I’m excited to (finally) share a rather big secret. A secret that actually pertains to nearly half the people who work at Starsky Robotics: In addition to developing self-driving trucks, we’ve been operating a full-fledged trucking business for nearly two years.
We’ve built a 36-truck regular trucking company to serve as a beachhead for our autonomous trucks. We’ve made millions in revenue, gained operational competencies worthy of industry credibility, secured key industry partnerships, and removed many of the key operational blockers of autonomous development.
What’s more, we’ve been able to prove that trucking, an industry many assume to be completely adverse to the slightest change in technology, is more than comfortable with the idea of unmanned self-driving trucks hauling their freight every day.
Getting the Driver Out of the Cab
As I’ve talked about before, trucking’s problem isn’t labor cost, it’s supply. While drivers do account for a third of the cost of operating the truck, fleets are desperate to hire all the drivers they can. As a commoditized service, trucking faces downward pressure on rates. This translates into a cap on how much carriers can pay their drivers.
That cap, even at $50–60k/yr, isn’t enough to get someone to spend a month at a time in a truck. As a result, the American supply chain has 60,000 fewer truck drivers than it needs. The only way to solve this problem is to get a person out of a cab.
Nobody Knows How to Manage a Truck, Until They Do
In the fall of 2016, I bought our first truck to test our self-driving technology on it. I had never bought a truck before, and at $50,000 it was the biggest purchase I’d ever made. As it turned out, I probably should have spent more money on it.
We installed our system on it with plans to take it to The Thunderhill Expressway for testing. By the time it crossed the Bay Bridge, smoke was billowing out of the engine. During the next two weeks, our hard-to-hire autonomous engineers sat on their hands while I called dozens of diesel mechanics (of whom there’s also a shortage) and finally got the truck running again.
The challenges didn’t end there. We had a major goal to hit by the end of the year, and our truck refused to cooperate. It would break down. Our temporary insurance was nearly up. Our driver was getting restless because he wasn’t regularly driving. The only part that was kind of working was that it was easy to find freight to haul, but we couldn’t get paid to do so because we weren’t registered as a trucking company. And then, right before we tried to do an autonomous freight run on the highway, I realized that the title hadn’t been transferred to Starsky, it had been transferred to me personally. And no one was officiating title changes on Christmas Eve.
I was going to have to tell our investors we missed our goal not because of our engineers, or because of freak weather patterns, but because we didn’t know how to manage trucks. Given that challenge, and our penchant for building a “real company” (as opposed to an R&D lab), we decided to start a trucking business to develop our operational expertise.
You Can’t Spell “Autonomous-Trucking” without “Trucking”
I’ve often said that there are infinite reasons why we need our regular trucking fleet. But, a few of them are as follows:
Trucks: Trucks are different from cars. They’re made in far fewer numbers and are much less precisely manufactured. A higher portion of trucks need much more maintenance than “average.” Even “average” trucks need to go to the shop every 60 days, which is operationally complex when you’re traveling the country. Two trucks of the same make and model might be different enough that you want to return one. With our regular fleet, we can test a truck first and make sure it works properly before we actually put our self-driving system on it.
Lanes: We also use our trucking company to identify freight lanes that will be perfect for autonomous vehicles in the future. Thanks to our extensive fleet, we have more autonomous-friendly lanes to choose from than we have self-driving trucks to deploy. This allows us to test new routes with manually-operated trucks, and keep our driverless trucks for carefully selected routes that we can survey and authorize.
Drivers: It turns out that eating your own dog food is really good for you. While building our fleet, we’ve experienced first-hand just how hard it is to scale a fleet of trucking drivers. Doing so while still trying to make a margin off of trucking, is a harsh lesson in why every worker can’t be treated with Silicon Valley largess. Given how hard it is to hire drivers and an industry turnover rate north of 100%, we could easily be in a situation where expensive autonomous trucks can’t be tested, because we don’t have enough qualified drivers.
Now, we use our growing trucking business as a captive source of safe, experienced and well-trained drivers to staff our safety and remote driver team. We’re able to get to know our drivers over the course of a few months to see if they have what it takes to advance into being a safety driver and eventually into a teleop driver.
We learned a lot in those months. We’ve had drivers quit because they can’t use our electronic log to track their driving hours. We’ve had drivers make a wrong turn and have had a low-hanging branch rip a hole in the side of a trailer. Last year, we even had a new driver take an off-ramp in Florida too fast and roll the truck onto its side. It’s not only fortunate no one got hurt in that incident, but that we were able to catch that carelessness and terminate the contract of the once-promising driver before we put them behind the wheel of a self-driving truck.
The only way we can scale our operations on the autonomous side is by continuing to scale our regular trucking operations 6 months in advance. By early 2020, we aim to have 25 autonomous trucks, which we think is possible only if we have 100 regular trucks. It simply wouldn’t be possible to find 25 highly-skilled safety and remote drivers without having a broader fleet of regular truck drivers from which to choose.
Getting Industry Support
Given that the hardest thing in the trucking industry is to find capacity for your loads, we’ve been able to partner with 15+ brokers who are regularly giving us freight to haul. Many even (wrongly) assume we’re already hauling it with unmanned trucks, and give us the green light to do so.
We’ve teamed up with some of the most recognizable companies in the industry, including Schneider Logistics. Like many successful, large logistics companies, Schneider has a fine-tuned demand-generating organization that could always use more capacity. We provide capacity where they need it so Schneider can service more of their customers’ freight. We’re now regularly hauling over 100 loads per month for their customers using our regular trucks.
What we’ve been excited to find is that, by using brokers as our distribution channel, they’ve been able to absorb our capacity as quickly as we can grow it — which is why our revenue increased by 50% month-over-month in Q1, as we were able to grow our capacity.
To be clear, there is so much demand in trucking, that if you have the right partners and increase the size of your fleet, you grow revenue. Right now, to grow our fleet, we need to convince notoriously fickle drivers to join us. In the future, though, we’ll be able to grow our fleet just by ramping up manufacturing. Which means our revenue will be limited only by our ability to make our product.
Trucks are complex engineered systems, though, and given how widely they travel, making sure they’re in tip-top shape is a challenge that sinks even great trucking companies. That’s why we’ve partnered with Penske Truck Leasing and Transport Enterprise Leasing (TEL). Through them, we’ve gotten access to some of the newest and best equipment in the industry.
They’ve been flexible and forward-thinking in their equipment allocations, allowing Starsky the ability to grow. Also, we work closely with Penske to automate the processes of getting a new truck up and running: making sure it’s properly licensed, permitted and taxed. Laying foundations with quality equipment providers allows us to establish a real competitive edge as we grow our autonomous fleet at an accelerated pace.
Bringing On The Right People
By the time we were raising our Series A in Fall 2017, we were regularly making around $50,000 a month in trucking revenue. It seemed like a lot, but when I started digging into the numbers, I realized that we were not on top of the unit economics. It became clear that we needed to bring on board an experienced head of trucking.
That’s why I’m so excited to (finally) announce that we’ve brought on an industry veteran, Paul Schlegel, to lead, optimize, and expand our trucking operations.
Paul joined Starsky after spending over 32 years in the transportation industry with companies such as Schneider National, Stevens Transport and Roadrunner Transportation Systems. Paul brings us a wealth of transportation and logistics knowledge that will ensure the growth, efficiency and the ultimate success of Starsky’s Trucking arm.
Paul was able to quickly jump into the weeds and fix our unit economics to make it possible for us to earn a profit from our trucking operations. His first priority has been to develop a trucking business that enabled autonomous trucks, but still worked well in their absence.
Paul’s been able to work closely with our existing team, and built a team of future-minded industry veterans in Dallas. They’re doing many of the everyday trucking things that don’t scale; while keeping a list of things that, with technology, just might.
The Future of Trucking
‘Wait, so is Starsky a trucking company or a robotics company?’
If you’re asking yourself that, you’re in good company. Industry thought leaders, elected officials, VCs, and many others have wondered just that. I’d say it’s not unlike asking if AWS is a web hosting business. Kind of. But kind of not.
We don’t want to build a traditional trucking company with thousands of sales people, hundreds of dedicated customer service representatives and mechanics shops across the country. As a former sales guy, you can believe me when I say I don’t want to hire any sales people.
I want our sales force to be an API, through which great brokers, like Schneider Logistics, can order their customers’ shipments to be moved. I want our customer service team to be an API, through which those brokers can see a real-time GPS location of their load and share that information with their customers. I want our mechanic to be Penske, who already has competency in managing the purchasing and maintenance of hundreds of thousands of vehicles.
With strong support from industry leaders, a driver team that’s making it happen every day, and great professionals on board; it seems to me like we’re on our way.
Keepin’ on Truckin’