Photo: Amazon
Will Smart Carts Pay Off Better Than ‘Just Walk Out’ for Amazon?
Amazon plans to remove its “Just Walk Out” technology — which allows customers to shop and leave the store without going to a register — from Amazon Fresh stores, replacing it with its Dash Cart smart shopping cart device.
The change, first reported by The Information, only affects 27 of Amazon Fresh’s 44 U.S. locations that use the technology. The remaining stores already use Dash Cart. Amazon’s Just Walk Out technology will still be used by the Amazon Go mini C-store concept, the smaller-format Amazon Fresh stores in the U.K., as well as the more than 130 third-party retailers that have partnered with Amazon on the technology, including several ballparks, airports, college stores, and cafes.
Amazon told media outlets that the decision to remove the technology at its Amazon Fresh U.S. stores was due to customer feedback.
“We’ve invested a lot of time redesigning a number of our Amazon Fresh stores over the last year, offering a better overall shopping experience with more value, convenience, and selection — and so far we’ve seen positive results, with higher customer shopping satisfaction scores and increased purchasing,” Amazon said in a statement. “We’ve also heard from customers that while they enjoyed the benefit of skipping the checkout line with Just Walk Out, they also wanted the ability to easily find nearby products and deals, view their receipt as they shop, and know how much money they saved while shopping throughout the store.”
The change comes after Amazon CEO Andy Jassy in February 2023 said Amazon Fresh openings would be paused “until we have that equation with differentiation and economic value that we like.”
First announced in July 2020 and majorly upgraded in 2022, Dash Cart operates on a technology platform similar to Just Walk Out. Shoppers using the smart shopping cart scan the QR code on their Amazon app to begin browsing the store. A combination of computer vision algorithms and sensor fusion identifies items as they’re placed in the cart. When done, customers are automatically charged and avoid the cashier as they walk out.
One challenge to Just Walk Out technology, first introduced in 2016, was that it often took hours for customers to receive receipts after leaving the store, largely because a team of more than 1,000 workers in India was rewatching and labeling videos to ensure accurate checkouts, according to The Information’s report.
Amazon told Gizmodo that the primary role of the outsourced workers was to annotate video images and train the AI, and they validated only “a small minority” of shopper receipts, although sources told The Information that about 70% of receipts required human reviews.
Critics of Just Walk Out technology in the past have questioned the costs and structural challenges related to placing video cameras and other hardware in store ceilings. Privacy advocates have also raised concerns about the Just Walk Out technology’s access to customers’ personal data, including their in-store behavior.
Dash Cart’s appeal also remains debatable. A Business Insider article from 2022 found that although customers like Dash Cart’s running tally of the items they pick up, usage had been well below Amazon’s expectations due to the cart being unsuitable for large items because of its size, prohibited from rolling to the shopper’s car, and having scanning issues.
After removing the Just Walk Out technology, Amazon Fresh shoppers will still have the option to check out with a cashier or use self-checkout lanes.
Discussion Questions
Do you see Amazon’s removal of “Just Walk Out” from Amazon Fresh stores as a blow or a temporary hurdle to the technology’s expansion potential?
Will Dash Cart offer significantly more appeal to in-store shoppers?
Poll
Loading …BrainTrust
Mark Ryski
Founder, CEO & Author, HeadCount Corporation
Cathy Hotka
Principal, Cathy Hotka & Associates
By removing Just Walk Out technology from it’s own stores, Amazon is sending the market a clear message: The solution is not currently scalable or sustainable. When Amazon first introduced Just Walk Out, there were many skeptics (including myself), about the cost and technical complexity this solution required. The disclosure that Amazon has a 1,000 human workers in India manually reviewing video is also telling. I do believe that this may be the end of Just Walk Out aspirations for Amazon, but it’s not necessarily the end of the road for other technology companies trying to develop a cost-effective, Just Walkout-like solutions. However, the lesson here is clear: if the mighty Amazon couldn’t make this work after all the time and money they invested, then it’s going to be challenging for anyone to solve. The Dash Cart pivot by Amazon is nothing more than plan B.
Agree! This was a technology with its challenges. It may evolve into something more implementable or it may fade away and assigned to R&D under the “it didn’t work” column. I have questions about the cart technology replacing its predecessor at Amazon Fresh, albeit fewer reservations than with the camera-based technology. Let us watch what this next Amazon experiment yields. The silver lining in the recent cloud is that the retail industry has at least one pioneer company with DEEP pockets and it may well influence the industry as it has in other areas.
Excellent points- 1000 people watching video feeds for transactions from the 40ish existing Fresh stores. That ratio indicates a person is scrutinizing every transaction. So, they haven’t eliminated cashiers they’ve just found a way to outsource them.
On today’s discussion thread at Ars Technica, someone calculated that with the video review, the actual labor intensity at the JWO airport shops is 9X of what a traditional human-staffed shop needs. Plus a human can perform stocking and cleaning duties and make marketing posts. For anything bigger than a vending machine, just hire humans!
I do not see this as a big blow to Amazon. It is part and parcel or experimenting, testing and adjusting – something that Amazon excels in. However, the truth is that Just Walk Out technology was never a silver bullet for Amazon’s grocery ambitions. For a start it is very complex and expensive technology. On top of this, it is not a differentiator that is all that meaningful to consumers. In larger stores the technology was also glitchy, which is why human observers were required to ensure customers were being billed correctly. While Amazon isn’t abandoning the technology entirely (it will still license it out and use it in smaller Go stores), it is moving to a different solution in Fresh stores in the form of smart carts. These are simpler than Just Walk Out, and they can also drive media/advertising revenue, but they are still an overly complex solution for the simple problem of allowing customers to walk out without checking out. At Tesco in the UK, a simple scan and shop solution has been in place for years: it’s not fancy or flash, but it works!
Good summary, Neil. Amazon for sure doesn’t see his as a big blow. It’s part and parcel in how they innovate. They’ve learned, found a sweet spot, and are now trying to convert/build a long list of small store formats as fast as they can.
As a grocery tech pioneer, Amazon faces a temporary hurdle as it experiments, learns and pivots from Just Walk Out to Dash Carts. Both technologies are shaping consumer habits by offering innovative options for efficient, convenient store experiences.
Dash Carts will appeal to shoppers by making grocery store visits more engaging, interactive and personalized.
Consumers gain real-time product discovery and details to make informed decisions in stores. Dash Carts could evolve with relevant ads, recipe suggestions, recommended products, promotions and allergen alerts to help customers as they shop. Shoppers also save time by skipping the checkout line.
Those in the anti JWO camp – who saw it as a splendid example of Theory Triumphant over Reality – will likely see some vindication; and it could well be. But really only Amazon has access to the numbers on how much money it made – or lost – and what kind of shrink issues emerged. Until those mystical digits show up somewhere, or until – really, unless – Amazon does another 180 and brings it back, we’ll all we speculating.
To be clear, there is nothing wrong with Amazon’s Just Walk Out technology; the friction here is consumer behavior is difficult to change, especially in shopping.
While the notion of just walk out sounds good, the reality is, it feels weird vs what consumers are used to; no check-out, no receipt, sometimes no bag or even human in sight. Shopping is an engaging experience, checkout (even self check-out) is part of the experience. Those 2-3 minutes with the cashier or scanning your own items one by one and leaving with receipt in hand (or on your mobile device) brings closure to the shopping experience. When you just walk out, you miss that and it’s weird. Nice. But weird.
What Amazon may find is there is a place for this technology but it’s not likely a standard shopping trip. However, airports and ball parks are not typical shopping trips and may be well suited for Just Walk Out technology.
This leaves the smart shopping cart technology of Dash Cart, another good idea but a sometimes weird shopping experience. The carts are often not agreeable to larger shopping trips or bulky items. So, it’s best appeal is small trips. But, if you still have self check-out and cashiers in Amazon Fresh stores, do you need Dash Carts too?
There is a reason grocery stores are slower to evolve technology than other channels. Consumers have certain basic expectations of a shopping experience and those basic expectations don’t change fast or often.
Someone is going to figure this out. Probably later than sooner. For now, we will watch the instacart “pivot” with interest, along with the assumption it, too, will fail.
I think there are two parts to this. One, did the JWO tech actually work? And it seems like the answer is no, or at least not nearly as well as advertised. Two, is this a process flow that consumers will like and use? We’ve engaged a bit on the RFID-based JWO for Fashion, and my immediate feedback was and remains: no consumer is going to want a screen-less checkout. It’s one thing when it’s $10 and you can do the math in your head, and if you miss the $0.50 promo that should’ve been applied, it’s just not that big of a deal. But when you’re spending $100 or $200 on a full basket of groceries or clothes, then you want to make sure you got every promo and discount, and every price charged was the correct one – before you actually pay. Also, more and more retailers are offering options to pay – redeem loyalty points being a big one. When there’s a lot more at stake, you don’t want to get to the car or get home and then realized you were overcharged not $0.50 but $20 – or that you didn’t get loyalty credit for your $200 purchase. THAT is a bad customer experience. It’s no surprise to move to the cart – consumers can see the price they’re being charged while they’re standing next to the shelf label. And they can see the running total to help manage their budget. That’s what was wrong about JWO when you tried to scale it to more than a few items at a time.
I still don’t get the whole digital cart concept. Why does the shopper need another screen? They have their smart phone so all they really need is a port to connect to the store’s network and an app, which they likely already have in some form. It really is similar to the kiosk craze we saw 25 years ago that went nowhere, with the exception of deli ordering. Why should a retailer spend all that money when digital carts are just another solution looking for a problem to solve?
It is an excellent idea to use your smartphone, and the basket has a port.
The evolution of technology adoption in any industry, retailing included, is bound to be plagued by “fits and starts”, as the consumer learns about the benefits and drawbacks of the solution, and tech leaders refine the experience based on iteration and refinement.
There are innovations taking place in the “smart cart” space that are compelling, both in terms of the rapid checkout enabled by solutions like Amazon’s, as well as the ability to integrate retail media elements, solution selling, and product location within a store layout that are exciting to merchants like me.
To expect these solutions to be perfect at the first iteration is to not understand how innovation works. My favorite Thomas Edison quote, “I have not failed. I’ve just found 10,000 ways that won’t work” reminds us that sometimes it takes more than one try to get a new innovation in any industry, including retailing, just right.
Absolutely agree. It’s an over-engineered solution. Plus those carts are very expensive!
Agreed. This isn’t pulling the plug. These innovations are only a first round version of what we’ll ultimately use more often.
Transformations are complex, even revolutionary ones.
Amazon is reducing its proprietary “Just Walkout” automated checkout footprint from their Amazon Fresh grocery stores. In its place, Amazon will offer its smart shopping carts, called “Amazon Dash Carts,” as a frictionless checkout experience in the stores. The carts allow shoppers to scan items as they go and then skip the checkout line with an automatic payment.
Amazon’s strategies contradict the global trends and the emergence of autonomous stores, where these capabilities extend beyond retail operations. The list includes grocery, fashion, electronics, convenience stores, and QSRs, which are keeping pace with the digital-first customer, seamlessly navigating across digital, physical, and social platforms and quickly shifting from discovery to purchase.
According to Business Insider Intelligence, the number of global stores with autonomous capabilities will increase from only 350 in 2018 to 10,000 stores in 2024, with revenues forecasted to rise significantly from $70 million to over $20 billion.
The Amazon Go concept may work effectively in smaller convenience stores with 10-15 locations. However, a larger-scale grocery operation is highly challenging and dynamic, as it must incorporate all the technologies and infrastructure necessary to make it work.
The Dash Carts are not the answer due to investment costs, maintenance, or potential damages.
The removal of “Just walk out” technology from Amazon Fresh stores will definitely make it more difficult for Amazon the sell the technology to other grocery stores. I suspect the reason for the removal of JWO technology was because the ROI was not positive. Dash Cart is a less expensive option, but it is still a significant expense. Personally, I still think using the customer’s smartphone to scan items while they shop is the most economical way to making checkout more convenient.
The elephant in the room is shrink. How much shrinkage is the retailer willing to accept theft at checkout as a cost of doing business without cashiers? Not that cashiers are the answer to eliminating checkout fraud; too often, they are enablers. More than a few retailers have implemented self-checkout and have seen their shink rates soar. The interesting dichotomy is between those retailers who admit they have a problem and have changed or eliminated self-checkout and those who have decided to blame their soaring theft rates for everything except self-checkout. The reality is somewhere in the middle, as it always is. Although they haven’t cited shrink as an issue in this case, based on my experiences with the tech, it has to be part of the equation.
If you want to know about smart carts, talk to Stop n Shop in Boston. 7 years in pilot
AI that needed 1000 workers reviewing videos was neither Artificial nor Intelligent. Nice try Amazon, but the view behind the curtain here is not exactly flattering.
Look at the photo. The problem with the smart cart is its very small size. I cook for two and regularly bring home 8-10 bags on my weekly shopping trip; this cart holds only two bags. Amazon needs to spend more time talking with customers.
Here is what Amazon knows: By 2030 or 2035, most stores of every type will have some kind of JWO technology. They are just trying to determine what their offering will be. The technology they are trying to resolve has nothing to do with their own stores and everything to do with everyone else’s stores.
Just imagine the business that could be generated for AWS.
Customers are walking away – even from self check out, clearly preferring the low tech/high touch cashier driven approach. Anyone surprised that JWO wasn’t any more successful? 1,000 people offshore watching? After putting very expensive technology into stores to support JWO, in the end there was “mass shopper watching” going on?
Everything new is a test or experiment. Amazon gave the Just Walk Out technology a try. They are smart to pull it and replace it with something comparable. This doesn’t mean we’ll never see a version of Just Walk Out ever again. Something like it will be in our future retail experience.
This change appears to have positive benefits for both customers & Amazon. It appears to have significant shopping advantages in terms of a viewable interactive screen that brings convenience & valuable information to customers (identified savings, receipt viewing, etc.). I am always reminded that Amazon is the king of market innovations & is willing to terminate those that don’t work. We shall see how this latest market experiment goes.
Amazon is only shutting down Just Walk Out technology in grocery stores. Small GO stores, stadiums, airports and licensed stores are apparently working well as the sweet spot for JWO tech expansion. Amazon can now focus on those formats for success.
The Dash Cart upside is customers are naturally scanning items by just placing them in the cart, thus removing steps. This is a quantifiable time saving equation benefiting both customers and operations teams. Heading to the front end checkout, it’s up to employees to confirm scans customers have missed. In the end, less money to deploy Dash Cart in stores with 40,000 items is a good move.
As a key player in this space, I am not ready to call this “a blow”. There is a time and place for every tech. We see “places” like stadiums, universities, hotel lobbies, that are quick to test and in some cases adopt this tech. Perhaps the “time” is not right for the “grocery” sector. But don’t underestimate the knowledge that Amazon walked away with from this endeavor. The insights derived from both “what it takes to run it” and “what people liked/hated about it” is for sure informing what they are doing with cart. Cart is definitely more “Capex light” than the total store, but I don’t see this as THE driver. This technology remains an important part of Amazon’s innovation in retail, with its application continuing to evolve based on consumer demand, competitive dynamics, and tech advancements. Think about it this way (using another effort that was “pulled back upon”)….Did Alexa fail? Or simply give way to a better version of her with GenAI now in play?
Technology is fine except when it is used to forcibly change a shopper’s behavior. Did shoppers ask for Just Walk Out or Dash Cart? Here is an idea to speed up the checkout process: Have the store hire more cashiers.
I see Amazon’s removal of the “Just Walk Out” feature from Amazon Fresh stores as a temporary hurdle rather than a significant setback to its expansion strategy. This decision showcases Amazon’s responsiveness to customer feedback and its dedication to refining the in-store shopping experience.
Amazon’s Dash Cart may offer more appeal to in-store shoppers because it addresses some of the limitations of the “Just Walk Out” technology, such as providing real-time product information and deals. However, its success will hinge on overcoming challenges like cart size and scanning issues.
Overall, Amazon, as an industry leader, remains committed to innovation in cashierless technology. While setbacks are inevitable, they are part of the journey toward enhancing customer satisfaction and improving operational efficiency.