Reality Check: Can Simbe’s retail robot, Tally, break through?

Reality Check is an occasional series at Protocol taking a look at the prospects of some of the more audacious emerging technologies. In each story, we question a particular business and get the operators to tell us what they need to go from funding to success — and what might stand in their way. This week: Simbe Robotics.

The supermarket hasn't changed much in the century or so since its invention. And for most shoppers, the routine from aisle to cashier works well. But as retail stores increasingly compete with Amazon, they face a common problem: They often don't know when something they stock isn't on the shelves. A recent study from IHL Group suggests retailers are missing out on around $1 trillion in potential sales by not knowing what's on their shelves, according to RetailDive.

This is where 6-year-old startup Simbe Robotics sees opportunity. Its pitch is relatively simple: Why change how your store operates when you can add a layer of intelligence on top of what's already there?

Simbe's 6-foot, 85-pound robot, Tally, maneuvers through a store's aisles and determines when products are missing from shelves, when they're on the wrong shelf, when they've been sitting on the shelves for a long while, and whether they're listed at the right price. The robot can scan in two ways, either relying on RFID tags affixed to merchandise or by using computer vision to recognize the products on shelves from a proprietary library of images, as well as the price tags.

Some stores, like the Decathlon sports stores that Simbe is working with in San Francisco, use the RFID tags, and others, like supermarkets Giant Eagle and Schnucks, use the AI approach. The company is currently operating robots in six countries, including with Carrefour and Groupe Casino in Spain and France. But it's time for a reality check: Can Simbe break through?

How are things going?

Simbe raised $26 million in Series A funding in September, and penned a separate funding agreement with SoftBank to help it build an additional 1,000 Tally robots. Much of the last six years at the company has been devoted to putting together a working product to show investors and retailers, co-founder and CEO Brad Bogolea told Protocol. "We were operating with largely a product- and engineering-centric organization," Bogolea said.

Since the Series A, the goal has shifted to scaling the company, getting into more stores, and building out sales and customer-service teams. Over 2019, the company's head count grew by 125%, Bogolea said.

One positive for Simbe is that retailers are realizing the potential benefits of automating various parts of their business to speed up sales. According to a recent report from market-intelligence firm Tractica, global shipments of logistics and warehouse robots are set to explode over the next few years, from 194,000 units in 2018 to 938,000 units by 2022. But it's difficult for retailers to know which companies can make a difference, especially in a nascent field like robotics.

Get what matters in tech, in your inbox every morning. Sign up for Source Code.

The plan for Simbe in 2020 is to install heads of customer service, sales, and the logistics of manufacturing its robots, using the funding round as a springboard to grow beyond the initial test deployments. Bogolea said the company is choosing to prioritize growth into new areas over profitability, but that could change. Right now, building a Tally is less expensive than "the cheapest car you could buy on the market new," Bogolea said, but that cost will come down closer to the cost of "a high-end MacBook Pro" when building at greater scale. (Back-of-the-envelope math on that scale would put the cost of the 1,000 Tallys at $3 million to $13 million.)

Are robots really the answer?

Some are skeptical of the field altogether. "I don't think for most people robots — apart from being interesting to look at — are going to be what makes for a better experience," Steve Rowen, managing partner at Retail Systems Research, told Protocol. Many retailers, especially grocers, operate with thin margins, so any bets on new technology are going to be a tough sell. Asked whether retailers are interested in novel technologies, Rowen said, "If it's Amazon, sure, if it's anyone else, no — these are not people who are known to be bleeding edge."

Bogolea acknowledged this challenge, saying the company has primarily dealt with retailers' CIOs and CTOs to set up small-scale tests of Simbe's bots. When the discussion shifts to the next level — rolling out Tallys across a company's entire chain — the CEO and the board generally need to get involved.

"I think one of the challenges many robotics companies have is helping their client base really understand the business case around this technology," Bogolea said. He said the company is close to rolling Tally out across one FMCG company's entire retail chain, but wouldn't disclose which.

How will Simbe make money?

"The power that we have is at every quarterly board meeting we can say we need X number of robots in business recognizing revenue today to break even," Bogolea said. "We know precisely what that number is, and we're always on the edge of it."

Simbe operates like many SaaS companies, in that its margins come not from the robots but from the subscription to the software it's selling. "The capture mechanism is instantiated through a physical robot, but we have demonstrated that our computer-vision platform can operate on a mobile phone. We've demonstrated it can be operated on a fixed camera. We have demonstrated that it can be operated on a drone," Bogolea said.

And that's probably good news for the company's long-term prospects. "Small cameras looking at shelves have a much better chance of telling you when something's out of stock than a robot that's going to scan through your store maybe three times in one day and probably save you one minimum wage worker's hours' worth of work," Rowen said.

The Tally robots tie into a retailer's existing inventory-management software, and offer up a near-real-time look into what's going on in any given store. Bogolea showed Protocol what that view looks like for Simbe — the company can see Tally robots operating in stores around the world. But he said that, for most retailers, the information Tally collects manifests as notifications in existing software. Many retail employees have handheld devices or tablets, and Tally can ping managers to let them know when something needs to be addressed. It can even alert suppliers — say if a store is running low on Coca-Cola products — if the stores prefer.

The information captured by these robots could also be useful to product manufacturers, market-research firms like Kantar or Neilsen, and companies like Instacart that shop for others. Bogolea said Simbe is working with its partners to think about how to monetize the data beyond inventory checks. "I think our ideal goal is really being able to maximize the value of that information across the entire value chain," he said.

Bogolea said he expects far greater penetration of robotics in "everything greater than convenience stores" over the next five to 10 years. That's the same time frame he gave for the company's path to success.

How might Simbe fail?

When asked how Simbe might fail to succeed in that time frame, Bogolea cited poor execution. "Most startups fail because they either run out of money or they're not solving a real problem," he said.

Simbe has competition from companies like BrainCorp, which recently told Protocol it's exploring inventory-tracking on robots it already has in facilities across the U.S., and Bossanova Robotics, which has raised around $70 million over several funding rounds. But Simbe also faces a market ripe for disruption. Companies like Fabric, Takeoff Technologies, Attabotics and Alert Innovation have raised a combined $240 million to replace traditional warehouse and retail facilities with micro-fulfilment centers, according to data compiled by PitchBook.

"Right now, we have about as many people telling us that robotics are a hammer looking for a nail as we do saying that 'I've seen practical uses of how I can free up my employees to engage with customers more,'" Rowen said. "For every single one of these sort of staid models, there are just some clever people sitting around in an apartment somewhere saying, 'How could I reinvent this industry?'"