Amazon recently announced that they were scaling back their camera based checkout-free system in their Amazon Fresh and Go stores.
The media had a field day with this news after discovering that Amazon were employing thousands of people in India to manually verify video feeds to track purchases.
This was not new to those in the industry. It has long been an open secret that camera based checkout-free systems like those used by Amazon, Standard, AiFi, and others require an army of humans to track purchases. Why? Because cameras offer up around 60-70 percent reliability in tracking items picked off the shelves.
Amazon’s Decision
The accuracy of camera based checkout-free systems fail more than 30 percent of the time, which means that they require humans in the loop to verify transactions, to ensure the correct basket and receipt.
The technology is getting better, but this significant cost of having manual checks for what is meant to be an automated system coupled with the fact that this means that consumers don’t get their receipt until well after they’ve left the store leads to a challenge for purveyors of this technology. Amazon’s execs have quoted the reason for them to scale back Just Walk Out as consumer feedback demanding a faster receipt.
However, it would be presumptuous to say that checkout-free cannot work. In fact, RFID-based checkout-free systems, better called auto-checkout systems, work incredibly reliably, with extremely high accuracy at 99+ percent.
At Robomart, we have developed the first RFID based auto-checkout system with extremely high accuracy during operations. The small added cost of RFID tags is baked into the margin of the products being sold by retailers and consumers get their receipt within 30 seconds of closing the door, making it a truly seamless experience.
The Benefits of RFID
Compared to Amazon’s Just Walk Out system’s that often makes the customer wait hours for a receipt, 30 seconds makes the experience truly seamless. Our team have been experimenting with triangulation of RFID signals to show consumers their basket as they pull products off the shelves. This level of innovation is made possible through a combination approach, leveraging camera vision alongside RFID to pinpoint the movement of every single pack inside a shop on a real-time basis.
There are still challenges for auto-checkout to work well in large store environments, whether they are RFID- or camera-based simply due to the scale of the shopping environment. However, auto-checkout has proven to be extremely effective for small format stores like Robomarts that are hailed by consumers, giving them the ability to shop right at home in a frictionless way.
Sometimes it takes a complete overhaul to enjoy the fruits of a new technology, rather than using it in existing environments in an attempt to make marginal improvements.
Embracing Frictionless Shopping
Dash Carts are Amazon’s supposed saving grace to the challenges of Just Walk Out. But they are yet another marginal improvement that saves just a few minutes. This is almost like putting out slightly less inferior technology to replace inferior technology. Dash Carts may lead to higher accuracy than Just Walk Out, but they require shoppers to scan items, and are thus not truly auto-checkout.
Additionally, the bugs and errors that will undoubtedly occur in the Dash Cart will not have the benefit of a manual verification so they will be much more prevalent, visible, and unforgivable by consumers when they do occur, much like the oft-error prone self-scan checkout systems that were once heralded as superior tech.
Just Walk Out pioneered and popularized checkout-free shopping. However, Amazon’s decision to scale it back and instead focus on Dash Carts is a mistake in my opinion. For a technology that is meant to be seamless, Amazon’s efforts have often gone the other direction.
It’s time for the industry to embrace RFID based auto-checkout and new autonomous retail models that enable the most frictionless shopping experience possible.
About the Author: Ali Ahmed is co-founder & CEO of Robomart, Inc.