'
Special Feature
Part of a ZDNet Special Feature: Data, AI, IoT: The Future of Business

Is Amazon Go's cashier-less shopping the future of retail?

Shopping is being disrupted by AI-powered automation, but the customer experience still reigns supreme.

In late 2016, e-commerce giant Amazon unveiled the ultimate grocery store for introverts -- a cashier-less store in Seattle where customers simply walk in, take what they need, and walk out, with their account charged automatically. Using artificial intelligence (AI), the Amazon Go store fundamentally changed what it meant to run to the store for some milk and bread.

Special feature

Special report: Data, AI, IoT: The Future of Retail (free PDF)

This ebook, based on the latest ZDNet/TechRepublic special feature, looks at the rise of e-commerce and the digital transformation of retail companies.

Read More

Amazon Go opened to the public in January 2018, promising no cashiers or checkout lines and a streamlined shopping experience. Amazon's system in the store is called Just Walk Out, and it relies on technologies such as computer vision, sensor fusion, and deep learning to determine what has been picked up by a customer before making the appropriate charges to their Amazon account when they leave the store.

SEE: Job description: Ecommerce tech analyst (Tech Pro Research)

There's no denying that Amazon Go is controversial for its substitution of human workers with technology -- much like McDonald's use of automated self-service kiosks to take orders. But the bigger question is whether or not this will last, and spread to other retailers.

So, is Amazon Go's cashier-less shopping the future of retail? The answer is complicated.

How did we get here?

Digital has already altered nearly every aspect of the retail business, said Forrester Research principal analyst George Lawrie, empowering customers and changing their expectations. Amazon has worked hard to create its own differentiated customer experience online, Lawrie said, and now it's extending that to brick-and-mortar retail.

Much like contactless payments (e.g. Apple Pay), the benefits to Amazon Go are simple: Speed and convenience. "The ideal shopping experience should be frictionless," Lawrie said.

And that speed and convenience is important to the customer. In the past year, 86 percent of US consumers said they left a store due to long lines, resulting in a purchase at a different retailer or no purchase at all, according to 451 Research data. "This results in approximately $37.7 billion lost in potential sales," said Jordan McKee, research director at 451 Research.

Additionally, 75 percent of shoppers who are early adopters of digital technologies (451 Research calls them 'spendsetters') said they would shop more in a store with a payment experience like Amazon's Just Walk Out.

Simply put, yes, Amazon Go's model is the future of much of retail, but not all of it. It will work for certain retail segments, but not all. It all depends on the context of the experience.

"In Amazon's case, the brand is super efficient. That reflects what they do, so it makes a lot of sense in the Amazon Go case," said Ray Wang, principal analyst and founder of Constellation Research. "But if you've got a luxury high-touch brand, you want to be able to provide both, because your customers might not be happy paying a premium to be completely automated."

For example, Wang said, sometimes customers want to do a mobile check in at a hotel and not talk to a concierge. But, for a five-star hotel, they may want a personalized greeter and more high-touch benefits.

There's also the issue of cultural differences in certain geographies, Lawrie said. In some areas in Northern Europe, for example, consumers tend to avoid credit cards and prefer to pay cash. "Cash on deliver is the most popular settlement for e-commerce purchases in Sweden and the Netherlands, for example," Lawrie said.

Still, the customers are only one half of the equation.

The business value

On the business side of things, Amazon Go-like automation could help reduce staffing and increase throughput, McKee said. "Additionally, some may realize higher average order values in increased customer loyalty," McKee said.

The employees who do work for the stores could also see increased productivity as a result of being freed up by the automated checkout lines, Lawrie said. But there's much more to it than simple automation.

"If you think this is just about automating and getting a kiosk up and running, you've missed the point," Wang said.

What is the point? Data, and lots of it.

Using AI to simplify the retail process allows for more sophisticated data capture, and for a company to better understand its customers. For example, Wang said, a company can potentially determine where a customer was standing when they decided to make a purchase, how they may have been feeling, what type of weather was happening outside, and even the customer's heart rate. This kind of data can be invaluable in understanding the context around a purchase.

As sensors mature and omnichannel retail expands, this data capture will extend into the home as well, Wang said, with self-replenishments and subscription models learning customer behaviors and mailing them products. As all of this is driven by AI, though, it will depend on how much privacy customers are willing to trade for convenience, value, security, and status, Wang said.

For companies that wish to get started with cashier-less shopping, McKee said it's important to remember that the experience hasn't reached maturity yet.

"Those eager to get started should conduct thorough proof-of-concepts and pilots before commercially deploying," McKee said. "Retailers would also be wise to talk to their customers to understand if and when they will be ready for this type of shopping experience."

Also, CIOs and CISOs should be ready for any new security threats that come with such innovations, Lawrie said.

"The most important element will be to guarantee that customer data is held securely and used only within the strict GDPR permissions that customers grant to brands and retailers," Lawrie said. "New technologies like facial recognition also need powerful networking and computational capability at the edge of the enterprise network."

Once the infrastructure and security tools are in place, and the technology has matured enough to provide a better user experience, it is only a matter of time before the Amazon Go model invades most retail markets. There will be plenty of holdouts, and luxury brands that still offer a personal touch, but for most retail experiences around the world, the future will be cashier-less.

Also see

amazongo.jpg
Image: CNET