Amazon bought Whole Foods in 2017, and the changes to the store since then have been fairly modest. By contrast, a new Amazon Fresh store just opened in D.C., and it really looks to me like the future of grocery shopping.
But of course, the Fresh product is a lot easier to launch since it can piggyback on the Whole Foods supply chain. And the existing Whole Foods retail footprint means that if Fresh is really working out, then Amazon can retrofit Whole Foods markets as Amazon Fresh stores.
The whole thing could turn out to be a huge bust, but I don’t think it will — especially if policymakers keep us on track for full employment and rising wages. And the benefits of the model could be pretty exciting: higher productivity throughout the retail sector, plus more convenient shopping for millions of people. This, in turn, is an object lesson in the perils of dogmatic anti-bigness — as opposed to pro-competition — as a worldview and a philosophy.
How it works
Before walking into the store you fire up the Amazon app on your phone and click on a button there that gives you a QR code. Then you scan the code at the entrance to the store as you head in.
Once you’re in the store, you can get a cart and grab paper bags or reusable bags (or bring your own reusable bag) and then browse the aisles like any normal supermarket. You grab stuff off the shelves and put it into your bags. Then when you have everything you want to buy, you just walk out of the store.
It works because there’s a huge array of cameras in the store (most people don’t look at the ceiling while grocery shopping, but you can see them up there) monitoring you and seeing what you take off the shelves. The All-Seeing Eye of Bezos then simply bills you for whatever you took when you head out the door and emails you a receipt.
The experience is a delight, and a friend reports that he even got a free watermelon for his trouble due to some kind of bug.
Advancing productivity in grocery sales
To know whether or not this is really working as a business, you would need to know a ton of stuff that Amazon doesn’t disclose. How many free watermelons are they accidentally giving away? How much do the cameras and other electronics they’re installing cost? What kind of rent are they paying for the stores?
But the basic business case for the Amazon Fresh store looks pretty obvious:
You reduce labor costs by not employing people to check out and bag the groceries.
It’s more convenient for customers than traditional check-out and dramatically more convenient than the self-checkout methods stores have been pushing people to.
Perhaps a little more subtly, you’re not taking up all this space with check-out facilities, so you can have a reasonably full-featured grocery store in a smaller space.
This last part is intriguing to me because the Amazon Fresh store is more compact than would be suggested by strictly eliminating the square footage of the checkout apparatus. The Amazon hypothesis seems to be that the actual optimal size of the store has changed with their technology.
And you can see why; the check-out mechanism is kind of a fixed cost that can serve a store of arbitrary size. So the check-out mechanism rewards the creation of very large supermarkets or even hypermarkets like Target and Walmart that combine grocery store functions with other merchandising. That model works very well in places where land is cheap and traffic is uncongested (I am writing to you from Kerrville, Texas, for example) but in more expensive urban and suburban destinations, it faces some limits.
With Amazon Fresh, there’s no fixed checkout infrastructure. But the more square feet you add to the store, the more ceiling cameras you need. So it makes sense to operate a larger number of smaller stores in a given area — especially if your company features a world-beating delivery and logistics operation. The smaller footprint lets you become the most convenient option for a larger number of customers.
Productivity is good
One of the many, many, many harmful aspects of the long, sluggish recovery from the Great Recession is that having the country underemployed for a decade created a kind of Luddite mentality where productivity improvements are bad. “Oh no, they are getting rid of the checkout workers!”
Today we’re seeing the emergence of a high-pressure economy that doesn’t have those properties. Instead, employers are reporting difficulty recruiting workers.
“Just pay them more!” goes the shout on Twitter, and in some circumstances, we are talking about high-margin companies that can certainly afford to pay more. But lots of companies deal with very slim profit margins, and if they’re forced to raise pay they will have to raise prices commensurately. There’s also just stuff nobody is going to pay for. It is nice to have someone come in and clean your hotel room. But if the Hampton Inn can’t hire people to clean rooms cheaply, they’re just not going to hire anyone to do it. It’s not a high-end luxury brand, and the customers don’t care that much about it.
To go from high pressure to sustainable high wages, you need productivity improvements that let certain tasks be done by smaller numbers of workers. That productivity doesn’t generate automatic wage gains by magic, but it does create the circumstances in which a high-productivity employer can take advantage of the tight labor market to offer higher pay and drive competitors out of business.
And that’s really what Amazon is threatening here. A model that will let them squeeze grocery stores into locations that other companies don’t see as viable for supermarkets and then hop in with a customer-friendly high-productivity model. And it’s hard to see how many other companies could have pulled this off.
Amazon’s flywheel of everything
It’s very hard for me to imagine a startup doing what Amazon is trying to accomplish with Fresh.
For starters, it’s the prior acquisition of Whole Foods that ensures Amazon has access to a full-scale supply chain for meat and produce even while operating their new stores at a relatively modest scale. But Amazon also has all these built-in customer relationships. “Just walk out of the store and trust us to charge you an appropriate amount of money” is an amazing feature, but it’s also a big ask to make of customers. Why should you trust them?
Well the reason you trust them is that a huge share of the population already shops with Amazon routinely, and we have mostly positive experiences with it. I can’t recall being charged for something in error by Amazon, and they are excellent at dealing with lost or stolen merchandise. Then on top of that, the whole thing relies on a lot of fancy computer stuff (AI? Machine learning? I will just call it “fancy computer stuff”) of the sort Amazon already built in order to power their Alexa devices.
But then the stores, once they exist, become a vehicle for boosting more Amazon businesses. They are locations for Amazon Locker, for example. But also yet another reason to get an Amazon credit card.
Oftentimes sprawling conglomerates lose focus, let things slide, and get disrupted hungrier, more nimble companies with narrower concerns. But at the moment, Amazon is really making bigness work for it. By having established beachheads in so many markets, they are able to execute double envelopment of competitors in new areas. It can feel almost unfair.
The antitrust “paradox”
Noah Smith wrote a post recently arguing that a group of economists from outside the traditional industrial organization subfield has been leading the charge for a rethink of American competition policy.
I agree that the stuff Smith says is happening is, in fact, happening. But there is also a different thing happening, namely a political movement in favor of a neo-Brandeisian approach to anti-trust policy that would focus more on bigness per se and corporate power. David Dayen, who’s now executive editor of the American Prospect, was an early mover in this space, and Barry Lynn, Sarah Miller, and Matt Stoller at the American Economic Liberties Project are its biggest institutional voices. Pierre Omidyar has put money into this movement. And Lina Khan, an early intellectual leader, has been elevated to chair of the Federal Trade Commission and is now a really powerful person.
And these people are saying something different from the economists. The revisionist economists, I think, are basically making two arguments:
For either technical or political reasons, merger enforcement has gotten too lax.
Consumer welfare analysis has tended to overlook non-price harms of lack of competition, such as labor market monopsony.
The neo-Brandeisians are saying something totally different. Khan’s seminal article is titled “Amazon’s Antitrust Paradox,” and the paradox is basically that Amazon is this giant super-powerful company that neo-Brandeisians would like to take action against, but there isn’t anything remotely resembling a case against Amazon that you could win in court. Amazon is a really big company, and Amazon does shady stuff sometimes (there have been a lot of labor rights exposés, they sometimes rip off products that other people sell on their platform, they drove Diapers.com out of business with a price war), but they operate in highly competitive markets. And what’s more, they are constantly entering new markets and increasing competition.
Because the competitiveness of the marketplace is just a different thing than the bigness or power of the corporations operating in it.
The perils of anti-bigness
To the neo-Brandeisians, I’m sure the threat of Amazon stepping into groceries and gobbling up a large share of the supermarket industry is terrifying.
And beyond the world of abstract ideas, a concrete issue here is that a lot of America’s older grocery chains are unionized. If Amazon drives them out of business, that’s bad news for the UFCW and bad news for the labor movement as a whole. This is one of the worst aspects of America’s broken labor relations paradigm — it is so hard to organize new unions at new companies that essentially any shift in the business landscape is bad for labor, which is not a healthy way to run an economy.
To me, though, Amazon Fresh is an object lesson in the perils of anti-bigness as an approach to economic policy. It’s important to have technological progress, rising productivity, and higher living standards. And not everything in life can be done by small groups of hackers working out of a garage somewhere. We need changes that involve interplay with the physical world.
Another area where, in part because of the pandemic, we are seeing technology advance is hotel check-in. The future is clearly a situation where you use your phone to provide the hotel with a credit card and to verify your arrival, and then an app on your phone serves as the key to your room and pass for the elevator. That’s more convenient for customers and also labor-saving for the hotel. But making a system like this work seamlessly is something that is much easier for management groups that operate at a big scale than for a fragmented ecosystem of independent operators. And as this advance rolls across the country, America is going to benefit from having a relatively consolidated hotel sector where Hilton, Marriott, etc. control the bulk of the rooms.
By the same token, it’s because Amazon is big — because it has an e-commerce platform and a cloud computing platform and a grocery chain and more — that it can take on a project like semi-automated supermarkets. A world without bigness is going to be a world where it’s much harder to make certain kinds of progress.