Amazon isn’t the only company imagining stores without cashiers
Amazon is trying out a new brick-and-mortar shopping experience. In 2017, it will open a grocery store in Seattle to the public where you don’t have to stand in a checkout line, not even a self-checkout, because you will use a phone app – Amazon Go – to enter the store and then the store tracks what you buy to charge you at the end of the experience. Hints of the project emerged in March 2015, when Amazon filed a patent on updating fulfillment center tracking technology, but the exact details were kept secret until this week.
Supermarket workers, of course, would be in a uproar should this develop further. And it will, as other companies are exploring their own options. While it won’t eliminate counter staff or stockers – automation of those jobs is a whole other story – it would mean there is no need for staffed registers anymore. There would have to be some such presence in the event of an error or other problem, like returns, but this next step in self-checkout technology would be even more disruptive than self-checkout, in part because it will be a lot easier to use than the current set up.
Amazon says the technology works by combining, “computer vision, sensor fusion, and deep learning.” In short, the store maintains a digital library of all the products in it, and when you remove something from its shelf, Amazon Go registers that it is you doing that and adds the item to your virtual cart. If you change your mind, it will register that you’ve put the item back and delete it from the bill. You then leave the store with what you have and pay via phone.
I don’t use self-checkout at my local supermarket chain, because the system still has a very hard time processing anything that isn’t in a container, produce in particular being a hassle. Even bar codes at staffed cash registers can be uncooperative. Self-checkout technology has, when it comes to registers at retail outlets, hit something of a ceiling. This is why the NCR Corporation, a hardware company that makes registers, barcode scanners, and ATMs, has been researching technology similar to Amazon’s, according to NPR, to make them more accessible and appealing to the people who still prefer going to a store to shop.
(Not everyone is inclined to buy groceries online and still wants the in-store experience.)
Amazon Go is better equipped to handle those kinds of items, and would be a real time saver in that regard. The patent application from last year suggests it would be able to deal with a common problem in retail, that of customers putting items they’ve changed their minds about back on the wrong shelf. Many of us, including your author, have surely left a bottle of soap in the cereal aisle or a can of soup near the checkout line’s candy display at the last minute. So how would Amazon Go know we’ve put it back? “Because the items have already been identified, the transition may be done automatically without any affirmative input from the user or delay to the user,” the description reads. This is based off of time-saving technology meant for Amazon fulfillment centers, where processing products quickly is critical.
Presumably, the cashiers at stores adopting such technology, like the kind NCR is looking at as Amazon Go goes through its beta, would be fired or reassigned. Shoppers would still want bag packers and some folks are likely to want to be able to pay in cash, so it is not like all jobs will be lost. Given how inventive shoplifters are, there really is no way the company can dispense with some level of human oversight. Wearing something that could confuse facial recognition software, for example, could have an effect on how the store processes shopper identification and create problems.
However the stores are likely to prefer keeping the bar on security on the low side, preferring that customers have a smoother buying experience that is likely to bring gains across the board, while they accept – as they do today – that they will face some losses to theft. Which option saves them more money, having a few steaks go missing, or paying more salaries?
The Fintech that goes into this initiative will need to be pretty secure as well, to prevent hacking or people gaming the system. The fact that it is in person limits some of the things hackers could do, but the main concern is basic data security. Home Depot, for instance, has been successful in instituting self-checkouts – which I do use there, unlike at my supermarket – but a data breach greatly embarrassed the company and will cost $13-20 million to settle. Amazon has had some account information leaks in recent years, including incidents where it had users reset passwords for undisclosed reasons, but has not yet suffered a major breach.