AI buzz is nothing new for Hungryroot
The AI-powered, direct-to-consumer grocer says it could go public and plans to increase its SKUs sixfold to 4,000 over the next few years. The artificial intelligence-powered, direct-to-consumer grocer says it could go public and plans to increase its SKUs sixfold to 4,000 over the next few years.
Hungryroot, the AI-powered grocery delivery service that does the shopping for customers based on their preferences, purchase history and other factors, is rapidly adding to its selection of SKUs and could go public, once the market improves.
That’s according to its CEO and founder, Ben McKean, who discussed the future of artificial intelligence in grocery at the annual Groceryshop convention in Las Vegas.
The company launched in 2015 with just six items. Over the next four years, Hungryroot gradually grew its inventory to 60. It secured $40 million from investors in 2021, and the direct-to-consumer grocer now holds roughly 600 SKUs in its inventory. McKean said he plans to increase that number to roughly 4,000 over the next few years.
“We believe that we have the financial profile of a company that could go public when the markets are ready or normalized …" he said. “We think we've really found something that's resonating with consumers. We think our technology is built in such a way where we can really scale it now and scale the impact of it now.”
Despite the low number of SKUs—the average grocery store has about 35,000, and even smaller grocers like Trader Joe’s have about 4,000—McKean said the company is “nicely profitable,” bringing in $182 million in revenue in the first half of 2023.
That positions the company to surpass $300 million this year, up from a mere $25 million in 2019 and $237 million in 2022. He said the New York City-based company’s profitability is driven by large orders and a business model that does not support on-demand purchases, like delivering individual orders of ice cream.
The average weekly order value for Hungryroot members is $125. “We get large baskets, which offsets a lot of the fixed costs associated with delivering directly to consumers’ homes,” he said. “And secondly, we don’t try to support the on-demand use case at all. We don’t believe in that. We believe that there is a use case for on-demand, but we don’t want to support that.”
Under the Hungryroot model, what’s most important is the day of the week each order is delivered, McKean explained. “As long as we delivered on that day, it doesn’t matter if it was fulfilled the day before or even two days before,” he said, noting that the company has three semi-automated fulfillment warehouses throughout the country. “So, our fulfillment costs are really attractive, and that’s led to the generally strong profitability of the business.”
While artificial intelligence was among the hot topics at this this year’s Groceryshop conference, the topic is old hat for Hungryroot, which has been working with AI since 2019. “So in 2019, when we decided to build a platform on AI, we didn’t call it AI, because that wasn’t the buzzword at the time,” he said, noting that the proprietary technology used to create its AI-powered shopping lists is known as a “large-scale constrained optimization algorithms.”
Such algorithms are perfect for solving problems that have a number of hard constraints, he said. “For example, if one of our customers tells us they're vegan, we can’t send them non-vegan foods, right? That would be a horrible mistake,” McKean said. “If we run out of inventory on something or we max out inventory, we can't put that in someone's cart, because we can't fulfill it. So, there's a number of hard constraints that are built into the algorithm.”
That algorithm is continuously optimizing for each Hungryroot customer and even takes into account external factors, such as the weather and time of year, according to McKean.
“That's a big opportunity for us down the road, but it's much more around customer objectives with their foods,” he said. “Eighty-two percent of our customers sign up with some sort of health objective, either trying to lose weight, lower cholesterol, even just eat more plants, and that's where it really helps to optimize around what are the foods that are going to work best for that person."
About the Author
You May Also Like