This is Dominic Pino filling in for Jim Geraghty.
On the menu today: Food.
Amazon’s Plan to Remake Grocery Stores
Amazon has transformed the way people shop with its extremely fast delivery service and Amazon Prime subscription. The company has the fifth-largest market cap in the world at nearly $2 trillion. It employs 1.5 million people, second only to Walmart among private corporations, and has annual revenue of about $590 billion, also second to Walmart. Only nine countries’ governments have annual revenue greater than Amazon’s.
Fears that Amazon is too powerful are compounded by the company’s continued adeptness at creating new technologies. If it can always stay a step ahead of competitors, its grip on the market for seemingly everything will never go away, Amazon opponents say. The Federal Trade Commission has sued Amazon on antitrust grounds, saying it keeps prices too high to keep out competitors.
This is the opposite of past antitrust arguments against Amazon, which were that it prevented competition by keeping prices too low, a practice called “predatory pricing.” The flip-flop in the arguments made by FTC chairwoman Lina Khan against Amazon should indicate that her opposition is really about finding a way to punish the company, rather than impartially enforcing the law.
Another reason some fear Amazon’s power is its ability to expand into new markets and take control of them. After all, Amazon started as a bookseller and then expanded to being an online marketplace for just about everything. Amazon has tried to expand into brick-and-mortar retail as well.
It has done so through acquisitions. Amazon bought the grocery chain Whole Foods Market in 2017. That acquisition has been reasonably successful for the company.
Amazon has also tried its own grocery stores. The chain is called Amazon Fresh, and unlike Whole Foods, its offerings are not upscale or all-organic, with prices similar to those of normal grocery stores. It has its own generic store brand alongside regular name-brands.
So why would Amazon want to compete directly with most grocery stores? It had a plan to “disrupt” the industry: no checkouts.
The idea was that an Amazon Fresh customer would be an Amazon Prime subscriber who would scan his or her phone on the way into the store, grab groceries, and “just walk out.” The items acquired would be automatically charged to the Amazon Prime account. That would easily allow for discounts for subscribers, not unlike rewards cards at other grocery stores, while avoiding the need for staffing cashiers and providing a unique, hassle-free experience.
“Four years ago, we started to wonder: What would shopping look like if you could walk into a store, grab what you want, and just go?” an Amazon promotional video said in 2016. “What if we could weave the most advanced machine learning, computer vision, and AI into the very fabric of a store, so you never have to wait in line?”
Amazon’s “Just Walk Out” technology was supposed to be a feature of its Amazon Fresh stores. It was also planning to sell the technology to other companies interested in it. Sports venues have installed cashier-less stations using Just Walk Out technology.
But in April, Amazon announced it was removing Just Walk Out from its Amazon Fresh stores. It turns out the technology didn’t really work for grocery stores, and it wasn’t very popular.
Just Walk Out Just Didn’t Work Out
In April, Amazon said it would be removing Just Walk Out from Amazon Fresh stores, and replacing it with “Dash Carts,” which are shopping carts equipped with scanners. The same idea — no waiting in line — prevails, but in a completely different form and without relying on storewide AI to track purchases. The Just Walk Out technology was more successful in smaller settings, such as convenience stores or kiosks, and the technology will continue to be sold to other companies that want it.
Even though Just Walk Out was first conceived years ago, Amazon seems to have made little progress integrating the technology into its full-size grocery stores. It was still reliant on human backup to accurately provide customer receipts, with a team of over 1,000 workers in India aiding the AI system to track customer purchases. Human aid for AI isn’t unusual, as that is how the systems are trained to improve, but Amazon was missing its own targets for automating the technology. In 2022, every 1,000 sales required 700 human reviews; Amazon was hoping for under 50.
The Amazon Fresh stores more generally have not been as successful as Amazon had hoped. CEO Andy Jassy said in February 2023 that the company would not be expanding its Amazon Fresh brand further until it figured out a store format that customers liked. There are around 40 Amazon Fresh stores nationwide.
This isn’t the first time Amazon has fallen short in brick-and-mortar retail. It tried to compete with physical bookstores with its Amazon Books subsidiary. The first store opened in Seattle in 2015. All stores had closed by 2022. Part of the company’s stated reason for those closings at the time was that it wanted to focus more on Amazon Fresh.
Amazon’s innovation on books was the online store with unlimited options. Competing with other physical bookstores, which lack unlimited variety but have mastered a quirky, fun shopping experience, did not comport with its corporate brand. If you want to buy a book from Amazon, you’re going to do that online, because the whole point of buying a book from Amazon is that you don’t have to go to a store.
It was a similar story for Amazon’s attempt at physical clothing stores, Amazon Style. “The Style clothing stores didn’t even make it two years,” reported the Verge in November 2023. Only two locations, in Los Angeles and Columbus, Ohio, ever opened.
The stores were supposed to provide an “all-new in-store shopping experience built to inspire,” according to Amazon. They allowed customers to scan clothes with the Amazon Shopping app on their phones to send them to a fitting room or a counter to be picked up. The fitting rooms included touchscreens that allowed shoppers to continue to select clothes from there. The stores also used machine learning to create customer recommendations based on their in-store shopping.
Amazon’s specialty in clothes shopping is the ability to order online and easily send back items that don’t fit. Again, the point is in not having to go to a store. And it doesn’t really save that much work to send clothes to a fitting room with an app rather than carrying them from the shelf or rack to the fitting room yourself. (You have to go to the fitting room anyway, and clothes aren’t heavy.)
These examples show two things:
1) Amazon is not invincible. The company makes mistakes. It is not inevitable, or even likely, that it will take over all of retail and last forever. It’s not even close to doing so. Amazon accounted for 4 percent of U.S. retail sales in 2023.
2) Failure is a major part of innovation. Learning what doesn’t work can be just as valuable as learning what does. One way that Amazon has been able to succeed is by taking risks, evaluating the results, and cutting its losses when the results aren’t good.
Amazon isn’t run by geniuses who never screw up. It is run by geniuses who do, and who understand when something isn’t working. They measure that by whether customers are satisfied. Customers didn’t think Just Walk Out or other Amazon physical stores were all that great, so Amazon has either stopped or recalibrated its efforts in those areas. For all its market power, Amazon remains deathly afraid of customers, who still decide whether it succeeds or fails — as they should in a market economy.
ADDENDUM: Speaking of a market economy, it was Adam Smith’s 301st birthday yesterday. Check out one of our essays from his tercentenary last year by clicking here.