Photo: RetailWire
Has Amazon figured out how to scale its Go cashier-free tech to bigger stores?
Amazon is known for doing things in a big way. So, it should come as no surprise that the e-tailing giant is reported to be working on a way to put deploy the technology behind its Amazon Go convenience stores in much larger store environments.
The Wall Street Journal reports that Amazon is working at a location in Seattle to test how it needs to adjust the technology based on a bigger footprint. Higher ceilings and more items to track are two of the challenges Amazon is looking to address in the test, which is set up to look like a big box store.
In the current seven Amazon Go stores, customers with the Go app are tracked by a variety of technologies including “computer vision, deep learning algorithms and sensor fusion” as they enter and move around the location. Items are automatically added to a virtual shopping cart when a customer takes them off the shelf. When all done, customers simply walk out with their products and Amazon bills their accounts.
The most obvious choice for future deployment appears to be Amazon’s own Whole Foods chain, but a physical store destination for the technology is not yet clear. Neither Amazon nor Whole Foods would comment on the report to the Journal.
The high cost of operating stores with the technology has also been raised as concern for the small Amazon Go locations, so it’s far from clear how that will factor into the tech’s viability in larger environments.
Another concern if the tech were to be deployed in Whole Foods is how it would affect the customer experience in the store.
“They need to be careful not to break what has made that business successful in the first place,” Dennis Keim, a Whole Foods customer told the Journal.
- Amazon Tests Its Cashierless Technology for Bigger Stores – The Wall Street Journal
- Amazon Go still plans to transform convenience in retail stores – RetailWire
- Will the tech behind Amazon Go redefine convenience at retail? – RetailWire
- What would 3,000 AmazonGo stores do to the U.S. retail landscape? – RetailWire
Discussion Questions
DISCUSSION QUESTIONS: What do you see as the biggest challenges Amazon faces as it attempts to port its Go technology to larger store environments? Do you see the technology as a fit for Whole Foods or do you think Amazon will seek to deploy it in some other big box concept?
Poll
BrainTrust
Mark Ryski
Founder, CEO & Author, HeadCount Corporation
Ken Lonyai
Consultant, Strategist, Tech Innovator, UX Evangelist
To me, the biggest challenge remains the technology investment required to monitor larger store environments. By some accounts, the small Amazon Go stores required hundreds of cameras to track shoppers – this could be thousands of cameras in a Whole Foods. I’m just not convinced that this is going to be practical for the foreseeable future given current technology. That said, I’m not surprised that Amazon is pursuing this. Testing is testing – there’s plenty of merit to experimentation. Rolling this out at scale is an entirely different proposition.
Reporters report a lot of things and often they have little REAL knowledge of what they’re talking about. Amazon may indeed be making the system work. I’m skeptical though. I’ve worked with the required technologies for years. There are inherent issues like shrinkage that Amazon is not going to share with reporters. And just like last week’s discussion about cashless stores, app only membership shopping (that’s what Go is) at this stage of society, is a failure waiting to happen.
They may roll it out, it may work good enough that it becomes a distraction for competitors sending them in a flurry to compete. It may be news cycle spin. I may be wrong. We’ll see.
Amazon Go is based on a lot of hardware that is put on the shelves, such as different sensors. This makes the solution less scaleable and more costly than solutions that are based more on computer vision and less on sensors. Additionally, Amazon Go requires the shopper to sign in with a mobile app and be identified, which may be more of a challenge in big box stores as well as in Whole Foods.
Cost, cost, cost. The technology either exists or can be developed. The question is: at what cost does one implement it?
Totally agree. Cost is the biggest challenge and not just for the deployment of the technology. What about the staff that will be required to make sure it is always up and running? That is certainly not some minimum wage worker. For my 2 cents.
Amazon continues to test the concept and push the concept to higher levels. No doubt we are heading into an era that includes a cashless experience that includes the Go technology. The concern for the experience in a store like Whole Foods is a non-issue. There will still be support at the meat counter, the seafood counter, the bakery, etc. There will be adjustments made in the technology that allow customers to just walk out with their product. I like that Amazon continues to push the envelope here. And I’m betting they will lease or sell this technology to others.
Amazon loves to be the game changer and the attention it brings. Go technology is the future but the future is not here yet, and they need to be careful. So many technologies incorporated today for convenience are far from perfect, and they only lead to customer frustration. Take for example self-checkout. How often when using the technology are there problems with items not scanning, not accepting a coupon or a credit card? The convenience is no longer convenient and has the customer frustrated needing to get assistance. Now let’s look at Go technology. It will be quite some time before this technology is so good and perfected that there will be few to no issues. So Amazon would be taking a significant risk if they attempt to implement this technology in Whole Foods too quickly.
Lastly, because of all the problems with hackers customers are becoming more concerned about using technology that has their personal information and Go technology may be one of them. Amazon has already made numerous changes to Whole Foods that in many ways have not helped their business, so I would caution about rolling this out anytime quick. However when Amazon does, it should make it optional and continue to have registers until such a time as the technology has no issues and customers are embracing the concept.
The real challenge in moving a store like Whole Foods to a cashier-less environment is the loss of shopper engagement. Whole Foods is a high touch retailer and that’s one of the primary reasons it’s been able to charge more for its products (I know they claim quality is the main reason but come on, twice as much for organic Gala apples…). Lose that and shoppers will really wonder if it’s worth the effort.
Having worked with video analytics in retail environments for years, RFID, and other biometric technologies, there is a long list of challenges that affect the scalability of the technology in Amazon’s Go store. I went out of my way to personally experience the Amazon Go store during a recent trip to Chicago. I counted no fewer than five Amazon employees (orange shirts) working a small store (less than 1,000 square feet). The Amazon Go app membership shopping experience is sterile — in order for the technology to work the products have to be very accurately displayed (it reminded me of a Swiss Migro grocery truck in the mountains). I simply don’t see a broad deployment in a large store at this time.
In all the years participating on RetailWire, a significant number of the issues covered always seem to converge on supply chain visibility and out-of-stocks. Interestingly enough, I counted no less than 20 items where the employees placed cards that read “so good it’s gone!” With all the technology, they still have out-of-stocks!
In yet another ironic twist, when I left the store there was a nondescript white van parked on the street not 20 yards from the Amazon Go store. There were at least 30 people lined up getting their online lunch deliveries through the passenger window. Just saying!
My spacious local Best Buy is about to become an Amazon Go store, so Amazon thinks it can scale. I’ll guess that Amazon’s plan is to create a new expectation of speed and lack of friction, and will double down on the technology to make its rivals irrelevant. It’s a big gamble…but Amazon has the money to pull it off.
The challenges have been noted in the article. However, I caution voting against Amazon. It has shown the ability to test and innovate often beyond what is imaginable by current retailers. Amazon Go’s retail productivity of at least $2700 per square foot selling area and 50 inventory turns/year, is unmatched in retailing. Consider that Tiffany averages $2900 per square foot of selling area, to get some perspective on the success of Amazon Go and the potential of exporting the Go technology to larger store environments. I predict Amazon will introduce the technology in a limited number of Whole Foods stores as a test, before rolling the concept to more Whole Foods stores and to other potential big box concepts.
New technology is costly. But, you can’t get to scale and efficient costs with technology if you let the going-in costs stop you.
Whether it is Amazon’s Go technology, or something else they learn, or something completely different developed by someone else, this type of technology in stores will be inevitable.
Cost is the biggest challenge to this ambitious goal from Amazon. I am very skeptical that Amazon will actually open the 3,000 Go locations they plan to by 2021 and I am just as skeptical of the viability of extending the cashier-less model to larger format stores. The cost of RFID tags which are the main component of Go technology are still high in comparison to the unit price of the items Amazon sells which creates an unsustainable margin challenge.
With so many stores struggling to remain profitable and most of these store underfunded for core technology projects, I can’t imagine that they would get approval to invest in the very expensive technology to make a cashier-less store a reality.
This is a great strategy to position Amazon as an innovation leader and increase their awareness with free advertising (news articles), but it is far from a reality. It will eventually happen, but it will take some time … and a drop in RFID pricing or the advent of some new technology that will broadcast or sense item identity.
The biggest challenges for Amazon in larger stores are that many more cameras are needed (which scale up in three dimensions as the scanned space increases), there will be more customers and employees getting in the way of the cameras, and the challenges of image recognition as the number of SKUs increases. In addition, how will Amazon know if the cameras missed an item that winds up in the customer’s bag or cart? There are many challenges for the technology working in a Whole Foods store. Just think of the deli aisle where meats and cheeses are cut to the customer’s requests and then wrapped. Or the bakery area where the customer places the goods in a bag themselves. Amazon would be better scaling into non-Whole Foods larger stores where the items are prepackaged and more easily scanned by the cameras.
Self-transacting stores are a bit like self-driving cars — both tech platforms have considerable potential, but the journeys will be long and require deep pockets.
There may also be a bit of road-kill along the way.
I’m a big advocate of in-store sensing for a host of reasons other than just-walk-out transactions. Inventory and order optimization, shopper behavior tracking, in-store messaging, shelf-management, and unified commerce top the list.
So the Amazon Go experiments are of keen interest even if they were eventually to abandon the checkout-free idea (which they won’t).
I get impatient, however, with speculation that the Amazon Go technology is destined for Whole Foods in the near future. Just consider: Why would Amazon propose to build 3,000 other store locations to test this technology when it already has 470 existing Whole Foods locations to play in? The answer seems obvious to me. It’s because they don’t want to screw up Whole Foods, a profitable brand and business on its own, by messing with the formula too soon. At least not until it has learned a great deal more about checkout-free shopping.
So I’d wager that the Amazon Go experiments will continue in a variety of formats, while the Big A gathers learnings it hopes will put it far ahead of most competitors. I give it a couple of years.
Along the way, it will continue to perfect and reduce the cost of in-store sensing technologies and the AI-powered cloud analytics that it requires. That will be a very big business on its own, and the trigger for some intense global competition.
Extending Amazon Go tech to large format stores like Whole Foods is highly questionable. It may not even be viable for Go over the short term given the cost. If the goal is to deliver convenience from a check out perspective, then there are other approaches that should be considered for larger format stores where lines can be a problem. For instance, self-scan options as part of the shopping experience with a quick check at the exit could have a bigger impact. Sure there are some adoption issues but regular shoppers who use it could be given higher service and special checkout to reward usage.
It will be TOO expensive. Period. But if Amazon can somehow figure it out for Whole Foods, competing grocers — who may not have the resources — would be in trouble.