Keeping track of inventory within a grocery store or other fast-moving goods environment remains a huge problem for physical retailers, especially as companies like Amazon and Walmart can quickly deliver goods to customers in their homes in a matter of days.
Many of these retailers are turning to robotics companies to help them track inventory levels, including Simbe Robotics. Its Tally robots have logged more than 15,000 miles of in-store inventory scanning, accumulating data of in-stock and out-of-stock items for large grocery, drug store and value chains around the world.
In recent months, the company (a 2019 RBR50 honoree) announced partnerships with grocery chains such as Schnuck Markets, Giant Eagle, and sporting goods retailer Decathlon to bring Tally robots inside their store environments.
Robotics Business Review recently spoke with the company’s CEO, Brad Bogolea, about these announcements, ways the company is adding value for retailers, and new technologies to track even more in-store items.
New locations, retail markets
Q: You’ve had a slew of announcements with partnerships in the retail space. Are there any retail locations where the Tally system wouldn’t work, or does any kind of situation within retail make it possible to use Tally?
Bogolea: When we designed the original Tally solution, it was really focused on packaged goods-based retailers, often referred to as fast-moving consumer goods retailers. These were traditional markets, grocery, drug, value chains, and mass-market hypermarket stores. That was the original sweet spot for the product, which is still our core business today. Folks like Schnuck and Giant Eagle, and a number of other clients we work with fit in that category. One of the unique things we did about a year ago was publicly announce RFID capabilities in addition to the computer vision side. That’s opened us up to clients like Decathlon, and some high-end retailers that focus on electronics, sporting goods, and higher-end clothing.
So our focus is still on larger format store environments, drug stores or value chains – the smallest store environments we’ve been in are larger than convenience stores. On the RFID side, you may have smaller footprint, boutique retailers, but those have a very high density of product. So, 50,000 or 100,000 tagged items in a space. Whereas, if that was a fast-moving consumer goods store, it would be much, much larger. So that’s kind of our focus from a market perspective. If you look at the retail market globally, there are a lot of retailers in that bucket. We target the largest 250 global retailers. Today, that client set maps to more than half a million physical stores.
On the customer side, we have a number of folks that we’ve announced – Schnuck, Giant Eagle, Groupe Casino, one of the largest European retail players. They have a global presence, which means they have stores throughout the world. We announced some work publicly with them earlier this year as well. In addition, we have a global partnership with SoftBank and their robotics group. So we’ve been doing work with them on the ground in Japan.
Q: Are the larger retailers or chains more willing to invest in robotics than the smaller ones, like a convenience store or a smaller mom-and-pop store? Is that why you focus on the larger ones?
Bogolea: Even someone like Schnuck, which is a regional grocery chain, they’re still a $4 billion grocer. For us, you get a network effect in targeting these larger players, as it allows you to essentially scale your business and scale your deployments faster. With this sort of robotic technology, whether it’s front-of-store or warehouse, there can be high up-front costs, startup costs or integration work, those things can get greater scale by focusing on those [larger] players.
We do see that there are specific stores that are relevant to Tally. Yes, Tally could add value in a convenience store, but they have a super small footprint. The payback period will be a lot longer than grocery, in an environment that small. You might be able to cover it with fixed infrastructure, at a more competitive cost. If you look at what people are doing with the Amazon Go model and other strategies, that may be a better fit for some of those markets.
Q: What are the biggest challenges for retailers that your system is helping to address?
Bogolea: The biggest problem is really helping retailers with the challenge around store operations. We often refer to this as the blind spot, helping retailers close the gap between supply chain and point-of-sale intelligence. This is such a huge problem and it affects consumers in these stores regularly.
For example, one out of 10 products in a customer order isn’t on the shelf. A quarter of Amazon’s revenue comes from people that went into a physical store to find a product and it wasn’t there. The inventory distortion-related issues is where we see the key focus and aligning around those pieces – is product stock in the right place, and does it have the right price.
That said, we’re seeing retailers really leverage this information for other folks across the value chain. So the data we collect is not only relevant to the retailer itself, but also to brands. Companies like Unilever and Procter and Gamble care about things like ‘Where are their products placed in store? Are they being effectively restocked? What is their share of shelf against their competitors?’ We see value of the data that we’re collecting, and we believe strongly that we’re building one of the most valuable datasets in physical retail.
In addition, it further extends across the value chain. For companies like Instacart and others doing click-and-collect, or click-and-deliver services, it’s not uncommon for one or two out of 10 products in a custom order to not be on shelf. Those folks typically don’t have the ability to go into the retailer’s back room to pick product. So our information is super helpful for them to understand things like product availability – is it actually on the shelf, but also helping them define the optimal pick path.
Some of our clients have been using Tally’s data and integrating it into their consumer mobile apps. I’m sure you’ve walked into a retail store before and went, OK, where in the world in these 20 aisles is this product? Leveraging Tally’s information to help people find it in store has been valuable.
We’ve also seen interest in using Tally as a platform for other capabilities, things like health and quality of produce in the store, spill and hazard detection, security, temperature, environmental condition monitoring, asset tracking within the environment. We see Tally as a platform to unlock more value per store. Those are capabilities that we’re thinking about for the future.
Q: There also seems to be areas within a grocery store where inventory tracking via robot might not be as easy – such as the produce section, bakery, frozen foods, the meat department – can your system handle those different areas within a store for the purposes of inventory management?
Bogolea: Our core focus across all of our production clients today are helping them understand what’s traditionally referred to as center store grocery. That’s the core shelving, both on the aisles as well as the end caps, and the area traditionally referred to as health care and beauty. In an average-size supermarket chain, typically we’re monitoring between 25,000 and upwards of 60,000 products just in that area. Traditionally, for these stores, that’s their cash cow.
Of course there are other areas, such as produce, deli, bakery, floral, these types of things. These are capabilities that we’ve been actively investigating. Looking at it from an R&D perspective, as you can imagine, looking at a display of apples is a very different computer vision problem than looking at a product that’s on a shelf.
Q: Could you take the same technology that you’re using and modify it for these environments, or would a different type of technology be needed?
Bogolea: I think much of it is the same technology, so Tally would remain in the same form factor. Additional sensor technology may be coupled. Much of what we analyze stores with today is capturing 2D and 3D data, which would be highly relevant for things like produce and other categories. But there’s lots of other sensor technology that you could imagine being leveraged in the stores. For example, lots of food producers use things like hyperspectral and multispectral imaging. Could that type of technology have value within retail stores, to assist with things like the health and quality of produce?
In other situations – the average supermarket chain typically spends about 10 hours per day, per store, checking on things like temperature and environmental conditions, where something like Tally could have value there. Many stores perform a store walk every hour where they’re looking for hazards. These are the type of future value streams that we look at capturing with the existing platform. It may require additional configurations or different sensor technology, but the Tally itself will largely remain the same.
Q: Do you see either Tally or a new robot that you develop turn into a multipurpose robot? For example, we’ve seen robots in stores that handle spill detection, or they clean the floors. Do you see a day where a Tally or similar robot can track inventory, but then also detect spills, and then clean it up? Or is there too much intelligence needed for that?
Bogolea: Our focus has really been on the data capture and data analysis. If you look at our road map and our vision, we really want to focus on building the most valuable dataset in physical retail. So much of our applications align around that and focus less on the physical manipulation of goods through the environment, where other people may be focusing in the space on picking a product, cleaning floors, or those sorts of things. Our retail clients not only get efficiency by deploying this type of technology, but there’s so much value in the data that we’re collecting. I think that’s what separates us from a lot of different robotics companies that are purely focused on efficiency or automation. We have those same components, but we have this data driven strategy.
In these environments, goods move very quickly. They’re called fast-moving consumer goods stores for a reason. But we have heightened expectations across the retail industry. With Amazon acquiring Whole Foods, we know they’re not going to be shy about technology-based investment. Consumers’ expectations are continuing to increase around instant gratification and satisfaction. We’re in this wave where retailers are beginning to heavily invest in technology and data-driven strategies, rethinking their business process and the customer experience. So we’re super-excited by it all.
Q: Robots in grocery stores have such a high visibility factor, because the customers are right there in the store. How much work do you focus on human-robot interaction? We’ve heard companies say that they want their robots to be invisible, they just want them to be seen as a machine – while others say they want robots to interactive with customers. What are your retailers telling about interactions between customers and the Tally robot?
Bogolea: We’ve logged more than 15,000 miles in physical stores during normal store hours. Something that’s really unique about us is we’re one of the first companies on the planet to bring robotics to an environment that most people frequent every week – grocery stores and retail store environments.
When we designed Tally, we were really focused on designing a solution that would blend seamlessly within the environment, but be intuitive enough that people understood what it was doing. Retailers have really branded Tally as a power tool to their retail associates and store teams to develop a better store experience. We haven’t dealt with a lot of the challenges around the robots-and-jobs conversation because people really see it as a power tool. It’s not taking somebody’s job, it’s doing a couple of mundane and monotonous tasks, and allowing them to focus more on customer service.
That said, the bulk of our deployments are in areas like the Midwest – it’s not Boston, it’s not San Francisco – people don’t see autonomous cars driving on the road. This is really people’s first interaction with this type of technology.
There are two buckets of retail clients that we have today. There are those like Schnuck and Giant Eagle and Decathlon, which have been very vocal about their work. So people have often seen it on the news or in the press even before they come into the store, and they clearly understand the value.
I would say that retailers that operate more under the radar, there’s definitely a bit more of a surprise, where some more education that has to happen. But overall, things have really been quite positive.
Q: What’s next for the company? Where do you stand in terms of either raising more money, or looking to grow revenue? What’s the plan for scaling up?
Bogolea: So much is positioned around growth and scale. We have scale projects operating on three different continents, and deployment work – meaning robots on the ground – with more than a dozen retailers. Much of our business is not only focused on scaling those existing clients – which many of them have plans to take this technology to all of their stores or nearly all of them – but to focus on the rest of the market. We talked about the Global 250 target – today we’re just servicing a really small amount. So there are lots of retailers to work with, not only in the U.S. but abroad.
On the product strategy side, we talked about the ways we can add value with either existing technology or slightly changing the product in some ways. We continue to focus our road map of how we add more value in these stores.
Q: What advice would you have for robotics companies in similar situations, or in a different market where they’re at the startup stage and they’re looking to scale up? What lessons have you learned along the way?
Bogolea: I think our history as a company is that we focused on being highly capital efficient. We really focused on the problem we were solving. There’s so much exciting work going on in robotics technology. It’s really easy to get enamored with the technology side, and you end up with the classic robotics problem of a technology in search of a problem.
So really encouraging people to think thoughtfully around the customer pain point and solve the business case associated with it, and then map or take that product or technology to scale. As an organization, focusing on those things has really helped us yield a lot of fruit.