Ryan Petersen is currently the CEO of Flexport. Flexport builds software that integrates all the different shipping vendor systems you might run into as you try to get a product from a factory in China to a consumer in Idaho: rail, sea, or truck.
It’s a good time to talk to Ryan since it feels like we’ve been in the middle of a supply chain crisis since the pandemic began in 2020. We’ve talked about the supply chain and inventory management on Decoder with a lot of our guests, and sorting out how to get products made and delivered on time is a pretty universal problem.
But, we haven’t really talked about how products get from one place to another around the world. That’s one of the problems Ryan himself has solved. So I wanted to talk to Ryan, figure out what Flexport’s role in all this is, what his bigger supply chain solutions would be, and why he’s leaving his job as CEO to be executive chairman and handing the reins to Dave Clark, who used to work at Amazon.
We talk about these changes, Ryan’s big ambitions, and just what the hell is wrong with the supply chain. I gotta say, Ryan keeps it pretty real — he might be one of the most candid CEOs we’ve ever talked to.
Okay, Ryan Petersen, CEO of Flexport. Here we go.
Ryan Petersen is the founder, current CEO, and soon-to-be the executive chairman of Flexport. Welcome to Decoder.
Thanks for having me.
There is a lot to talk about here. There is a supply chain crisis and everything seems to be in a shortage, and you make software that can help with that. You also recently made a big decision to bring in a new CEO and step up as executive chairman. This is a podcast about decisions so I will want to get into that, but let’s start at the beginning. What does Flexport do?
Flexport is a technology platform for global logistics. We make it easier to ship anything anywhere, all over the world. That is the vision. We don’t actually ship anything anywhere and we don’t go to every country. We mostly ship containerized freight, so ocean freight, air freight, et cetera. What is freight? Anything that is too big to fit in a FedEx or UPS truck. It can be consumer goods, electronics, apparel, all kinds of stuff. We are getting really diverse.
The core problem is that global trade is one of the most important things in the world; it is the circulatory system for the world economy. Forty-seven percent of GDP is now international trade, yet it is still a nightmare to do. We build technology to make this easier for companies to manage that process and give them visibility. “Where is my stuff? When is it going to arrive? How much is it going to cost? How do we lower the costs of those transactions and make the world a better place?”
You started out by saying, “We don’t ship anything,” and then you said you shipped a bunch of things.
No, we don’t ship everything.
Flexport primarily makes software, right? You’re not in the business of actually moving containers from place to place.
It depends how you define things. We make software that connects companies that need to ship stuff with companies that have assets to move those things. You actually contract to ship things through Flexport. We have rates for our ocean freight to ship anything between 112 countries. Air freight is the same. We can ship all over the world and make that easy for you. It’s a hard problem from a software perspective because we have to go to the supply side and solve their needs as well. Supply side means trucking companies, ocean carriers, airlines, warehouses, and customs brokers. We connect this whole network of those companies with over 10,000 importers and exporters that have cargo to move.
I am assuming you don’t have a lot of single guys with ships, but just assume I’m some guy with a ship. Can I sign up for Flexport and say, “I’m going to China and I have room in my ship, so fill it up”?
You are right that there are no guys with ships.
I have a dream where I own one of those giant container ships and just spend my days on the sea.
Listen to Decoder, a show hosted by The Verge’s Nilay Patel about big ideas — and other problems. Subscribe here!
It’s a cool idea. The first article ever written about Flexport was a Bloomberg article that called us Uber for ocean freight and it’s the model that you just said. I think it confused people for many, many years about what we do, so I want to be careful. Today, there are only 11 or so ocean carriers. These are the people that own container ships that represent 95 percent of global trade on containerized freight. We have contracts and partnerships with all of them. We have integrations to get rates and service schedules in order to book freight with them.
Let’s say you have a container you want to ship. The ocean carrier goes from port to port, but your container is going from a factory to a door. Someone has to coordinate the trucking and the customs clearance as well. That is where Flexport comes in.
We also have partnerships with hundreds of trucking companies that run our software to assign loads to their drivers, to dispatch, and to deliver. We clear customs in over 100 countries to get goods in and out. Stitching it all together is the art of freight forwarding. I often joke it should be called freight email forwarding, because usually people aren’t building software for the supply side and connecting them in; they are forwarding emails, PDFs, Excel attachments, et cetera. We are trying to modernize things so the data can flow between these companies. The fundamental problem is that all these companies run different IT systems. How do I actually connect them together? It’s a really challenging problem.
If you have some experience in this field, you might say, “Okay. To start my direct-to-consumer business, I’m going to get a factory in China to make hair dryers, and then they are going to send them to me.” Suddenly, you have to deal with how they are going to get to you and how they are going to get to customers. That is very challenging. If you have an abstracted view of that, you might then say, “Whoa, we should just build some software to standardize a bunch of these procedures and automate it. With a click of a mouse, let me make this thing happen.” However, you have 11 different companies that are huge and probably don’t want to have lower friction from switching between them. You have whatever number of trucking companies, customs protocols, and offices in however many countries. How on Earth did you get them all to play ball?
We don’t have everybody. Honestly, if I had known everything then that I know now, I wouldn’t have done it. It would have sounded impossible, as in, “You could never pull this off!” There is a naïve, yet helpful optimism for a new founder charging into an industry they don’t know much about. You can say, “I’m going to solve all these problems,” and actually do more than you think you can. It almost had to be an outsider to the industry to tackle this, because an insider would have understood everything and said, “It’s not possible.”
What was the first win? When did you know it was going to work?
Before I even launched. I came at this from a customer perspective that was almost exactly the scenario you were describing. “I have these hair dryers,” but in my case it was motorcycles, mostly from Taiwan and Mainland China, and a little bit from Brazil. It was my older brother’s company. I was working for him. We were buying and selling them and we hated our freight forwarder. George Bernard Shaw says, “All professions are conspiracies against the laity.” Freight forwarding is the poster child for that. They have all these code words in Viking English. It’s like, “What are you talking about? How much is it going to cost? When is it going to arrive? What documents do I need? What do these documents mean?”
I said, “Hey, we should just build software to do this. We can tell the person what the documents are. We can have a wizard that is basically a TurboTax for importing.” Only after I started did I realize, “Oh man, now I have to go contract freight from these ocean carriers and find truckers to deliver it.” I did not understand all the problems that were actually there.
Saudi Aramco signed up for my website. I thought, “Even the biggest company on Earth wants to use us.”
Before I ever launched, I had a website that offered the service, to see if people would want it if it existed. One day, Saudi Aramco signed up for my website. It’s the Saudi national oil company, and one of the biggest companies on Earth. Just like your example of trying to import hair dryers, I thought, “Wait a minute. I thought I was attracting Amazon merchants! If we could do this service, we would make a lot of money. Even the biggest company on Earth wants to use us.”
We don’t ship oil, even to this day, so Saudi Aramco is not a customer — but the possibility is all there. It’s like, “Okay, now we just have to learn how to do the business.” That was in 2009, so it has been the better part of a decade now. That is 13, 14 years of constant iteration and improvement to be able to do the job, and I still don’t think we have it all solved.
Imagine starting a software company, setting up a landing page to collect email addresses, and seeing something like @saudiaramco.com. That’s one thing. Even building this software is a lower lift than making this company. Did you have to go knock on doors of the freight companies and say, “Hey, I need you to let me use your backend IT systems”? Or were you saying, “We will just figure out how to read the PDFs and take from it there”?
A lot of chicken-and-egg problems had to be solved. They don’t want your business until you’re big, and they are not going to give you any good freight rates until you’re big. But of course, I can’t get big if I don’t have a provider and good freight rates. It’s just chicken-and-egg problems in every direction.
The way we solved it was by getting customers who really wanted software to manage their supply chain, and by building that software. I won’t reveal customer names here since I don’t have permission to, but we would tell them, “Look, this really is beautiful. We have all the data you are going to want here, but you have to ship with us for it to work.” They would say, “Yeah, if you can get this price then we will ship with you.”
We have to have these conversations before we can go to the supply side and say, “Hey, ocean carrier/freight forwarder, if you can get this price then I will bring you the world’s largest company that is interested.” A big company would then come on board and say, “They want this software, they will ship with us, so we just have to get this price.” And the price wasn’t crazy either; it was half of what we were paying at the time. We also had to get people because we had a lot of manual stuff. Humans doing things. There are still humans doing a lot at Flexport now, but we just chip away and automate those interactions. We integrate systems and find ways to streamline things. There are hard problems to solve in every direction.
I think the audience for Decoder is every business school student in America. This is the real meat of the problem. What did your pitch look like to both of those entities? You go to someone who wants to ship something and say, “Look how beautiful this software is. Don’t you want to use it?” They would say, “Get me the rate and we will use it.” You would then go back to the other side and say, “Give me this rate and I will bring you this customer.” Can’t they just get the customer anyway?
At its simplest, the operation is going back and forth. You are basically trying to get a market to form, and market-making is hard. Most times, you need liquidity in the market in real time. You need to match supply and demand right now on that eBay sale. You can’t wait three months to find the seller or find the buyer. In this case, it’s asynchronous, so I could get the buyer and say, “Hey, look at this software. Let’s co-develop this software. Let’s sit at a whiteboard and be like, ‘You want a dashboard that lights up when this shipment is this late, et cetera.’ Okay, cool. Let’s design that.” Now the software is there.
It’s if/then statements. “IF this software works, and IF you hit this price, THEN I will buy.” Okay, cool. Now I have to go find a provider that would do that. They couldn’t get the freight because the buyer sees them as a commodity on some level, but doesn’t see the software as a commodity. They want the software and are willing to have different providers if the software and the data is there.
One lesson here is that you do have to take the supply side. They don’t want to be a commodity. They don’t wanna be in a race to the bottom, like Kayak, for who is the cheapest freight provider. You have to really spend time with the supply side to understand their needs. In that case, it’s like, “Hey, they want to be able to differentiate. They want to be able to charge more for better services. Okay, cool, let’s create that product.” We have created a lot of products for the supply side, where they have premium, last containers loaded on a ship or the first off, and get paid a lot more for that. Prioritizing containers, fast lane type of things. That helps them make more money.
I think one of the core lessons of building a platform is you have to look in every direction, and make sure everybody is winning. It’s really hard to do. But that is the definition of winning, right? The definition of success is like, “Everybody wins; nobody has to lose.” It’s a good framework for evaluating if a company is going to succeed at the end of the day. Is one of the stakeholders really losing? Is everybody benefiting from this? You can apply that to a lot of companies to figure out, “Does this have staying power or do they need to really invest in their compliance, their government relations, their neighborhood?”
It’s like Airbnb. Everybody wins in Airbnb except for the neighbors. It’s like, “Okay, what are we doing about that?” They spend a lot of time saying, “Hey, no parties allowed. No this or that,” to try to solve that kind of flaw in the model.
Stepping back again, this is extremely reductive. When you say that ocean carriers want to differentiate, it sounds like you are describing classes of airline service and you want to expose them in your software. That way I can pick Ocean Liner Comfort Plus and get that for slightly more money. They want to compete on that level instead of just getting from here to there.
It’s really the transit time and price. Each one of those has a standard deviation, which is your reliability — that is, your ability to hit that price without surprising them later that there is a different price or your ability to hit that transit time. You might have a really fast transit time, but half the time it’s super slow. That’s not very reliable. They want to be able to differentiate on that. They do not want to be in a race to the bottom on price, which is what has happened to the airline industry basically. You have Kayak and people tend to just choose the cheapest flight.
It’s a big split, right? There is a race to the bottom for half the plane, and then skyrocketing costs at the top of the plane. The airlines will happily tell you, “This is what everybody wants,” because half the people get cheap flights and the other half will pay any number of dollars for a larger seat.
We actually do a lot of air freight. It is interesting to talk to these companies and see that they are actually really sophisticated on yield management for the passenger, which is all of the stuff you are describing. In the belly, they are 20 years behind. They are getting there. A lot of the same people have moved from the passenger side of the business over to the cargo side. They have started optimizing, putting in yield management philosophies and processes. It’s not there yet.
The same is true for software. The airline industry was one of the first in the world to adopt computers. Way back in the 1950s, way before the internet, they started using computers to do ticketing and stuff. By the ‘80s, still before the browser, you had these online dial-in systems for travel agents, and the databases couldn’t keep up. That is a big deal when you’re selling a ticket. If you sell it twice, it creates huge problems for them. It’s Google Flights now, but ITA came out and built this — just using a relational database — built a real-time database for them to be able to present ticketing options to people. That was a huge win way back in the ‘80s on the airline ticketing side.
That never happened in the ocean carrier world and never happened in the air cargo world. Maybe it is starting to happen right now and Flexport is a part of that story. These companies are modernizing pretty fast. Though in my view, it is 20 years after it could have happened with the rise of the internet.
Is that because if you blow it with an airline customer, then you have an angry airline customer? If you blow it with a shipping container, then you just have a shipping container?
Certainly the cargo is less angry when you leave it on the side of the runway for two days. It doesn’t speak up. That makes it harder too. The passengers can walk themselves through the terminal and get their own passport and visa. With cargo, you have to do it all. Everything has to be managed door to door, but the cargo doesn’t get as upset when it is delayed. You have a lot more optionality for how you route things. The customer cares a lot more about stops, for example.
That is actually one of the things I really want to push on here. A lot of the process is still manual, right? Someone has to pilot a ship across the ocean. Someone has to drive the truck to the port, load the container, unload the container, and get the hair dryers out. How do you actually make sure any of this stuff is where your software thinks it is?
The physical process is manual. What you’re trying to do is get it so that the data moves seamlessly without humans having to do things. Even then, there are parts that are manual where people are like, “How do I get data to and from each actor out there in the field that needs to do something? How do I eliminate layers between that operator, who really knows what’s happening, to the freight and their dispatcher or manager? Then ultimately our company and our customer?” There are all these layers to communication.
We do this with software. The simplest example to explain is our trucking software, where we have a dispatcher app. It is a web application that is run by over 140 trucking companies in the United States. Each of them has between 10 and 200 trucks — I think the average is around 100 trucks. These are doing port pickups and airport pickups for us. It is a dispatcher app and their drivers get a mobile app. If they want, they can also get a hardware device that plugs into the truck. The mobile app is actually pretty good though. You can get GPS off of that.
The driver is then told where to go to pick up the load and what load it is. They can log data, take a picture of the container, enter the container number, and get the customer to sign the delivery receipt. That data can flow not just to their dispatcher — which is novel for them, since they haven’t had tools this good before — but also to our team, and then all the way to our customer. We can say, “Where is that truck? When is it going to arrive?”
At some level, you are trying to eliminate the need for this freight email forwarding, which was required before. In the beginning, before we built that software, we would call and email these truckers to dispatch them. “Hey, I got this load, come pick it up.” It was bad for getting real time data, like, “Where is the stuff?” It was also labor-intensive. It was costing us 42 minutes per shipment of humans having to interact with those truckers. So it’s a huge win. It is a better customer experience and saves a lot of money.
This is a transition that Flexport made? You had people calling the truckers?
“It was like building a rocket ship while it’s in flight.”
Yes. In the beginning, we didn’t have that software, we had to build it. It was like building a rocket ship while it’s in flight. First, get the customer. We are a customer-centric company. Not all the solutions are tech. If we have to do stuff with people, okay, we do it with people. Then we will come back to build tech to replace the stupid things that we are doing, like having to email truckers. We should not have to do that. The tech solution was a big bet for us. We were like, “Oh, we’re going to build software. They are going to run it to manage their fleet and they are going to use it even if it’s not a Flexport load.” These trucking companies use our software to assign all their loads to their truckers. That was a big bet we made that paid off.
How many people does Flexport have? Are you scaling up when you have people that you need to make phone calls then scaling down again? How does that work?
We tend to grow out of it. Hopefully we don’t have to scale down and go through it. It’s pretty tough on culture if you are laying people off. We have 3,000 people. We try to automate the work so that as we grow, we don’t need to hire as many people. We are making good progress on that. We think we can probably automate about 80 percent of all the work that happens coordinating a transaction in the next few years. It’s probably that every five years you automate 80 percent, but you are never really able to automate everything. That is my suspicion of how it plays out.
What is the pricing model here? Are the truck companies getting the software for free and you just hope that you can get Flexport loads in there? How are you making money?
We make money by charging customers to ship freight and then providing all the services that you need around that. It’s not just a freight delivery, but also customs clearance, insurance, and a financing business. We lend companies money to buy products. Those are the core business lines. There are lots of additional services we could layer onto that. We charge for the trucking software we do, but not very much. It is not an important revenue stream for us. It’s more so that they know we are serious about supporting it.
I hear this from every CEO who comes on Decoder. Once you have a software customer, your investment in software starts to skyrocket because your users expect a lot out of you. You have given the software to the trucking companies. You’re charging and saying you’re serious. Do you need to have customer support people for the trucking companies in case the software goes sideways?
We do. It’s an area for us to make some big strategic decisions. I would really like to see us invest and go big there, to start adding more services for these truckers. What do we need to help them run a good business? The group that builds that software should only think about the truckers to make amazing stuff. Maybe it’s cheaper trucking insurance or financing. There are a million things you could think of that a trucking company might need to lock in that supply and make them more loyal to Flexport, and to be available to move our loads when we have loads to move.
Today, we build software that helps our core customer move freight, and I would really like to see us make strategic bets there. The problem with these companies is that you just don’t have enough people. We are not cash-constrained, but having the talent to go and do all that stuff is a lot of people.
All right, you opened the door to talk about your finances. You said you are not cash-constrained. It says on my sheet here you just started to be profitable in 2021. What does the revenue look like? What does the profit look like?
Last year, we did $3.3 billion in revenue. We have a take rate on that. I think last year’s take rate was 18 percent. That is what the net revenue was. Bottom line, we made a little over $30 million in profit last year, if I have the number right. We have $1.2 billion of cash in the bank and liquid assets, so we are good on cash. We are trying to maintain and not go back. We burned a lot of money to get here. For context, we raised a total of $2.35 billion. If you have $1.2, that means we spent $1.1. There is a lot of anxiety that comes with that. It is not fun to set that much money on fire. I do not want to go back to burning money. I am really proud that we are profitable. We might burn a little bit of money this year as we invest, but I never want to go back to the world that we were in a few years ago.
Here’s the big Decoder question. You have been at this for 12, 13 years. You have been the CEO, you are the founder, and you have just turned the corner to being profitable. Depending on how this year goes, you will be profitable again. You actually just decided to stop being the CEO. You hired Dave Clark, who is from Amazon, to be the CEO, and you are going to move yourself up to executive chairman. How do you make those decisions?
This is a different kind of decision, so I am not sure I have a generic framework. I wasn’t planning to make this decision. In fact, I love being the CEO of Flexport. I think I’m pretty good at it, since I got us this far. There are some things I’m not that good at that I think directly align with what we need to do over the next decade to become one of the most important companies in the world. They are not my strong suit.
I am a creative, intuitive type. I am very good at coming up with ideas. You show me a problem, I will probably come up with a creative idea. Now, I don’t know if you should do it or not, it might be too crazy. But I do come up with creative ideas that other people have not thought of. I have a unique ability for that.
What Flexport needs right now, though, is rigorous day-to-day to get a little bit better at the stuff we did yesterday. We are going to launch new stuff to improve our velocity, but a lot of it is operational excellence. What is the transit time from China to the US? What is the standard deviation of that? What are we doing to cut that every single time to make it run a little bit better?
It’s not my forte. I’m learning it really fast. I am finding it intellectually interesting, but I’m having to learn on the job. Dave has got this like the back of his hand, and can free us up to be more innovative and focused. If we are not good in the core operations — and I think we have some real opportunities of improvement there — then we are just never going to be the world’s most important logistics company and logistics platform.
It was just opportunistic. I was starting to look for a COO who could be my right hand to do that stuff. I caught a whale when I found out Clark would be interested. I looked at the org and was like, “What should I own versus Dave?” I couldn’t really find anything. I thought I would be a lot better off with him there, so I decided to step up to exec chairman and be his business partner rather than his boss.
I will come back to the org chart, because I love an org chart question. Let’s start with the more general question though. How do you make decisions?
I personally do not love making decisions. I think that is another area that Dave has, because he actually loves to make a decision. He hates to make a decision tomorrow that he could make today. Most decisions where he’s correct are reversible so you don’t have to overthink things.
I am more consultative. I want to hear other people’s opinions, delegate, and let them make the decision. It has served us quite well. We built a culture where it is not all dependent on me since I don’t make all the decisions. It can slow us down though, because a lot of times someone at the top just needs to make decisions so we can move faster. I don’t think I am the best decision-maker on earth. I like other people that make decisions and if you’re smart, I will trust you on the stuff.
Before Dave comes into this, how is Flexport structured now with you as the CEO?
I am the CEO, and I have seven direct reports. We have sales. We have procurement, which goes out and procures ocean freight, air freight, trucking, et cetera. We have operations which executes the transactions. We have account management and a network optimization team. Those are the core. Each of those teams has their goals and then there is a tech org that supports it. After that, you have governance layers, finance people, legal, and compliance.
Correct me if I’m wrong. Your tech org doesn’t report to you?
It does. The CTO reports to me.
Okay. That is under the CTO then.
Yes, but he has teams that are serving each of those functions. They are paired up to make sure that we deliver tools for the customer and for the supply side. For a lot of the supply side, it’s like what I mentioned with the trucking software. We have warehouse software, and a lot of it is integrations into other IT systems. There is a whole team that does integrations and an ecosystem team that connects our APIs with the rest of the world. Then operations builds software for the operators.
So your head count is 3,000 people. I’m always curious, is that mostly software people? It seems like every software company is escalating.
No. I wish we had more. We have less than 600 people making software out of the 3,000. Call it 20 percent or so. We have a lot of sales and account management and a huge amount of operations. These are people who are coordinating these shipments, clearing customs, and doing data entry. We still have to ingest a lot of docs.
We have built a lot of great optical character recognition tech that does machine learning and ingests documents, but there are humans that have to review these things. Compliance is super important to make sure that you are filing the correct documents that customs needs. There are large teams that do that work too. You would be surprised how many people end up working in finance, HR, et cetera, when your company gets big.
Do you sell a clearance time? Are you like, “Look, we guarantee that our team will knock out this customs compliance in 48 hours”?
Yes. We do have service level agreements (SLAs) for all these things. I don’t think we are good enough right now to do that, but that is one hundred percent what all the teams are building towards. They are measuring it and saying, “Hey, what is our P90 transit to P95 transit time?”
What is a P95 transit time?
P95 is “5 percent of shipments will be worse than this.” Ninety-five percent of shipments achieve this metric of transit time on this lane. What we want to get to is where we sell deferred freight that is cheap but unreliable. Might go really fast, might go really slow, but it’s cheaper than standard. Then there is a premium or even express rate where you are paying a lot more for transit time. The last couple of years have been really hard, man. It has been very hard to guarantee anything when there is a 40-day backlog at the Port of Long Beach.
We are putting in the systems and building towards that vision where we are really selling differentiated services to allow people to pay more to move inventory they really care about and pay less for the stuff that doesn’t need to get there on time. The logistics industry hasn’t really had this. You see it in parcel, but you do not see it in freight. With parcel like FedEx, you pay one price for overnight, a different price for two-day, and a different price for five-day. In the freight world, there isn’t as much distinction. That is what we are building towards.
FedEx is just end-to-end. They control their whole network so they can just do whatever they want.
That’s right. They own it, door to door. Maybe there is one hand-off to another local company in a different country, but it is basically one IT system, scanners, and the whole way control the network. They are running on a schedule, and that plane is flying whether or not you move your parcel.
In freight, that truck is not driving if you don’t ship the freight. It’s just different. It is harder to predict things. You don’t own the assets, you can’t schedule them. You are building lots more heuristic machine-learning models to say, “Hey, we think this ship is going to arrive.” We found that machine learning is really working on the estimates of when ships will arrive. We are better now at knowing when the ship is going to arrive than the company’s own schedule. We are more accurate by using satellites, machine learning, and estimating when the ship will arrive. We are getting there.
I want to talk about ships, and I definitely want to talk about the Port of Long Beach. The last question is the org stuff. Dave is going to come in as the new CEO. If he’s like, “This org chart is dumb, I’m blowing it up,” are you going to let him, as the founder and now the executive chairman?
We are already doing that. I don’t want to reveal everything because I haven’t told everybody at the company yet, but there will definitely be some changes. In fact, I will preview. Dave will have more of the business lines directly reporting to him and will elevate some of the teams to report directly to him. There will be more people on the directs to the CEO. I look at it and think it’s much better for the company, but much worse for the CEO.
It’s like 10 directs and then you run out of capacity, right?
He is comfortable with doing more, but I think I’m too lazy. I think my laziness has held us back. It’s things like that. You’re like, “Oh yeah, this would be better for the company. I don’t want to do it, so it’s perfect! You are going to be better at this job than I would be.”
Then you are executive chairman. Are you sipping mai tais and taking boat rides around the Port of Long Beach? What are you doing?
“I am comfortable with more uncertainty than almost any human I know.”
TBD, honestly. I am comfortable with more uncertainty than almost any human I know. I am one hundred percent committed to making the company and Dave successful. I assume I am going to find some really awesome ways to add lots of value, and I have a long list of what those might be. We are co-CEOs until March 1st, but I am not thinking about that at all right now; I am thinking about the next couple of months and how I will help Dave be successful. After that, we will find a way. I think I will be more free to come and hang out with you to do your podcast again, go to conferences, and meet with the world stakeholders.
I can only be in so many places and I am running this business every day. I can’t be out there talking to a customer, visiting an airline or an ocean carrier, or doing deals with the ports and the governments of the world. Hopefully it frees us up so we can do more of that really valuable work that’s hard to do when you have day-to-day responsibilities internally.
As you are talking about it, the model that actually comes to mind is Eric Schmidt at Google, who was the adult supervision CEO. Larry Page was like, “It’s my company, I’m the CEO now.” So Eric said, “All right, I will be the chairman. My job is to glad-hand politicians and ease the regulatory environment.”
I think it’s different than when Eric first got hired because Larry stayed and ran products in engineering. Nothing is going to report to me. It will all report to Dave.
Let’s talk about the Port of Long Beach. You would be on the show anyway because Flexport is a really interesting business, and I think trying to make software that moves atoms around is really challenging. You also have a CEO transition going on. There are lots of reasons for you to be on Decoder. But you are on Decoder today because you took a boat ride around the Port of Long Beach and said, “What the fuck is going on here?” Let’s be honest, it blew your profile up. Literally the thing that stuck to me was that you were like, “Just stack them four high instead of two high and we will ease the supply chain crisis.” They did it and it worked. Tell me about that. A software CEO doesn’t usually go and look at the atoms and say, “We are organizing the atoms wrong.” How did that happen?
We are more than a software company. We see ourselves as a logistics platform and we are responsible for moving these containers. Those are our containers that are stuck. We have had real operational problems and I was trying to unpack and figure out what is actually going on down there.
You have to go in person. I think it’s a really important lesson for every CEO, no matter what business you are in. You have to see it firsthand. Talk to that customer. Talk to that vendor. Talk to that frontline team. You will learn so much from getting out there. I think the Japanese call this “gemba.” Managing by walking around. Nothing too unusual, honestly.
This is our job to ship containers. I have to go down there and see what is going on. The tweet for the boat tour went super viral and I unpacked something really interesting. I found this catch-22 where the containers couldn’t get unloaded because there was no place to put the containers. It was this backlog that was piling up.
A week before that, I was trying to understand what was going on and I actually rented a taco truck and sent it to the port. I wanted to hear what the union had to say. I figured, “Okay, how do we get the union to talk to us? How do I learn about what’s going on here? Send them some tacos.” It took me so long. I want to give a shout out to Jean-Marc (JM) Sorhondo, senior director of west coast Americas operation, because every single person I talked to at Flexport thought I was joking about the taco truck and refused to do it. But JM was actually like, “You know, he is the CEO. He says to get a taco truck, so I’m going to get a taco truck.” So we did it.
Was that you out there with voice memos on your iPhone, talking to people?
We went down there. We got the truck. The union came.
Wait, you keep saying “we.” Was it you or did you send a team?
No, I wasn’t in LA that day. Our team, local op JM and this guy, Nathan Strang, run the operations for the company that were done in LA. They brought in and interviewed all these union workers and asked them what the problem was. I hadn’t heard their side of the story. They were blaming the truckers saying, “These truckers are not making their appointments. You need an appointment to pick up a container. Truckers aren’t showing up, and if they don’t show up, then the container just jams up to a bottleneck here.”
I was like, “Oh, that’s really interesting. I wonder why these truckers aren’t showing up?” The reason I rented a boat was not so I could see the port from the water. It doesn’t look that different. I took the CEO of a trucking company and I brought him on the boat. If you ever want to interview somebody and really get the truth out of them, you bring them on a boat. Because right now, I could hang up on you if I don’t like you. But on a boat, you are totally stuck.
Amazing. Decoder on a boat, baby. Let’s do it.
I took him out there and I was just like, “Oh, what? This is what’s happening?” I started unpacking the thing that way. I still don’t know that I have the full picture of everything. It is such a complex moving system as to why it all doesn’t work and which parts break down.
Your tweet thread — which we will link to so people can read — was very good. It was basically like, “Look, here are all these container ships that are stacking up outside the Port of Long Beach. They can’t unload because there aren’t enough trucks to move the containers. There aren’t enough trucks because the trucks are full of empty containers, and that is because we are not sending enough stuff back to China or wherever. Just stack the empty containers higher and create some capacity.” I didn’t do it justice.
It’s pretty much that simple. We can all picture a traffic jam. You have a four-lane freeway and cars are just flying down the road at 90 miles an hour when there is no jam. You have a lot of throughput and the cars are really moving through. The same number of cars can move through that freeway at very different speeds. There are multiple equilibrium points. One is in a flow state where they are all going 90, and the freeway can handle this much. They could also all be going 20 miles an hour in a traffic jam.
Fundamentally, you had a traffic jam with too many containers, with containers getting in each other’s way, so now they are not able to unload more containers because the yard space is full. The trucks couldn’t pick up containers because there is a literal traffic jam of trucks outside and they are missing their appointments.
At the end of the day, if you don’t have a place to put these containers, there is no truck that can come and pick it up. The Long Beach City zoning law only allowed containers to be stacked two high. On the ship, they can stack 14 high, so it’s possible to stack them higher. I tweeted about this and said, “Hey, all we have to do is stack these things six high, but make it four high for safety.” I tweeted this at 6AM the day after my boat ride. By 3PM, the Long Beach City Council had changed the zoning law. I am told this is the fastest change by government as a result of citizen action in the sum total of all of human history, actually. Not just in the American republican democracy, but all the way back to the dawn of civilization.
This is good. We are going to hold you to it. The producers are already researching your claim. We are going to do a thunderous debunk dunk on TikTok. It’s going to be great. Let’s talk about that for a minute. I understand you want to say it is a logistics platform, but at the end of the day, you make software that connects people. Why isn’t the operator of the Port of Long Beach seeing this problem? It doesn’t seem like that hard of a problem to see.
You are now asking a problem about institutions more broadly. What is going on? What is happening? How do we get some more skin in the game? How do we get people believing they can do things that are outside their job description? How do we get them to solve things? I don’t think this is specific to Long Beach Port or even a public sector thing. I think you see the same in big companies. What’s with the inertia? Why can’t they just step up? Why is it that two guys in a garage can do so much more than the 1,000-person engineering team? What happens? There is something about the nature of large institutions where people are a little bit too divorced from the customer that is experiencing the problem and they don’t have enough skin in the game to solve it.
The reality is that the world during the pandemic experiences so much change so fast and institutions are not made to handle that kind of change. Take a port. The port always moves the same amount of containers every year. It just goes up by a couple percent. Global trade has grown three or four percent annually for the last 800 years since the Mongol invasion. You don’t expect overnight.
The number of containers went up by 20 percent in one year. You don’t have 20 percent extra capacity sitting around in case there are that many more containers. Everything breaks down when you have an institution that is used to doing the same thing day after day, year after year. A company like Amazon is used to 20 percent to 30 percent growth in a physical world. They have built systems for handling that. Flexport is getting there. A port just doesn’t have that, and I don’t think we should expect them to.
I will give you another example. When the pandemic hit and we saw there are not enough masks in our hospitals, I found that to be totally unacceptable at a civilization level. We owe this to our first responders, to our doctors. If you asked me to go in there without a mask with some weird disease that might kill me and that no one knows anything about, to serve these patients, I don’t know if I could. These are real heroes that were willing to do it. If you don’t have masks and the doctors don’t show up for work, civilization collapses pretty soon after that. That was very unacceptable to me. I rallied a team at Flexport and we stepped up.
A big part of why there were not enough masks is because all the world’s airplanes were grounded during the pandemic. They were not really flying to China. Fifty percent of the world’s air freight flies in the belly of passenger planes. If those passenger planes are grounded, there is no air freight capacity. We found there were lots of masks available in China. They have ramped up production, but we don’t have air freight to get them in. The rest of the world, as far as I can tell, looked at that problem, put their hands up, and said, “Eh, fuck. I guess our doctors are going to suffer. Let’s watch this on TV and see what the people are saying.”
If you listen to the problem statement, 50 percent of the world’s plane flies in the belly of passenger planes and those are grounded right now. The solution is so obvious. Look at all the planes that are grounded. I managed through investor networks and connections that I have been fortunate enough to build over the years. I called the airline CEOs and I was like, “Hey, can we use your planes? We are going to go get some masks to save America’s hospitals.” One hundred percent of them said yes. United Airlines gave us free flights. Atlas Airlines gave us a 747 for free. We were getting Dreamliners for a 200K round trip. Ask your super-rich friends if that is a good deal on a private plane, a round trip to China on a Dreamliner. We flew 83 planes, completely full. We filled the overhead bins and the seats. In the end, we shipped 500 million masks to America’s hospitals.
It was like, “Wait a minute. Why are we the ones doing that? We’re not supposed to be in that industry.” The value in that lesson for the whole world is naïve optimism. Try it. Let’s see if we can solve the problem. You can do more than you think you can. It was very inspiring for everybody at Flexport to see, “Whoa, this is working. We actually made this happen.”
It happened because you could make some phone calls. There is a point at which the people in charge have to act like the people in charge. It seems like one thing you do is you just ask the people in charge to do a slightly better job. Stack them six high.
Maybe I’m in charge now. It is very terrifying for me. I’m like, “What, I’m in charge? What the heck is going on?” We need more people to step up and believe that they are capable of being in charge, I guess.
This leads me to one of the recommendations you made. You made a lot of recommendations on how to rethink shipping and maybe solve some of the supply chain crisis. One of your recommendations that I have read is that the United States government should treat our ports more like a strategic national asset as opposed to what they do right now, which is that every city gets to run their own port. At the same time, if every city runs their own port, some ports might be better run than others, as opposed to the inertia that we have felt — particularly, the inertia of the Trump administration in responding to the pandemic and the inertia of the Biden administration. The recommendation is that we should think about it more like an asset and control it more, but the reality is guys like you are just going to make a bunch of phone calls and get it done. How do you split the difference?
It is a really tricky problem. I don’t think the federal government is that competent and should run all the stuff. I would call myself a libertarian if I had to put myself on a spectrum, but it is impossible to be a libertarian if you are a realist and a pragmatist when it comes to ports. Not everybody with a beachfront property should go build themselves a port. I think that it is an absurd argument that we should let somebody who happens to own a beachfront property open a container port on their beach. That’s crazy. Physically, geographically, there are only a handful of sites that can be ports. It’s a tricky problem.
Even if the local governments were equipped in doing a good job, they just do not have the right incentives. Take the city of Oakland here in the Bay Area. Oakland has a really bad murder rate and terrible public school system. I don’t want to speak too much about all their problems, but they have problems to solve. Healthcare, schools, all these things. Yet the Port of Oakland reports in and is part of the city’s mandate. If the Port of Oakland is causing problems in the hinterland, it is causing problems for businesses that are based in Iowa. It is not their problem. You know what I mean?
“Frankly, we have not invested in our ports. I don’t really care who does it or how it’s done, but they need to be invested in.”
You call this an externality in economics. How do you solve that externality? How do you build something? Frankly, we have not invested in our ports. I don’t really care who does it or how it’s done, but they need to be invested in. Our ports have a productivity rate, a throughput and a labor rate — meaning labor per container moved — that is worse than the ports in Kenya. Meanwhile, Singapore has this fully robotic port that is coming online. What is possible would blow your mind. I would like to see this infrastructure be invested in. I don’t really care who does it, but it needs to happen.
You said Biden administration inertia. Until yesterday—
What happened yesterday? What did I miss?
I just mean the concept of yesterday — until very recently. They passed the gigantic infrastructure package. There is $17 billion for import infrastructures in waterways and $25 billion to airports for repair and maintenance. As with everything with the Biden administration, there is a heavy push towards carbon-neutral technologies and climate electrification. Did you see that package come through and think, “Okay, this is the first step to solving the problem”?
I personally don’t think it’s going to do anything. The way it works is they send money down to the states and then the states have to commission studies. I think by the time it gets to putting a shovel in the ground to build a new port, $0 will be spent. That is my working assumption, and I will tell you why.
That is very much the libertarian in you, but go ahead.
We have famously had all these backlogs, with hundreds of ships waiting off the coast. In Oakland, for the last four years and even as we speak, there has been an empty port terminal that no one is running that you could just turn on to start unloading ships.
Have you been tweeting, man? What are you doing with your time? I’m told that you are the fastest and most effective city council zoning operative in the history of human civilization.
“I called them and offered to run the port.”
I called them and offered to run the port. I was like, “I will turn this thing on.” They won’t let me. I wasn’t serious since I think it would be bad business. I wasn’t sure I really wanted to do it, but I was experimenting to see what’s possible when pushing the boundaries of the simulation. They are not interested and I don’t actually want to run the port. I was mostly just figuring stuff out, so maybe they didn’t take me seriously or something. Why aren’t we turning on this port? What is going on?
You haven’t been able to find an answer?
It’s not like we don’t have the money. If you want $17 billion to improve support, start with the one that is right there that is not working. Let’s run it.
You are not the CEO anymore, so let’s say Biden appoints you Port King of America and you have absolute authority. What are your moves?
Fundamentally, the ports have a very difficult relationship with their labor. The union is very highly paid and very brittle. Getting paid a lot doesn’t bother me, I think that it’s fine. They are very brittle in terms of who will do what and how their org structure works.
For example, the way that it works today is that a port terminal operator, this private company who leases the land from the government, leases the port terminal and says to the union, “I need 2,000 workers tomorrow.” The union furnishes those 2,000 workers, but they only do it one day at a time. Each day they say how many workers they need for the next day, then the union provides that many workers. Different people every day and no team structure or management process.
There is no ability to say, “Do you know how to operate this machine? Have you been trained on it? I am your manager every day. Here are the metrics for yesterday. Here is how we are going to get a little bit better today. This is what we are going to work on.” The fundamentals of running an operation is like, “How did we do yesterday? What are we going to do better today?” They don’t do any of that.
From a starting point, if I’m the czar — which comes from Caesar — then I am in charge. I would do a big buyout and take care of the union. I don’t want politics. The unions have done a lot for the world, getting us the weekend and all this stuff. There are good reasons for them, so I don’t mean to be disrespectful, but we have to move forward. We need technology and the union doesn’t want it. We need safety, we need efficiency, we need robotics, etc. If we are the czar, we have infinite money. We are going to buy them out of their contract and take care of them. Maybe not take care of their kids and their grandkids the way the current system works, but you are all getting a retirement package and going home.
Now we are free to run this thing like a well-oiled machine with full robotics. Let’s go invest in awesome robots. Where is the giga crane? We still unload these ships one container at a time. There has to be a better design. Can’t we get 100 containers and grab a carousel or a ski lift that grabs a container every second? Maybe a Pez dispenser design. We flip the ship upside down.
A Pez dispenser?
I don’t know! Flip the ship upside down and dump all the containers out. Let’s go, man. This idea, this invention, this current system, it was invented in the 1960s. We have to have better ways to unload these ships faster. There is a lot to be done there, but it’s a government scale project and infrastructure. Even buying out the union is probably a $10 billion proposition.
I understand why you would want to buy out the union. You just want to clean the slate, start over, build a new structure, and hire people who are bought into that structure.
Take care of them. I want to get out of the fight. They are not going away if you just say, “Hey, we can’t work like this.”
I will throw some other examples at you. First of all, I manage a unionized newsroom. We have open conversations with our union here.
All right, all right.
It is different. It’s a much tinier group of people, and there are a lot of relationships there. Let’s look at car makers though. Most of the car makers in America are radically reinventing themselves as EV manufacturers. Every time Ford announces a new electric vehicle, they have the UAW standing next to them saying, “We are very excited about this transformation and this new factory with more robots in it.” They have managed to close that gap.
Why is that? What is it? How do they get them interested in robots?
I guess that’s what I’m asking you. Is it a bottleneck for a labor force that seems disinterested or protective? Are people not inventing stuff that captures the imagination enough to get people to buy?
That is entirely possible. By the way, the union agreed to do containerization, which was probably a much bigger change than what we would be proposing going forward. There used to be that backbreaking work of unloading these ships by hand. Then we switched to containers and they eventually embraced it. They fought it for a long time and embraced it eventually.
It is possible and is the change that we would want to push through. I just don’t know if we have time to debate it all and try to get it solved. The current status quo seems pretty far removed from that. Look, you asked what I would do if I was in charge and dictator of the world, right?
You just picked one thing that seems really interesting.
I think that a lot of the solutions I would want to do would not be super popular with the union and rather than getting into an argument about who is right, I think they would agree. They would be happy to just get paid and get bought out. If you are out there in the union and listening to me, I am not anti-union. Actually, my proposal was awesome. I just told you I was going to give you a ton of cash.
Just to be clear, I am not anti-union. I am anti- “adversarial relationships between management and union.” It’s breakdowns like, “Hey, let’s be a team here. How do we get a process that lets us all get better every single day?” I am not blaming the union. It could be the management’s fault that they don’t have that, but it’s clear that they don’t. They are treating labor as this fungible thing. You say, “Hey, I need this many workers tomorrow,” as if all workers are the same and it doesn’t matter how they’re organized or what team they are on.
The current system over there is not working. That is my outsider’s view. I have never worked in the port. I have no idea what I’m talking about. I’m just a guy on a podcast getting asked questions I never thought about.
That is the whole point of Decoder. Do the ports know you feel this way?
Yeah, and they make fun of me. As they should. I’m some outsider that doesn’t know what he’s talking about with some ideas about how they should do things. I’m sure I am wrong on a whole bunch of stuff. I have also seen how they do stuff and I know that I could do better. Say what you would.
Now, I don’t know if I could do better in their circumstance. If I was thinking from first principles and we were starting fresh, I could say, “Okay, look. We got this nice piece of land on the water and a good harbor. How are we going to move containers through at a higher throughput than what we do today, given all the technology that exists in the world?” That would be self-driving trucks, cranes that could be as big or powerful as you want them to be, tighter process, and machine learning to read all the container numbers on that. There is just a lot to be done that I’m pretty sure I could do better.
Do you look at other countries like China or Singapore and say, “Oh, these ideas are actually in production”?
For sure. They are already in production. Rotterdam in Holland has had fully automated ports for over 30 years now. I actually think we could outdo all of them and create some amazing ideas that are way ahead of anyone in the world. Even just implementing stuff that Singapore, Shanghai, Shenzhen, and Rotterdam have done would already put us way beyond where we are today.
From your position, do you sense any of that desire to change because of the pandemic?
No, I don’t think so.
Broadly, what do you think actually untangles this supply chain crisis that we are in?
It is untangling itself, but not the way that we wanted. Remembering why this thing started, there were 20 percent more container imports. It’s nobody’s fault. Honestly, the ports always run that way and it’s not meant to have capacity for 20 percent more. The reason there were 20 percent more container imports is that all the consumers in the world were locked down and not going to restaurants and bars or traveling. They were buying stuff because you have to get your dopamine from somewhere. They were just buying more stuff and shipping more goods.
Well, that is changing back really fast now. People are going out, travel is booming, restaurants and bars are booming, hotels are booming, and the volume of goods has come down really fast. The bottlenecks are unblocking themselves, but not through anybody improving the port operations or any of the other bottlenecks. It’s just that there is less stuff shipping.
That is not what we really wanted. We in the freight industry want to see more stuff getting shipped all the time and more volumes. We would rather say, “Oh, we were able to dig ourselves out by upgrading our operations.” The economy has shifted back to the way it was.
Last question. I ask everybody this. What is next for Flexport after you move up to executive chairman?
We are going to be much more than a freight-forwarding business. We are getting our house in order to be the best provider of freight services from door-to-door anywhere in the world, but you’re going to see us be much more than that.
Shopify is our newest investor. We are building a lot of really cool stuff for e-commerce, to connect directly all the way to consumers’ doors through the Shopify ecosystem. We will be partnering with more awesome companies to get to the front of customer stores. You will see us add more services. We want to have an app store for global trade and all the things that you would want to do if you are trying to do business anywhere in the world.
My dream is to be able to go to some of our awesome brands that we ship for, and look at their current distribution network and help these guys go global by saying, “Hey, you guys are selling in the United States and maybe in Western Europe, but why aren’t you selling in India? Why haven’t you added South Korea? Didn’t you know there is this button you can click over here to go live on Flipkart, Rakuten, Coupang, and Mercado Libre?” That is my dream sale, helping them grow their revenue. You can make more money if you are helping someone grow than if you are a cost center for them. We are going to build all kinds of cool stuff like that in the coming decade, with Dave at the helm and me as his partner.
That’s great. Well, Ryan, thank you so much for being on Decoder. This was really fun.
Of course. Also, if you are out there in the union, again, please, I have no horses in my house. Thank you for unloading our containers. I love y’all. I’m just trying to get you a big payday.
We will have you back when you are appointed czar of all ports in America.
The Verge union, please keep writing nice things about us. We love you guys.
I will let you take that up with them. The number one thing they negotiated for is editorial independence, so you have your own problems there.
All right. All right.
Brother, thank you.