Transcription
What are you doing? Considering getting up? I'm going to take a shower and shave and take you to a lavish lunch.
[Music]
If you love listening to this show, please consider giving a rating and a review on Amazon Alexa or wherever you listen. We want to continue bringing you this amazing content, and part of our ability to do that means that we need a big audience. They may not seem like much, but rating and reviewing the show will help more people find us, just like how you found this show. Simply, on any podcast platform, search for the show, scroll down to the bottom, and push 5 stars. It's that easy. Thanks for supporting the show.
Today, I'm joined by Brad Begolia, who is the CEO and co-founder of Symbi Robotics, which makes the Tally robot that handles shelf auditing and analytics to ensure merchandise is always stocked. Welcome, Brad.
Scott: Thank you so much for having me.
Yeah, so you've spent many years in the energy and wireless sensor industry before starting this current endeavor. I'm curious to know and get your sense in terms of how has the smart sensor industry evolved over the years?
Yes, absolutely. So I got my first start in sort of IoT back actually in 2005, I think long before the term was really coined. But, you know, back then, I felt strongly we had a deep energy problem and very little visibility into how and when we use energy in our homes and in these sorts of things.
With the advent of sensing technology like low-cost wireless sensor networks, it presented a really unique opportunity for homes, small businesses, as well as large energy utilities to really rethink the way they monitor and manage energy.
So prior to starting this company, I spent about 10 years bringing various types of products and solutions to market in that space.
That's terrific. And certainly, when it comes to smart meters and you're talking about things around utility energy consumption, it's amazing how information and knowledge informs us, and not just the consumers, but also the industries as well, so that they're making better decisions.
Very much so. You know, in using energy in your home, we used to often refer to it as like driving a car without a speedometer. You know, you don't really know how you're consuming energy until you get your energy bill at sort of the end of the month.
So, you know, technology and greater visibility into these types of problems bring a lot of awareness and behavioral change.
Yeah, that's a great segue to talk a little bit about Symbi Robotics. If you could talk a little bit about the combination that eventually led you to launch this startup, as well as at this point, you know, the robotics is jam-packed with capabilities, including sensors and different types of modalities to be able to understand what's happening on the shelf system. If you could just start to get into some of those details.
Yes, absolutely. As I was kind of finishing my tenure in sort of the energy space, I was part of a really large company and really wanted to get back to my entrepreneurial roots. A good college friend of mine had been working in robotics throughout his career.
Through my interactions with him, I really started to understand that we were kind of on the precipice of this robotics revolution. During that journey, he ended up getting a job at Willow Garage, which was this world-renowned robotics think tank that was actually here in Silicon Valley.
I had the fortunate opportunity to spend some time with him there and just see all of these emerging robotics applications. However, very few of them were really being applied to real-world business problems.
So, long story short, Willow disbanded in about 2013, and this individual, my good college friend Mirza Shah, we ended up getting together and starting Symbi. We felt strongly we were at the beginning of this robotics revolution.
Yeah, it's a fascinating story. And of course, you know, robotics has so many applications for it. I'm curious to know, as you guys have, you know, you could have started with the sensor only, but you've chosen to have it in a form factor that has robotics as hardware plus software and all the firmware and everything else that's involved.
How has that journey been? Maybe harder than a typical SaaS or software-based company?
Yes, absolutely. You know, the real challenge with robotics today is much like the early days of sort of mobile or the PC era. Robotics companies are really forced to be full stack, and as you described, really build sort of everything.
Fortunately, there has been some great open-source software development work, particularly that work that came out of Willow Garage, and sort of the Robot Operating System and other tools that have begun to make it easier to build, deploy, or sort of manage robotic applications.
But we felt strongly in the type of retail spaces that we operate in, really the largest physical stores on the planet, the business case for fixed infrastructure, you know, fixed sensor technology didn't really work today.
It may work long term in the future, but using a fully autonomous mobile platform essentially allows you fixed infrastructure on wheels, and you’re able to capture these large environments very effectively and very efficiently.
Yeah, so with Symbi, can you get into the specific pain points that you guys are trying to solve for retailers, and why do we need robots, and why can't humans do that work?
Yes, absolutely. In the global retail industry today, we often have this problem we call the retail blind spot, and it has to do a lot with understanding the true state of merchandise or inventory within physical stores.
But in stores like supermarkets, they often have 50 to 100,000 unique products that they sell. Given these products flow from the manufacturer to the retailer's warehouse, you know, on trucks ultimately to the stores, there's lots of room for distortion to actually happen.
So it's actually a very challenging problem for retailers to truly understand what is physically on their shelves, you know, what's stocked, what's in the right place, you know, does it have the right price.
Today, this is a global problem that actually costs retailers somewhere between one and two trillion dollars annually for not having clear visibility into this problem.
What this means in short is, you know, often a product a customer is looking to purchase is not on the shelf, and the retailer doesn't always know that.
So, you know, really how Tally addresses this problem is it frees up sort of the mundane or monotonous task of trying to leverage human labor to perform these audits. Those audits are very time-consuming.
So we knew a robot could actually do this much, much better, faster, and really become a power tool to these store teams to allow them to focus on higher-order customer service tasks.
So in terms of the capability, can you go into a little bit more depth around the kinds of features that the robot has and how is it able to ascertain what the quantity and what the situation is happening on the shelves themselves?
Yes, absolutely. So Tally today operates really in the largest physical retail stores globally. You can think of this as your large format grocery stores, your mass merchant stores, drugstores, value chains, sporting goods stores.
Tally really has two primary sensing modalities to analyze the state of product or inventory and store operations. In grocery-style stores that sell lots of packaged goods, we heavily make use of computer vision.
So as Tally is driving up and down those store aisles, it's capturing high-quality 2D and 3D imagery of the store shelves, which we then use computer vision and machine learning to understand specifically which products are not stocked, are the price tags correct, are products misplaced, these sorts of things.
The other sensing modality we have is actually leveraging RFID or radio frequency tags. So Tally has the ability to excite passive RFID tags that may be used to monitor higher-value merchandise such as clothing or sporting goods.
So we've operated in these types of stores as well. Now, in the future, again, we're not quite there yet, but when sensors become so ubiquitous and so costless that every package or maybe even non-packaged things like, let's say, an apple can afford to have a sticker-based sensor, so that we can track and geolocate the location of these things, do we still need to be able to somehow tally them or quantify them?
Yes, you know, most retailers would love to get to the world of sort of a fully human scanless environment as far as sort of taking inventory.
The real challenge is many of the products we're talking about here, especially the products you see in the grocery space, are maybe not applicable, at least in the near term, for some of the sensing technology given the small gross margins on things like a can of baked beans, a roll of toilet paper, you know, these sorts of things.
You're not going to put an RFID tag on these items. The other challenge that retailers have is really unifying all of their manufacturers to do this.
So what you found is for solutions like RFID, it has been largely cost prohibitive for many of the grocery markets. But you see great penetration in higher-value retail stores like clothing, sporting goods, and home goods.
So we see a far greater opportunity for penetration for that type of sensing in those environments. And I think that's the reality, is that that fractured, inefficient market and supply chain is going to likely continue, so you really need to help by addressing it this way.
Now, in terms of some of the kind of hard quantifiable bottom line or the value, can you get into a little bit of specifics of how companies are deriving benefit in terms of cost efficiency or better inventory and higher sales velocity?
Yes, absolutely. So our primary customer is, of course, the retail, and we typically sell in at sort of the senior leadership level to these types of businesses.
They look at rolling this type of solution out ubiquitously across their stores. The type of return on investment or business case value that they're able to achieve is, first and foremost, actually freeing up their labor to focus on more meaningful tasks.
So in your average supermarket today, we typically see retailers spend anywhere from 30 to 100 human hours per store per week performing these types of shelf auditing tasks.
We can now better enable those employees to more intelligently restock based upon our information and spend really more time with customers.
In addition, through better data, you're closing these gaps in the supply chain. A product may be missing due to a shrink-type scenario. The product may be in the back room. How do we get that product back on the shelf faster?
So we have demonstrated proof of our ability to actually increase sales and margin contribution to these retail businesses through better data.
And then lastly, this data is actually incredibly valuable to multiple stakeholders. It's not just the physical retailer itself, but the biggest brands and manufacturers in the world, people like Unilever, Procter & Gamble, really want to understand how their products are being executed at the store level, how their products may look in comparison to their competitors.
So we actually believe strongly this is an incredibly valuable data set to other parties like these brands, as well as folks like the Instacarts of the world that are picking all of their online grocery orders in physical stores today.
Having accurate inventory information is actually very critical to execute on that business.
Yeah, I think you make a really excellent point about the fact that, you know, product manufacturers to the fast-moving consumer goods brands, it's oftentimes as much as, you know, they try to get the data through the various distribution channels, it's very difficult to actually get truly real-time accurate data.
So this gives them almost direct access to the data that they've been really desperately looking for, and that then informs not only the manufacturing, the supply chain, but also the product innovation in terms of extensions and so forth.
Exactly. Let's get into a little bit about changes in retail now. You know, retail, I think if you think about COVID-19, you mentioned some of the delivery players that are out there.
So it's shifting quite a bit in the sense of, you know, retail overall, and again, I'm not talking about this food category, but has been hit hard even before the pandemic, and now it's been hit harder.
We're seeing just a ton of bankruptcies, refilings, restructuring, and then, of course, the shopping behavior has changed quite a bit.
So, you know, if I go to Whole Foods, I actually see more delivery workers than the actual shoppers that are shopping for themselves in stores.
How are all these different changes affecting your business?
Yes, absolutely. You know, so COVID as a whole has been a massive force in the retail space today.
From a big picture perspective, grocers, which is really our primary market, are seeing all-time high sales. The reason for that is, you know, the vast majority of people that used to eat at work or their kids used to eat at school, you know, more people are eating at home.
So there's been an increased demand throughout these environments, and because of that, there's been much more pressure on sort of the supply chain.
So I would say as a whole, COVID-19 has actually made an even stronger case for the need of better data and automation within these physical stores.
Other aspects of retail have suffered more so. You know, if you look at sort of the soft goods sector or businesses like sort of Gap and Old Navy and sort of Victoria's Secret, with many of the shelter-in-place restrictions, these businesses were viewed as non-essential.
As such, you know, they've had to sort of close their doors and really rethink their strategies. But if you look at the North American grocery markets, you know, it's a more than a trillion-dollar market today.
And although people like Instacart and other online grocery services are increasing quite drastically, all of these orders today are served out of physical retail stores.
We think this is a very meaningful trend for our business because to do that work effectively, you really need accurate inventory and product location information.
Yeah, absolutely agree. I couldn't agree with you, especially on that last statement.
The one thing that I wanted to kind of explore a little bit is that, though again, you know, the grocery format isn't going to go away anytime soon, and certainly the delivery platforms are not going to look to develop their own physical assets.
That's going to be capital intensive. But we are seeing some move from more towards kind of automated warehouses.
So, like, let's say Walmart having a smaller grocery pickup or delivery within a kind of an urban center, or certainly Amazon getting into more of the food, especially with the acquisition of Whole Foods and delivery.
How are automated warehouse models potentially going to change the landscape, or will it have any type of impact on your business whatsoever?
Yeah, it's a really great question. You know, today in the grocery space, as we described, really all the global orders are really being picked out of physical stores.
I think one of the real challenges with grocery, and I think one of the reasons, you know, Amazon pursued Whole Foods effectively, is to do fast-moving consumer goods and perishable products well, you really need to be as close to the customer as possible.
So what you often find is these large industrial warehouses that typically serve as e-commerce centers for soft goods, home supplies, these sorts of things, they can operate large distances from their target customers and sort of manage delivery through those means.
But I think what Instacart and others have sort of highlighted is the need for these distribution centers to be as close to a store or to people as possible.
So the current trend we're seeing a little bit is some retailers are actually converting some of their stores to dark stores, and they actually become pick facilities to serve these types of online orders.
What I would say about the state of robotics today as it relates to products that are in a physical grocery store is you have such a myriad of different form factors of items.
It's incredibly difficult for most mobile manipulation robotic appliances to actually pick these types of products effectively.
So I think for a long time, we're going to still see human in the loop, and over the next sort of five or ten years, I think the physical stores themselves will be the primary location to serve these orders before micro-fulfillment becomes more cost-effective to actually deploy at a larger scale.
Yeah, I think that's a pretty accurate forecast of where things are going to go.
The question I was going to ask, there were a couple of questions. One is the role of supply chain and how critical it was, especially in the early phase of the pandemic, where we saw low inventories of certain goods, both perishable and non-perishable.
And it's gotten to a point where it's even affecting our diet and our lifestyle. So I think if you look at the consumption of meat as an example, it's dipped down quite a bit.
So it's interesting to see how supply chain and the area that you guys are working on is so critical to the nourishment and the movement of goods and making sure that the needs actually get met.
So you guys are definitely feeling a very critical role.
The question comment that I was going to ask you is, you know, aside from some of the more technical things that we're talking about, how is the robot Tally being perceived by the customers who actually walk in the aisles?
Yes, we have robots in more than six different countries today, so we've had the fortunate opportunity to actually bring robotics to an environment that most people frequent every week, and that's sort of their neighborhood grocery store.
So as you can imagine, you know, when most of these folks go out to buy milk, bread, or juice, or sort of produce, they don't expect to see a fully autonomous robot in this environment.
But, you know, I think our team has done a remarkable job in thinking about how do we design a solution that blends seamlessly within this environment, you know, is small and slender, very intuitive, and staying out of sort of shoppers' way.
We have logged more than 50,000 autonomous miles in these physical stores, so we've had lots of various shopper interactions.
But at the end of the day, you know, Tally is really designed as a sort of a shy robot. It's just there to sort of analyze the shelves and, you know, really sort of stay out of customers' way.
So, you know, although there has been some surprise, I think in today's world, as people are beginning to understand the future of autonomous cars, and they have a robot vacuum cleaning device at home, you know, these types of applications are really becoming far greater accepted.
And particularly on the store team side, you know, the store employees in these stores are really recognizing Tally as a power tool, much like a power tool would be to a carpenter, because it allows them to actually do their job more efficiently and drive greater accountability and visibility across the supply chain than they actually had before.
Well, I think the other thing is, for more perception-wise, it's interesting for shoppers that have had the good fortune of encountering Tally themselves, is that it gives the impression that the grocery store, the retailer, is really upping its game and is on top of things to see such capability right on the floor itself.
So I do think that it's actually welcome.
My last question for you is in terms of lessons learned. You know, if you could share any type of product or project failure and the lessons learned that you can share with others.
Yeah, absolutely. You know, as a company, we've had many lessons learned and, of course, sort of many failures throughout our journey.
As you described sort of early in the call, you know, companies like ourselves really have to manage far more complexity than other businesses that we may see in sort of Silicon Valley today.
Some of the ways that we've done our best to really manage these types of aspects is, you know, in the early days, we spent a lot of time with peers that were building robotics companies throughout Silicon Valley.
None of us were sort of directly competitive, and we actually found there were great forums for sort of knowledge sharing because many people were sort of running into the same particular challenges.
The other challenges that we encountered throughout our journey is, you know, when we started the business in late 2014, 2015, very few institutional venture capitalists were actually investing in robotics at that time.
So we had to be very scrappy and resourceful, even though our team had prior career successes and wins. We had to take a unique approach to fundraising sort of early on and largely target angel investors and some alternative sources of capital to really sort of get the business off the ground.
Today, there are far more investors interested in the world of automation and also sort of what the future of retail and commerce looks like.
So those were some of the other challenges that we've had to navigate.
As a company, we've had a lot of principles within our business where we've always had this principle of sort of robots always running.
If we're sending robots to clients sort of halfway around the world, we really need to ensure that we're not going to have to send people regularly to service them and these sorts of things.
So we developed a number of thoughtful initiatives like robots always running internally to really align our teams and think about how we scale this technology to really be enterprise-grade for our customers.
So, you know, those are a few of some of the challenges that we encountered as we grew this business.
Great! Well, thank you so much for sharing those insights and experiences.
So today, I've been joined by Brad Begolia, who is the CEO of Symbi Robotics. Thank you for joining, Brad.
Scott: Thank you so much for your time today.
All right, if you've enjoyed this episode, take a moment to rate our show on any podcast platform that you listen to. Scroll down to the bottom and push 5 stars. It's that easy.
And as always, thanks for listening.
[Music]