Cashierless stores are fast becoming a reality, thanks to Amazon’s aggressive expansion of its Amazon Go chain and an emergent class of startups looking to nab a slice of the nascent sector. It’s a win-win for retailers: With a checkout-free arrangement, customers don’t have to contend with lines or self-checkout machines, and stores save substantial overhead in payroll — theoretically, at least.
One of the space’s newest entrants is Grabango, a Silicon Valley startup founded by Pandora Radio cofounder and chief technology officer Will Glaser. Since its incorporation three years ago, Grabango has amassed 17 patents related to checkout-free technologies and signed on four U.S. clients that it claims serve over 600 million shoppers per year.
The early momentum has investors impressed. Grabango today announced that it has raised $12 million in a Series A funding round led by Propel Venture Partners, with participation from Ridge Ventures, Abstract Ventures, Commerce Ventures, and Founders Fund. This brings the 37-person company’s total raised to $18 million and will see Ryan Gilbert, general partner at Propel, join the board of directors.
Glaser, who serves as CEO, told VentureBeat that the cash infusion would fuel product development and Grabango’s expansion to new client stores.
“Our goal is to forever change the way people interact with things,” Glaser said. “We don’t want to be the center of attention, but rather for our systems to work so smoothly in the background they are completely taken for granted.”
Grabango isn’t the only company that’s working on frictionless shopping experiences. In June, Japanese telecom company NTT East teamed up with startup Earth Eyes to create AI Guardsman, which attempts to prevent shoplifting by scanning camera feeds for suspicious activity. In November, Intel and Tencent took the wraps off a pair of artificially intelligent (AI) products for retail that can track customers and perform complex object detection.
But, according to Glaser, Grabango’s product is much more scalable than many of its competitors’ offerings. It’s designed to work in stores with more than 100,000 square feet of floor space, he said, and to track hundreds of shoppers at once. (Current store deployments range from 2,000 to 50,000 square feet.) Moreover, the system can keep tabs on “hundreds of thousands” of SKUs, obviating the need for store employees to manually inventory store stock, and it integrates with existing point-of-sales systems and store operations platforms.
The ultimate goal is high fault tolerance, Glaser says. The bulk of processing is performed on-premises by a dedicated PC that doesn’t require an internet connection. Grabango’s item-tracking cameras — which were designed in-house — are lightweight and relatively inexpensive to manufacture, reducing the cost of install. Stores benefit from greater vision coverage, which helps Grabango generate better three-dimensional floor plans of stores and reduce the chance a purchase isn’t recognized as the result of occlusion.
The system’s machine learning algorithms — which were trained using reinforcement learning, an artificial intelligence (AI) training technique that uses rewards to drive agents toward goals — can recognize objects with 99.9 accuracy, Glaser said. That means they make a mistake one time out of 1,000.
“Everything is mission critical at a store — the system needs [to] never go down … [As] we iterate [on the system,] we look at successes and failures, and every time there’s a failure, we use that as an opportunity to train the machine,” Glaser said in an interview. “All [of these ideas] together — the large scale, the high-reliability focus, the low cost to deploy, and the integration with existing stores — is why we consider ourselves enterprise-class technology.”
These aren’t the only ways Grabango is setting itself apart from the competition. Unlike, say, Amazon Go, which requires customers to download an app and log into an account before they begin picking things off of shelves, Grabango-powered stores retain their existing cashier setups. Shoppers have two options: (1) checking out conventionally with cash, a credit or debit card, or SNAP benefits or (2) opting into a cashierless experience with Grabango’s app. In the latter case, checkout is automatic — customers can walk out without scanning a barcode or confirming that they’ve finished shopping.
Grabango’s market is currently in the U.S., where Glaser says its focus will remain for the foreseeable future. The optimistic plan is to capture the top 30 grocers and 10 convenience store chains in the U.S. as clients in the coming years.
To do so, it’ll have to go head to head with a bevy of new startups, most of which have opened proof-of-concept stores or partnered with brick-and-mortar retailers. Standard Cognition became the second company to open a cashierless store in San Francisco, following hot on the heels of Zippin in August. Trigo Vision recently inked a deal with Israel supermarket chain Shufersal for 272 cashierless stores. And, of course, mega retailer Amazon has Amazon Go locations in Seattle, Chicago, and San Francisco.
But Gilbert’s convinced that Grabango has a fighting chance.
“Grabango hits the target for brick-and-mortar retailers aiming to deliver a smoother shopper experience and a better bottom line,” he said. “We’ve looked at all the checkout-free technology providers, and Grabango is the only true enterprise scale provider in the market today.”