore loyalty.

“You can bring natural language processing and chatbots forward to take a load off the call center agents, so they can focus on what requires human-to-human interaction,” Barrowman said.

Or consider the possibility of a smart machine following along in a conversation between a customer and an agent, he said. The machine would be able to point the agent to prompts and information that would help him or her help the retailer’s customer.

Yes, the phrase “customer experience” is used enough to elicit eye rolls from time to time. But the concept comes up repeatedly for a reason. Keeping customers happy is the key to customer loyalty and lifetime value.

And after three days at, it’s hard not to conclude that embracing the right machine learning is a big step down the path to making that customer happiness happen.

Photo by Mike Cassidy

Mike Cassidy is Signifyd’s lead storyteller. Contact him at [email protected]; follow him on Twitter at @mikecassidy.

Mike Cassidy

Mike Cassidy

Mike is the head of storytelling at Signifyd. A former journalist and a retail geek, he covers ecommerce and the way technology is transforming digital commerce. Contact him at [email protected].