Taking Stock of Stockouts

Even before the coronavirus pandemic turned empty shelves into an industrywide crisis, grocery retailers had grudgingly accepted that a certain level of out-of-stock items in their stores is just one of those things they have to live with.

To be sure, they knew out-of-stocks were bad to experience, in general: While some shoppers who can’t find the one item they want when they want it will self-substitute, others will leave the store entirely and take the rest of their basket with them. Some may never return.

Undetected out-of-stocks, in the meantime, can trigger entire categories to sag and profits to vanish. And with a lack of visibility into the condition of stores, the wide variety of reasons that out-of-stocks occur, and the time-consuming processes to address whatever errors they do find, a sense of helplessness has pervaded.

“The problem absolutely exists,” says Jason Wirl, director of solutions engineering and marketing for Itasca Retail Information Systems, a West Des Moines, Iowa-based technology solutions provider. “It’s been 8% [out-of-stocks] on regular price items [and] 15% in promotion in grocery for years and years and years.”

That’s not even the bad news.

“The more that grocers are moving toward a digital business, the bigger the problem out-of-stocks are,” says Kevin Sterneckert, chief marketing officer of Symphony RetailAI in Addison, Texas. That’s because many stores today are now doing double duty as local e-commerce pick centers, introducing new demands on inventory and often leading to uncomfortable compromises. And increased capabilities of competitors offering e-commerce is also ramping up, with Amazon leading the charge to one-day fulfillment promises for some shopping baskets. That’s leaving very little room for a stocking error to go unpunished.

“The stakes go up in this new world,” says Richard Schwartz, founder and CEO of Pensa Systems, an Austin, Texas-based technology firm, calling out-of-stocks a “$1 trillion black hole” for retail. “If a customer buys online and goes to the store to pick it up, heaven help the store if the product is actually out of stock and unavailable. Studies show the retailer may lose the customer for life.”

Not surprisingly, each of the sources mentioned above has a stake in helping retailers recognize and begin to solve out-of-stock issues. While approaches differ, each is relying to a degree on newly available technologies such as artificial intelligence (AI) and computer vision now making their way to retail stores. The good news: They say it works.

‘Find and Fix’

Sources break down the technical approaches to out-of-stocks into two general camps: “find and fix” and “anticipate and prevent.”

Modern solutions for “find and fix” are literally rolling out to store floors today in the form of robots that patrol store floors, photograph and process shelf conditions in a flash and then generate alerts. Pensa is exploring drones to do similar work, and other vendors are advocating for shelf-mounted cameras. Retailers such as Schnuck Markets and Giant Eagle, using Simbe Robotics’ Tally units, and Walmart, a major client of Bossa Nova, are finding these units doubly effective because they not only ensure shelf and pricing integrity but also create better opportunities for store employees to serve customers, outsourcing time-consuming tasks.

Photograph courtesy of Simbe Robotics

“If a product is unavailable at the time our customer wants to buy it, we’ve missed an opportunity and disappointed our customer,” Giant Eagle spokeswoman Jannah Jablonowski said in a statement accompanying the announcement of a Tally introduction in the Pittsburgh-based retailer’s stores last year. “Tally helps us address these challenges by providing more precise and timely analysis of the state of in-store merchandise and freeing up staff to focus on customer service and guest interaction,”

A robot, drone or other smart camera can scan about 80 feet of gondolas every minute, says Sterneckert. An employee would take 10 to 15 minutes to find, report and remediate a single out-of-stock.

“Because it takes so much effort and time, it’s something that doesn’t get done. Because there’s pressure on labor at retailers to make sure they’re working on the highest [priority], this is one of those things that falls to the bottom,” Sterneckert says. “It’s done after you’ve moved merchandise, after you’ve taken care of the customer, after you brought in the new truck, after you’ve put the new merchandise out—after all those things have been done, then you work out-of-stocks and you usually don’t have enough time left in your shifts to do it right.”

 ‘Anticipate and Prevent’

For AI-enhanced demand forecasting, reductions in out-of-stocks are a byproduct of an advanced approach to supply chain management that can bring lots of other benefits. Wirl of Itasca says computer-generated ordering solutions “cure the disease” of out-of-stocks and contribute to improved availability, lower levels of inventory and other benefits.

“What we say is we don’t Band-Aid the issue. We don’t cover up symptoms. We cure the disease. We send an order that’s accurate by SKU and by store, and all the distribution center or the [direct-store delivery] vendor has to do is meet it,” Wirl says. “Because it’s coming from consumer demand, there’s no squabbling on who’s more right on the order. The consumer has told us what they want, how much they’re going to buy, and that's what we need to put on shelves.”

The company’s Itasca Magic solution is in place chainwide at retailers such as Lowes Foods, Tops and Raley’s, and most recently, at Weis Markets and Superior Grocers.

AI-enhanced demand forecasting brings a level of precision and agility to store sets that was previously unavailable, says Sterneckert of Symphony RetailAI.

“You need technology that has artificial intelligence and can make its recommendations without human intervention,” he says. “Meaning that the system identifies that you need three facings of product X and two facings of product Y in store one and the reverse in store two—and no product X in store three. Think about Walmart, for example: It has 4,500 stores. Even if they just opened up the plan-o-gram and closed it for each individual store, and that took 60 seconds, they don’t have enough time in the year to do a plan-o-gram for each store.”

Sources say the benefits of turning over ordering to AI-enhanced systems can get to the root causes of out-of-stock issues—whether it’s an error on the shelf or further upstream in the supply chain, or extraordinary events putting stresses on demand or supply, such as the surge in stocking up in the early weeks of the pandemic.

Raley’s, for example, was able to use Itasca’s ordering software to manage the supply challenges associated with temporary store closures due to the forest fires in California last year. And Sterneckert tells of Symphony’s system helping a client manage unseen ripple effects—namely, increased demand for bottled picante sauce—brought about by a spike in avocado prices.

“Most of the forecasting in the industry will look at a category or a demand group and give you calculated demand for an individual item within the category, or all items for profit, but it doesn’t take into account the rest of the store. They weren’t designed to do that,” he says.

But when avocados shot up to $4 each last summer, “it changed demand, but people were still entertaining and having outdoor parties. So what are they doing instead?”

The system determined it could shift more units of picante sauce to stores, Sterneckert says, “because if you don’t do that, you’ll run out of picante sauce and not know why.”