Imagine visiting the grocery store, grabbing whatever you want and just walking out, completely eschewing the checkout process and the associated hassles of waiting in line and paying for items. It may sound like shoplifting, but it’s actually Amazon’s Amazon Go grocery store, and it uses machine-learning techniques to automatically track and bill consumer purchases and eliminate the traditional checkout process.
While currently in the experimental stage, Amazon Go may soon be coming to the real world, with news that the online retail firm has filed trademark applications related to the store in the United Kingdom. The trademarks relate to Amazon Go’s slogan, “No Queue, no checkout. (No, seriously.),” according to a report in Newsweek.
The company currently operates a pilot Amazon Go store for its employees in Seattle. Before entering the store, the employees tap a button on the Amazon Go app and then scan their smartphone on a device located near the entrance.
The employees then select products from the store shelves, with their activity monitored by a system that watches their activities. Amazon says the system bears similarities to a self-driving car, combining computer vision, sensor fusion and deep learning. These machine-learning technologies are sophisticated enough to understand complex behavior, such as when a customer takes an item—and then places it back on the shelf.
When the employees leave the store, their Amazon account is automatically billed for the purchase.
This automated approach to brick-and-mortar grocery sales may have huge revenue potential.
The global grocery store market totals $800 billion, according to analyst John Blackledge of Cowen & Co. With this massive opportunity, Blackledge predicted Amazon Go could become Amazon’s biggest revenue source in the future. This would represent remarkable sales growth, given that Amazon posted revenue of $136 billion in 2016.
Amazon’s ambition to expand its presence in physical retail has been discussed in the past by company CEO Jeff Bezos. The company in 2016 said it planned to open hundreds of new bookstores.
“We’re definitely going to open additional stores. How many we don’t know yet,” Bezos said during an Amazon shareholders meeting. “In these early days, it’s all about learning, rather than trying to earn a lot of revenue.”
The move into brick-and-mortar stores makes sense for Amazon. The company is the dominant online retailer in the United States, accounting for 43 percent of sales in 2016, according to Slice Research. However, 90 percent of worldwide retail spending still is generated by physical stores, representing a massive growth opportunity.
Using machine-learning technology originally developed for the online segment, Amazon may be on the verge of revolutionizing the traditional retail business.
Tyler Schulze is vice president, strategy & development at Veritone. He serves as general manager for developer partnerships, cognitive engine ecosystem, and media ingestion for the Veritone platform. Learn more about our platform and join the Veritone developer ecosystem today.