Amazon Go and Autonomous Stores: Don’t think of them as a technology story

Lots of comment about the Amazon Go store.  And the ten variants of ever-more autonomic stores in China.

Four thoughts for discussion and debate:

  1. From a shopper’s view: it’s about time.

As I visit the stores (and read reports from the front), I’m reminded of the essential 2006 article (“Retail Store Execution: An Empirical Study,” Fisher, Krishnan, and Netessine, Wharton, December 2006) that identified the four most important determinants of store performance.

Stock position (specifically, in-stock); the speed, accuracy, and security of transactions; the ready availability of human help; and, the ability of that human help to truly help—to solve problems and create solutions.

From an operator’s point of view, it’s a neat and rigorously-researched four-point statement of priority.

But from a shopper’s point of view, it’s a neat listing of what shoppers have preferred ever since the first merchant’s rug was spread in the first bazaar.

Which is what Amazon Go and the Chinese stores are working to deliver.

2. From an industry perspective: thank you. (Gulp.)

When I reflect upon the work of Amazon and its Chinese brethren (Alibaba,, and Suning, specifically) what impresses me most is not the technology but the damn-the-torpedoes mentality that is determined to break through.

Take POS.  We’ve known for years that the in-store transaction process is one of the most painful parts of the decision journey. Yes, we’ve envisioned a world of RFID-tagged SKUs in self-reading carts or rolling through reader-rich tunnels or gates.  And there has been no shortage of efforts to improve accuracy, speed, and security within the existing POS paradigm.

But we’ve rarely thought about blowing the whole thing up.   And designing not point of sale—with emphasis upon “point”—but immediate and easy and secure payment.

Congratulations to those lighting the explosives. You’re forcing us all to improve.

There are three big differences between those tests of yesteryear and the pilot stores of today. The first is the technology. Of course. Today’s is more powerful, ever faster, ever smaller, ever more connected, enabling faster learning and deeper inference.

The second is that this is about the best practices of the pure-play dot-coms – and not incremental improvements on the store of your childhood. Pure-play best practices of convenience, speed, and ease. Pure-play best practices now being hammered, store-by-store, into brick and mortar reality.

The third is that these are not pilots built to prove technology’s value.  Hardly.  Understand them as first deployments. Soon to be followed by second, third, fourth, and thousandth deployments.

Each one better than its predecessor. Each one a new competitive hurdle.

3. From a value proposition perspective: these stores are more proof that retail’s most valuable SKU will be ease and speed.

Today—admittedly—it’s a most valuable SKU for a certain demographic. A demographic that’s younger. More digital. More affluent.

Yes, it’s spot-on for a certain demographic. Which is probably the demographic you want.

4. I wonder if there are five critical components to these new stores.

The first, the means of instant and secure individual identification. QR codes to registered smartphones, facial recognition, fingerprints.

Second, the connection of the secure individual identification to a payment-services affinity platform. Like Prime. Loaded with payment (account/card) details, shipment addresses. Communicating recommendations, reminders, suggestions, and the occasional offer.

The value equation: 1+2 = on your second step inside, you’re ready not just to shop, but to buy.

Third, the right-time, event-driven autonomic identification and management (from stock to shelf to basket) of individual SKU’s.  With apparel, you’ll use RFID.  With FMCG, you’ll want to use video and image recognition – but we have a long way to go before truly scalable deployment.

The value equation: 1+2+3 = the shopper knows what she has. The retailer knows what she has. The level of accuracy doesn’t change when she changes her mind.

Fourth, the re-imagination of associate value and business processes. Yes, there’s a role for associates. Re-stocking shelves and four-ways, of course. Pick-pack-ship. But with other labor-heavy processes (transaction) now automated, there’s also time for true, value-adding shopper service. The help—in Fisher, Krishnan, and Netessine’s vision—actually helping.

The value equation: 1+2+3+4 = higher in-stock and better shopper interaction.

Fifth, the autonomic and secure check-out. Without significant shrink. A combination of one, two, and three: instant personal identification, the instant (and secure) link to the payment/affinity program, and autonomic item identification.

The value equation: 1+2+3+4+5 = in, helped, out, happy.

Say hello to tomorrow.

Say hello to the new levels of expectations for the store.

Tell me what you think.


Published on Categories RetailTags , , , ,
Jon Stine

About Jon Stine

Global Director Retail Sales at Intel. Jon Stine leads Intel’s global sales and strategy for the retail, hospitality, and consumer goods industry sectors. His CV includes leadership of North American retail consulting practice for Cisco Systems, and a prior stint at Intel, where he founded the company’s sales and marketing focus on the retail industry. His perspective on technology’s value in the industry has been shaped by advisory and project engagements in the United States, across the European Union, and in India, Australia, and the People’s Republic of China, and from 15 years of executive sales and marketing experience in the U.S. apparel industry, working with the nation’s leading department and specialty stores. At Intel, his current areas of research and engagement include the future of the store in this new digital age; how and where retailers turn data into competitive advantage; the role of technology within the new cross-channel shopper journey, and, the critical business and IT capabilities that industry success will demand going forward.