I’m not usually one for predictions with due dates. I see the trends, where the dominos are falling, but spotting precisely when a trend will happen is harder. This time, though, I’ll go out on a limb because two events this week have combined to make me think that Amazon will buy Starbucks within the next two years.
Event #1: On 2/28, Cecilia Kang’s New York Times article, “Here Comes the Full Amazonification of Whole Foods,” described Amazon’s rollout of its cashierless “Just Walk Out” technology to a new Whole Foods in the Glover Park neighborhood of Washington D.C. (I thought at first it was in Washington State and nearly hopped in my car to drive north from Portland.)
Kang signed into her Amazon account biometrically, using her palm print, then:
For the next 30 minutes, I shopped. I picked up a bag of cauliflower florets, grapefruit sparkling water, a carton of strawberries and a package of organic chicken sausages. Cameras and sensors recorded each of my moves, creating a virtual shopping cart for me in real time. Then I simply walked out, no cashier necessary. Whole Foods—or rather Amazon—would bill my account later.
Just Walk Out is not new: Amazon rolled it out in 2016 with its robot bodega “Amazon Go” stores, and it debuted a fuller version in Amazon Fresh supermarkets last year.
In my experience, Go is more successful than Fresh. Go is like Trader Joe’s: small footprint, few but quality items, great for grab and go. Fresh was challenging. The shopping cart was hard to push around because of all the heavy sensors and the computer underneath, and the items in the store were packaged for the computer’s convenience, not the customer’s. There’s a lot of monitoring work that the customer has to do, looking at the screen on the shopping cart to make sure the items were logged, making sure they disappear if you put things back. It didn’t feel like a time savings: more like I was just shifting the time I would have spent in checkout into the time I spent in the aisles. Kang’s article suggests that the Whole Foods in Glover Park has similar challenges, but I suspect that Amazon will smooth these things out in time.
The most interesting sentence in Kang’s article came towards the end: “Amazon said it didn’t plan to use video and other Whole Foods customer information for advertising or its recommendation engine.”
Really? Amazon always has a double strategy. Always. If it’s not planning to use the data for advertising or recommendations, then what is the behind-the-scenes strategy for Just Walk Out?
The obvious candidate is that it will make Just Walk Out available to other retailers, following its strategy with Amazon Web Services (AWS). Amazon created AWS to serve its own needs but deliberately overbuilt and then started selling its excess capacity to other businesses. AWS is now the most profitable part of Amazon. Just Walk Out as a Service is already in trial, per Kang, at a convenience store in the DFW airport and at Shaq’s Big Chicken restaurant in Seattle. But the immense amount of hardware needed for Just Walk Out makes the profit margins on rolling it out as a service too low for Amazon’s taste.
It has to be the data. Amazon says it’s not planning to use videos of customers shopping at Whole Foods or “other customer information” for advertising or recommendations, but that leaves ample room for other data-driven insights. And what precisely does Amazon mean by “advertising?” I suspect it means that it won’t sell that data to the third party advertisers that make up the customers for Amazon’s $31B ad business. But there’s a difference between advertising and promotion, and I think that Amazon will use that data to determine what to stock at which Whole Foods, what to put on end caps, what to put on sale, and where to test drive new products.
None of this gets us to Amazon buying Starbucks, which leads us to…
Event #2: On March 2, Amazon announced that it would close all 66 of its bookstores, pop-ups, and 4-Star stores, shifting its retail attention to grocery and fashion. (It did not mention its planned department stores, announced last summer.) It’s not a surprise that Amazon would close down an underperforming business unit. With its Always Day One philosophy, Amazon still thinks like a nimble startup even though its market cap is (as I write this) just shy of one-and-a-half trillion dollars.
But why did the Amazon Bookstores underperform? I suspect it’s because Amazon is too focused on digital efficiency to have embraced the high-friction nature of a great analog bookstore experience. There are no tables and chairs at the Amazon Bookstore a few miles from my home. There is no invitation to stick around for a while, thumb through books, take your time. The geography of the store is planned for traffic—how a person will navigate the store and get to the cash register—rather than around cramming as much inventory onto the shelves as possible. All the books are face out rather than spine out, making them easier to see, but there aren’t as many as in a conventional bookshop. In other words, the Amazon Bookstore is a space but not a place. Great bookstores are third places where people gather, linger, chat, and ponder. Amazon doesn’t want to be a great bookstore: it wants to sell books.
Amazon will sell more books by breaking its Amazon Bookstores into component parts and distributing them throughout Whole Foods. With its immense data engine, Amazon will know what books are likely to sell in a particular Whole Foods in a particular neighborhood. Already a few end caps in Whole Foods sell clothing and toys. It would be easy for Amazon to have a few bookshelves throughout a Whole Foods containing a mix of best sellers and algorithmically curated titles that silently cater to the interests of its regular customers.
Let’s say there’s an Amazon Prime subscriber who shops at Whole Foods and who over the last year has read three historical narratives about famous rowing teams (like The Boys in the Boat). If Amazon has bookshelves in that subscriber’s Whole Foods, and a new rowing memoir has just come out, then it could ship a copy to that market for the customer to stumble upon while shopping. Since Amazon also knows what items the customer buys regularly, it could put that new rowing book on a shelf it knows the customer is likely to walk past.
The customer would experience this as a delightful moment of serendipity even though it’s really a big win for AI. Amazon can do this at scale for all its regular customers at all of its Whole Foods locations, letting algorithms pick likely-to-sell titles and organize the fulfillment behind the scenes.
The challenge is that books are higher consideration items than groceries, and book shoppers often need a few minutes and a chair in order to make a purchase decision.
That’s where Starbucks comes in. Starbucks is the master of dwell time. It has created third places all over the world where people can plug in their laptops and tap away. Amazon and Starbucks are already collaborating on hybrid Starbucks/Amazon Go locations in New York City, so the companies already know that they can learn from each other. Acquisition is the next step.
Amazon knew that Prime subscribers and Whole Foods shoppers overlapped when it bought Whole Foods in 2017. A similar overlap exists with Starbucks. A Starbucks inside a Whole Foods would be at the back of the store near some of the bookshelves and the Amazon pickup/return desk.
The reverse is also true: if Amazon owns Starbucks then every Starbucks becomes a tiny bookstore with a slender selection of bestsellers and algorithmically curated titles silently intended for that location’s regular customers. Fusing Starbucks customer data and Amazon customer data would give the company even more insight into the customers’ needs, wants, and habits. Amazon would even replace Google as the Starbucks wifi partner in order to get access to that data (where people are surfing) as well as to promo other products available right there at Starbucks. Every Starbucks would also become a new point in Amazon’s immense distribution network, helping it to achieve faster delivery in urban areas.
As I write this, Starbucks market cap is $104 billion, which means it’s roughly one fifteenth the size of Amazon. Buying Starbucks will be a huge acquisition for Amazon, much more ambitions than its Whole Food acquisition, but doing so will add deep dwell time expertise that will amplify the value of both its grocery businesses, its planned department stores, and its new hi-tech Amazon Style clothing stores.
I’m sure that Amazon’s Alexa digital assistant will fit in here too. I just haven’t figured out how yet.
Leave a Reply