Engineering the Future of Scooter-Sharing — with Deb Gangopadhyay, CTO of Beam

Today we have the pleasure of interviewing Deb Gangopadhyay, Cofounder and CTO at Beam, a scooter-sharing start-up that launched in fall 2018, raising $6.4m from top VCs including Sequoia, Founder’s Fund, ZhenFund, Cherubic Ventures, etc. Deb shares his perspective on a number of topics, including: his journey from Yale to Silicon Valley to Singapore, last mile vehicle types of the future, product development and scooter customization, how Beam collaborates with manufacturers on vehicle R&D, hard engineering problems and opportunities, and more.

*Article originally posted via The Harbinger China

Before we dive into some specific questions, I’d like to learn a little bit more about your background. Your journey starts at Yale, and you initially moved out to Silicon Valley to be a product manager. Now you’re out in Southeast Asia, working on Beam. Tell us a bit more about your story.

Deb: Sure. I actually went to Yale with my co-founder Alan, and we graduated the same year. I initially studied biochemistry and biophysics and was thinking about becoming a doctor. But I always had an interest in starting a company and the business world… so after Yale, instead of becoming a doctor, I worked in finance on a number of non-user facing engineering problems. A lot of quantitative work around economic models and simulations. It was a great learning experience, but I quickly learned that I wanted to build things for people.

That’s what led me to Silicon Valley. I joined a company called Pocket Gems, where I was building mobile games, some of the top grossing, most popular mobile games in the world. And that’s where I actually learned how to build things for people in the highly competitive mobile apps space. Eventually, I realized that I wanted to start working on some of my own stuff, so I built a couple of products for eCommerce, B2B SaaS products for marketing automation that did fairly well and afforded me a bit of flexibility.

In early 2018, I traveled to a few places in Asia, mostly for fun. I was visiting my cofounder Alan in Singapore, but also visited Vietnam and Thailand. I also explored Europe for a few months. When I finally returned to San Francisco in June, after five months of travelling… I got a call from Alan, who tells me: “hey, we should start a scooter-sharing company,” and at the time I said no. I’d been travelling for so long that I simply wanted to unpack and settle back into Bay Area life.

But as we looked into the financials and I delved into the technology, I got really excited to look into a product that was very IoT-driven. A lot of the technical challenge was, how do I take this vehicle and properly integrate it into the software and into the consumer app? That was a type of challenge that I’d never looked into before. Getting a chance to work with Alan was also super exciting because we were close friends and had established a lot of mutual trust. And fortunately for us we had very complementary skill sets to start a company together.

So within a few weeks, I was convinced, and we were flying to China to look at some scooter manufacturers and exploring ideas. By the end of those few weeks, and with some help from another college friend Adam Bao who was familiar with the supply side and VC environment, we had confirmed access to some scooters, obtained several funding commitments, and so I actually went straight from that trip to Singapore, and then it’s a whirlwind ever since.

So you’ve been traveling a lot over the past couple of months, all around the world. Before we dive into Beam, the company you started with Alan, tell us what it’s like in Singapore and moving to Southeast Asia after stints in Silicon Valley and the East Coast?

Deb: It was a fairly … I would say sudden move, in the sense that I … even while I was here for one or two months in Singapore, I still hadn’t really admitted to myself that I had moved here. It went from exploring Beam and figuring out if it’s going to work to “wow, I’m actually living here”. And I think that if I hadn’t traveled earlier this year, if I hadn’t gone through Asia just for fun, or through Europe, I don’t know if I would have been comfortable just making the move, packing up my bags, going to China, and then moving to Singapore to explore this idea with Alan. All the travel helped me become more open-minded, and it’s made being in Singapore that much smoother.

Tell us a little bit more about Beam and what you’re trying to accomplish.

Deb: On a surface level, it’s easy to look at Beam as a scooter-sharing company. And perhaps that’s what Alan and I thought it was when we first looked at it. But over time it’s grown into something a little bit broader. We believe that Beam and the platform we’re building is actually going to be the best method for people to get from point A to point B. Today it might be on e-scooters, shared e-scooters, but in the long run, the concept that we really believe in is self-serve shared mobility.

What that means is that people are actually sharing vehicles. So, we believe that in the future, almost all transportation is going to be done on a shared vehicle, and that car ownership for example, or the ownership of vehicles is going to be a luxury. So, just like riding a horse or owning a horse today is a luxury, and people do it for fun rather than for transportation, we believe that people in the future are going to own Ferraris and Bentleys for fun, but most transportation is going to be done on shared vehicles.

There’s a couple reasons for why we believe this to be true. So, if you look at the last revolution in transportation, we think of that as ride-sharing, right? So, if you go to most cities where you have taxis and ride-sharing, ride-sharing actually has greater market share already. The reason we think self-serve shared mobility is the next revolution in transportation, and already why it’s growing faster than ride-sharing is because it has kind of two distinct advantages.

One is that you don’t have a driver, which means that you don’t need to pay for someone to drive you around. It’s always going to be cheaper. Secondly, if you have enough density, which in many cities with bike-sharing, you had the density where you could walk outside of any building and be at a vehicle within 10 to 15 seconds. What that means is you’ve eliminated something rise-sharing can never really eliminate, which is ETA.

So now you have a faster, more convenient, and cheaper way of getting around, and that’s why we believe that self-serve shared mobility is going to be the next revolution in transport. And it’s already starting to grow faster than ride-sharing, starting to show those signs.

That concept makes sense. Help us understand this… the initial vehicle that you’re launching is the electric scooter. Are there going to be other types of vehicles?

Deb: Good question. So today, the electric scooter is a really good vehicle to start with, and there’s a couple of reasons. If you think about riding a bike, if you remember back to when you first learned how to ride a bike, it was really hard right? If you had never learned how to ride a bike then, and you were to try to learn how to ride a bike for bike-sharing, you probably wouldn’t be able to do it. You’d give up!

But for this L-shaped electric scooter to suddenly come out of nowhere and grow in popularity, there are a few reasons why. It’s very low to the ground, and that makes it easy to learn how to use very quickly. So if you’ve never ridden one before, or even seen one before, you can learn how to use it fairly safely in 10 to 15 seconds, and it’s more convenient than a bike, because it has a motor. So for those two reasons, that’s the reason why we’re picking this as the vehicle to start with.

But that doesn’t mean that it’s the vehicle that is going to be the main one that we continue to use in the future. What we believe is that the right vehicle depends on the cities that we’re in. Also, it won’t be a single vehicle. I think it’ll be a mix of vehicles.

So, if the roads are less developed … so if we go to countries like India, or Indonesia, which are huge markets, this L-shaped electric scooter is not going to work well there because the wheels are too small, and the vehicle is so small that in such dense areas, it’s unsafe to ride something so small. So something more like an e-moped or e-motorcycle is probably a much better fit for those markets.

Just to explore this concept a little bit further, you mentioned different types of vehicles for

different markets, and it seems that the scooter is currently primarily used in areas where the terrain’s a little bit more flat or there’s not as much snow or rain or hail. Is that characterization accurate? How do you look at other markets where the conditions are harsher?

Deb: I definitely think that any kind of vehicle where you’re exposed to the outdoors, so meaning that there is no cover, is definitely going to receive a lot less usage when there’s like snow and rain, or any kind of climate where you need cover. Whenever it is snowing or raining or just really cold, the usage of these vehicles will also drop significantly. You’re already starting to see that with e-scooters in the U.S. in some cities. So long run, there will be vehicles that are still self-service, that will still protect you from cold weather and rain and snow.

What could these vehicle look like? Are these round electric pods like you see in the Jurassic Park movies?

Deb: I think a lot of these vehicles are still a bit conceptual. They’re starting to come out a little bit with prototypes, some of our suppliers are also working on this. If you think about a car, one car moving a single person is extremely inefficient. A car might weigh 2,000 kilograms, and a person might weigh 50 to 100 kgs, and that’s an extremely inefficient ratio. But if you think about an e-scooter, maybe it weighs 15 kgs, right? That’s a much more efficient ratio for energy consumption. So a pod will hopefully be much smaller than a car, and weigh much less than a car. It’ll be likely electric, and it’s going to be much more efficient in terms of moving a single person around than a car, and it’s going to also protect from weather much better than an electric scooter.

In terms of these future vehicle concepts, would these be products that you guys build in-house? Would those be manufactured by other suppliers, and you would just be acquiring them and then using them for sharing purposes? How do you look at that?

Deb: Right now, I think the bleeding edge of the development of some of these products is actually done still by the manufacturers. The reason this is true is because they’ve been building these products for a long time. They were building them long before they became popular. They were actually building this L-shaped electric scooter for many years, and then the sharing market suddenly came about.

At the same time, these platform companies that are coming out for electric vehicle sharing… they’re actually starting to build a lot of in-house capabilities for R&D. I think that the bleeding edge of this technology will probably for the short-term be continued to be led by the manufacturers and the OEMs, but not too long from now, companies like ours will be driving a lot of the R&D and creation of the types of vehicles that are going to be on the platforms.

We’re talking about the future here… but today could you tell us a little bit more about your current product? So the Beam scooter, the hardware itself, any apps, IoT or systems?

Deb: Yes. Well today we’re pretty fortunate that we were able to secure some supply from Segway-Ninebot. They have the leading scooter, we believe, on the market for scooter sharing. There are numerous good new manufacturers that are up and coming, building quality scooters, and in fact building custom scooters for the sharing market. But the advantage Ninebot has is that their scooter has been the most used in the sharing market, compared to any other electric scooter that’s out there. What that means is that they have collected a lot of data on improvements to make to optimize the vehicle for sharing.

For example, they’ve focused on things like in increasing battery capacity, increasing durability, and building in anti-theft features. This can be very simple, from making the screws anti-theft screws, to things more like battery cages and things that protect the batteries from being stolen as well.

And how much influence do you guys have over their product development?

Deb: It’s collaborative. The manufacturers will have a R&D department that is more experienced in customizing the components, compared to us at such an early stage. So, the best thing that we can do right now is to provide them with data about what components are breaking the fastest, and most often. It comes down to communicating across what our biggest needs are in terms of improving the vehicle for sharing. It could be that our biggest need is to improve on the anti-theft. It could be that our biggest need is to actually improve the battery capacity. This also varies from market to market.

Right. So it seems that the manufacturers are starting to customize their scooters, or develop some of the improvements that you guys request. How much engineering work is there on your side for the Beam team?

Deb: Today much of the work is integrating the vehicles and making sure that the basic locking, unlocking mechanism works very efficiently in terms of the amount of time it takes, and it’s highly effective. So, our number one priority is actually making sure everyone … whenever someone tries to start a trip and actually unlock a vehicle for a ride, and whenever someone ends a trip and locks a vehicle for a ride, that it works 100% of the time.

There’s so many interesting factors that go into affecting our ability to do that. You know, the actual internet connection directly to the scooter, or to the phone, it can vary depending on where you are. So, even solving that very basic problem is somewhat tricky, and that’s actually probably one of our biggest software focuses this early at … such an early stage sharing company.

Can you tell us more about the process by which your engineers work with the manufacturers?

Deb: The way I like to think about it is that it comes in three phases. Phase one is, we are going to be providing them very pure data. So let’s say the tires are popping. We’ll prioritize this and share back … how long on average it takes for a tire to pop, and what percentage of our scooters are affected by this? Then it could be the folding component. We might provide data on how frequently that component becomes loose, and what percentage of our scooters are being affected. For every piece that can actually impact the usability, there will be some data, some very quantitative data that we can provide to the manufacturer. That’s phase one, and that can help drive some of their R&D to improve the vehicle for sharing.

Then there’s phase two. We’re still an early stage company, but we’re already starting to look into how we can build some of the best hardware R&D capabilities in-house. So we’ll actually start looking at the components of the vehicle and the overall design, and start to customize it ourselves from an electrical engineering perspective, and in that phase two, we might actually start sending designs to the OEMs to actually customize and improve the specific components in ways that we think are going to best improve it for the sharing market.

Then there’s phase three, where we’ll likely work with the OEMs, but perhaps we’re actually customizing most of the vehicle. We’d provide a very custom design for the vehicle and decide who the suppliers are for the parts, and how those parts actually come together in the best way.

In terms of these spaces, and in terms of this engineering work, how do you compare to some of the other existing players in the market? Some of the names, such as Bird or Lime, are quite well-known by a global audience, because these companies are relatively big and raised quite a bit of money. Are they doing something similar? What phase might they be on in the three steps that you mentioned?

Deb: I don’t have all the details on what they are building. But it’ll come down to being a mix of trying to build some kind of in-house hardware or R&D capability. I think that some companies may actually focus more on that, while others may focus on trying to find more supply from a variety of OEMs. So I think it’s a choice between kind of building a custom design and really putting a lot of belief behind it, and then only finding OEMs that will produce the design that you’ve made, or it’s about kind of finding a variety of OEMs that you think are good and can produce good designs on their own.

For us, we’re pursuing both approaches, but currently relying more on the OEMs in terms of R&D and design. But we also want to start pursuing the route of doing our own R&D and design sooner rather than later.

It makes sense that the various scooter companies have different approaches to engineering and collaboration with the suppliers and manufacturers. So you mentioned earlier about the importance of the locking and unlocking mechanism. But from an engineering perspective, there’s a number of other things that are of interest as well. I’m no expert on scooters, but it might be things like the battery, perhaps durability, or even geo-location. Could you just talk us through some of the other key points to understand when it comes to engineering for these vehicles?

Deb: One thing that I think is very unique to this business is that you have access to a very large number of IoT, GPS, and cellular-connected devices that are scattered across a city. What that means is that you can start using these devices … let’s say if they have Bluetooth, to actually start verifying the locations of each other through things like Bluetooth triangulation, and that actually gives you, I think, an ability to improve the geo-location accuracy, which has a huge impact not just on the rider side, but operationally. You could go from having an operations person … you can help them go from finding two vehicles an hour to three vehicles an hour, which makes a massive difference in operational efficiency and cost.

There’s a lot of room for innovation in this space, and some of the companies have already started looking into how to run the operations of charging the vehicles very effectively and efficiently. What the companies in the U.S. found out very quickly is that to scale this business fast, you must have a gig economy. So, you must allow anyone to sign up and get paid for charging the vehicles. The reason you need to do that is because your vehicles are distributed throughout an area and you need to collect them every one to two days for charging, so if your workforce is also distributed throughout an area, you have a pretty large advantage and can start scaling and launching vehicles very quickly.

Once you start building this gig economy, there are so many interesting optimizations that you can do. It kind of goes through the whole process of someone finding these vehicles, collecting them, taking them home, charging them, and dropping them off. And you want to be as granular as you can in terms of thinking about how to optimize the system.

So, if a vehicle hasn’t been ridden in a while, you might want to pay more because it’s in an area of low demand. Or it might be behind a tree or hard to find, or maybe it’s broken. You also want to include incentives to get the scooter charged faster. And when they’re dropping it off, you may want to look at historical demand data, and pay more to the charger when they drop it off in areas of high demand versus areas of low demand. So there’s a lot of interesting innovation that can be done in terms of optimizing how this gig economy is run.

You definitely see that in the U.S. and it’s really cool how a gig economy could be built on the charging piece of scooter sharing. A lot of value seems to be created there. I was also curious though, because you guys are looking at multiple markets. How might your start up model vary by market?

Deb: Firstly, as mentioned earlier, the vehicle mix may vary by country. The actual mix between whether we have in-house people charging the vehicles versus gig economy might vary as well. The reason there is because if the labor cost is quite low, it might actually start to make sense for us to employ people for longer hours to actually go collect multiple vehicles. It may actually cost less to do that than to have a gig economy where we’re incentivizing random people to go and pick up the vehicles.

Makes sense. You see that in China quite a bit as well, whether it’s Mobike or ofo, oftentimes there’s these trucks that come by with an entire team of people just picking up bikes and reallocating them, so that does happen. Deb, I know today you’re limited on time, so appreciate your time for talking through some of the engineering issues that are top-of-mind for you during this period. Perhaps next time we can have you on again to discuss other things you wanted to cover and share with our audience.

Deb: Sounds good. Thanks for having me.