How checkout-free retail is building scale while many stores remain locked down
Checkout-free stores still feel experimental, but the coronavirus pandemic is pushing more retailers to contemplate how to reduce human interaction in larger settings.
Amazon’s “Dash Cart” isn’t replacing Amazon Go, a convenience store that relies on cameras and sensors to detect which items customers pick up as they go through the store. But the Dash Cart's smaller profile makes the concept applicable to the 30,000-square-foot layout of a traditional supermarket, which would be much more complicated to retrofit with cameras and sensors in every aisle.
The idea of compressing this technology into each shopping cart may be the solution that brings the innovation to traditional stores, at least for larger layouts.
“Once we go to supermarkets at around 30,000 square feet, shopping carts will become the options that we will evaluate,” said Steve Gu, a co-founder and CEO of the Santa Clara, Calif.-based autonomous retail developer AiFi.
Gu's company has deployed about a dozen stores in Silicon Valley, Los Angeles, San Antonio, Amsterdam, Dusseldorf, Paris, Poznan and Shanghai. These stores are mostly less than 1,000 square feet, but AiFi is in the midst of a major rollout of 330 stores in partnership with Albert Heijn, Carrefour, Zabka and Loop Neighborhood. These stores range from 800 square feet to 10,000 square feet. AiFi is additionally venturing outside of supermarkets into hotel lobby stores and gas stations.
The coronavirus has given checkout-free technology developers a chance to hasten their plans. Consumers aren’t as willing to engage with checkout terminals or stand in crowded lines, so an option that allows shopping with very little engagement has more appeal than it did just half a year ago.
But the work to ensure older full-sized supermarkets can accommodate checkout-free shopping can’t necessarily be done as an emergency workaround. The economy likely won’t allow supermarkets to retrofit their entire chains with embedded cameras and sensors. Gu has already reported a coronavirus-related bump in deployments in China, particularly since the stores can sense physical distance between shoppers and ensure consumers and staff are not in stores at the same time.
The technology can work in stores of all sizes and types, Gu said. It’s more a matter of the cost of deploying the cameras in larger stores. AiFi is working to make the technology more scalable. The technology should also become less costly over time, he said.
“And putting the technology in a shopping cart may not solve every problem,” Gu said. “What do you do for people who don’t use shopping carts, but go to a larger store for a small amount of items?”
Berkeley, Calif.-based Grabango, which is one of the approximately two dozen firms building autonomous checkout technology, contends most checkout-free systems are designed for stores that do not exist, and as such have to be built out. Working within an existing chain is harder. Grabango’s clients include Giant Eagle, a supermarket chain that’s in the midst of a multiyear rollout of checkout-free technology, a strategy that includes testing different options for different store layouts.
The issue with most checkout-free technology on the market today is that the systems weren’t built for existing stores; they were created for specific layouts and store sizes that need to be built from the ground up, such as Amazon Go.
"To maintain high accuracy in a larger footprint store, most tech providers struggle to expand their sensor fusion approach in a cost-effective manner, failing to generate a positive return on investment for retailers,” said Andrew Radlow, chief business officer at Grabango.
To counter that challenge, Grabango uses computer vision and machine learning to manage tens of thousands of SKUs, which Radlow contends can work in stores that are more than 100,000 square feet.
“Maximizing low-cost sensor deployment to create overlapping coverage that guarantees the highest accuracy while maintaining the lowest cost per square footage is key to expanding checkout-free technology to larger stores,” Radlow said.
Standard Cognition CEO and co-founder Jordan Fisher in May told PaymentsSource the autonomous store industry isn’t yet ready to handle all store sizes, but progress is being made.
The San Francisco-based Standard Cognition earlier this year acquired Checkout Technologies, a Milan-based firm that uses AI to build retail technology in an effort to speed up that progress. The combined companies are honing the sensors’ ability to recognize items and shoppers in different environments. And like a lot of companies in autonomous retail technology, Standard sees Amazon not as a competitor but a signal to traditional retailers to step up their game.
"Computer vision technology is becoming more powerful and accurate every day; in the near future, there's no reason why large stores couldn't be outfitted with the same camera-based approach used at small stores," said Evan Shiue, vice president of strategy and finance at Standard Cognition. "Custom carts that require customers to work harder isn't an approach we would emulate, but it's always interesting to see Amazon innovate. It reminds retailers that they need to do the same."