The Costco-like online-only retailer launches Smart StockUp, a reordering program that predicts when consumers are running low on grocery items and sends reminders to order those goods.

It’s a common woe of grocery shopping. Discovering you forget to buy more soap—mid-shower.

Online retailer Boxed Wholesale aims to help consumers avoid such conundrums with a new feature called Smart StockUp. Smart StockUp predicts when shoppers will most likely run low on particular household products and makes it simple for them to reorder by sending them a reminder about the items they are likely running low on. 25% of Boxed consumers tried the new feature during its pilot test, Boxed says. The feature was released to all Boxed customers yesterday.

“It all boils down to convenience [and] not needing to keep a mental checklist of the things you have to stock up on,” says Will Fong, chief technology officer at Boxed.  

Boxed, No. 326 in the Internet Retailer 2017 Top 1000, is a web-only merchant of groceries and household products. Similar to Costco Wholesale Corp, (No. 8) it specializes in selling in bulk products. The retailer says it generated more than $100 million in online sales last year, up from $50 million the previous year. There is no membership fee required for consumers to buy from Boxed, delivery is free for purchases of $49 and more and packages usually arrive within two days. Boxed also sells its own brand of affordably-priced goods. The startup sells about 1,500 products.

Boxed is also taking Smart StockUp one step further with another just-released feature called Concierge, which not only predicts when a shopper is running low on goods, but preemptively fulfills the order and sends a customer products without them needing to engage at all. Concierge is geared toward B2B shoppers. For example, a company might use Concierge to reorder office supplies or snacks, Boxed says.


B2B orders make up 20% of Boxed revenue or about $20 million, Fong says. “We’ve rolled out Concierge to a small test group of B2B customers and are seeing strong results,” he says. “So far, there have been zero returns on these automatic shipments. We expect to roll out Concierge to B2B customers more broadly in the near future.”

It all boils down to convenience [and] not needing to keep a mental checklist of the things you have to stock up on.

A recent study from Oracle Corp. examines how consumers feel about retail technologies that use their personal data to make decisions on their behalf. It finds some shoppers my be leery about companies ordering for them. For example, 58% of consumers have a positive attitude toward having their grocer suggest a shopping list for their approval based on purchase history, social and environmental data, the study finds. Yet 54% of respondents feel it would be invasive for a grocer automatically charge and ship items based on purchase history, social and environmental data. The survey, polled 709 consumers in January.

Boxed uses machine learning to scan customer data and predict what shoppers are going to run out of. Machine learning describes the ability to program an algorithm that teaches a machine what it should be looking for so that computers can learn over time when they are exposed to new data. The machine-learning algorithm was created by Boxed’s in-house team of data scientists, who built it over the course of  three months. “Moving forward, we’ll continue analyzing our customer data to refine the algorithm,” Fong says.

Boxed is no stranger to using automation to drive sales and efficiency. Boxed says it spent tens of millions of dollars on a new automation system, which launched earlier this year, to triple the output of its 140,000-square-foot warehouse in Union, N.J., without needing more space or workers. Boxed has warehouses in New Jersey, Dallas, Las Vegas and Atlanta.