Why Global Supply Chain Continues to Thrive with Ryan Petersen of Flexport

Welcome to The Future of Supply Chain. Each episode, we bring together leaders across the supply chain space to discuss the role of technology and business model innovation on the future of supply chain. The Future of Supply Chain podcast is presented by Dynamo. Dynamo is a pre-seed and seed stage supply chain investor. To learn more about Dynamo and this show, head over to www.dynamo.io. www.dynamo.vcpodcasts or subscribe on the platform of your choice. Now let's get into the show. Here's our host, Santosh Sankar.

Hey ladies and gents, welcome back to the Future Supply Chain Podcast. I'm your host, Santosh Sankar, and joining me today is a man who needs no intro in supply chain, Ryan Petersen, co-founder, CEO of Flexport. Welcome.

Hey, it's great to be here.

So ordinarily I kind of go through this, like, tell me who you are, you know, what you're building, kind of what you're focused on. Flexport's a household name to your credit. I think for most people who are involved in sitting around supply chain, and I'd argue even kind of those who are kind of thinking about their consumer spending, their e-com activity, So with that, I'm going to kind of jump right into the conversation if that's cool with you. And a bit of a recap of this year, because I think you've had a very unique year as you've kind of navigated company building. And in late '22, you announced plans, ultimately transitioned out of day-to-day, appointed a new CEO in Dave Clark. And that didn't go as planned. It's what it is. You've stepped back into the CEO role. And you're now steering the ship forward at Flexport, no pun intended. And you're also, I guess as you're coming in, you've made a lot of changes very publicly at that. As you think about how you've managed your organization, we have 2 M&A kind of targets that you're now digesting and delivering, Convoy. And they've each had kind of their own opportunities, challenges. And I present all this because I'm kind of wondering, how are you approaching now this next phase of Flexport's growth on the backdrop of your core business seeing lower rates, lower volumes, and kind of steering through all that turbulence?

Yeah, well, it's exciting times. We're— you asked, you know, how am I managing it? Well, it's probably the same way as usual. A few big changes in my own Mentality, behavior, which is really about what we focus on. And I've joined Founders Fund as a partner as well, and it's, which is one of the top venture capital firms out there. And it's given me a really good fresh perspective on, not that I never thought about Flexport from an investor's perspective, but when you're really part of the team that's doing that, looking at companies all day, you take a new, a little bit of a different approach. And I think balancing those is good. It's And I guess the big difference is it's really about return on invested capital at the margin. You know, like what's that next dollar of spend going to return to the company? And then looking at really all the dollars of spend and going, hey, how does this add value for customers in a way that gives you a competitive advantage that allows you to earn excess returns, let's call it, that allows you to make a higher return on invested capital? Now, I still come back to the way I've always managed the company, which is obsessed on customers. How do you drive quality? How do you define, measure, and drive quality results for the customers. We have made 2 pretty important acquisitions. One of those we made kind of in, I want to say April of this year, we closed the deal, maybe March of this year. We negotiated it for a while. So I can't remember the exact date of the close, but we brought that team on in April. That's the Shopify Logistics Group, which we've, and they had acquired a company called Deliver.

Yeah.

Which is now Flexport Omnichannel. So this is our last mile sort of fulfillment, retail distribution, and last mile network focused on the United States now, but we're going to take it global over the next few years. And then we acquired Convoy, which is digital truck brokerage with 400,000 drivers on a mobile app. I didn't realize until we got involved in In the acquisition talk, just how much technology they'd actually built. I knew Dan, I knew the business, we'd partnered with them, but to see the level of, to just get it, managing 400,000 drivers in a compliant, secure way, the cargo security, the amount of, you know, running 6 different background checks on every load, every time they assign a driver, anytime the driver veers off the assigned route, making sure that, you know, it sends an alert to security. So both of those offer really awesome things for Flexport. One way to think of that strategy is it's a one-stop shop. You can ship anything anywhere in any quantity by any mode as one of the core parts of Flexport's vision. Another way to talk about it is true end-to-end where from order, placing that order out to your factories, receiving that order from your customers, placing a replenishment order to the factories around the world, and then delivering it all the way back to those customers' hands. That's like true order-to-cash lifecycle that we can manage for customers. So the strategy, those things are both in the vision. I think they were things that, you know, given how hardcore freight forward, global freight forwarding, our core business is, I had sort of said, yeah, we'll get to that in, you know, 5 or 10 years. But when acquisition opportunities like that come along, And Flexport being a pretty good place from a balance sheet and leadership standpoint to be able to take those deals on, it was kind of a must-do, even though, yeah, we are going through a lot of core transformations as I come back to the company. Well, one of those we did with Dave and Convoy came, just did a few weeks ago.

And I would add like Convoy was specifically just the technology also that you ended up purchasing.

Technology. We acquired the technology. It was a technology company. Um, and around, we hired 50 of the top engineers, you know, the ones that we thought could really build and extend the system. Um, so we feel pretty good about that. Um, oh, and we brought on some of the sales group, you know, so we're going to, we're going to sales and account management group. Um, we're talking to customers, talking to carriers. It's mostly mom and pop carriers. 400,000 drivers with 60,000 carriers. So you can look at the ratio and see, oh, okay. It's mostly Almost small, all small kits, 4 or 5 truck carrier kind of thing. So we are going to light the business back up and timeline for that is TBD, but not that far in the future. We're aiming for Q1, bring back some of the good customers, turn back on the marketplace. And so, yeah, I think, I expect, you know, we acquired the technology, but I think we're going to get more than the technology as we bring that back to life.

Sure. And so as you've said, you always had this end-to-end vision, and given the strength of your balance sheet and the opportunities that you were faced with, right? Was part of capitalizing and acting on these opportunities presented to you informed by kind of being active as an investor as well and being able to kind of see something more holistically and And really what you've done is like you've accelerated the TAM that you're ultimately going after and solving that customer pain point more holistically.

No, these 2 particular acquisitions are both, they came to Flexport because of personal relationships I'd built over many years. And in specific with the CEOs of Shopify and with Convoy, became friends with them. And so You know, maybe you could tangentially say, oh, I'm an investor and meet more people or something. But no, I've known these guys since way before I was a venture capitalist. And yeah, both cases. Well, in the Shopify case, Tobi called me, wanted to, was looking for a better way to drive the Shopify logistics, now Flexport omnichannel, drive that business faster. Saw the expertise we have on our team in logistics. And saw what a good combo it is there so they can go more focused on e-commerce. And so that's where that deal came from. The Convoy one, I called Dan not to acquire the company, honestly. I called to check in on it. I'm friends with the guy. I was like, oh man, this is, you know, they had just kind of effectively shut down the business. And yeah, I called him to see how he is doing. And then quickly he introduced us to the bank because it was at that point they had taken on some venture debt and it was bank owned. So then we called, the deal was negotiated not with Dan, but with the bank. Dan kind of facilitated the intro, but yeah.

So you've made mention over the last several months, like your ambition still is eventually over a period of time to return to profitability. Flexport has been profitable in the recent past, and I think even kind of your most recent interview with the Journal, you made mention of that. But you ultimately do also want to take this company public. It has also been a statement you've made. So with that in mind, kind of what are the focal points? Like what is top of mind You've certainly been extremely customer obsessed, and I think that is clearly one ingredient to the recipe for success. But are there other things that you're prioritizing as you're going through this kind of growth stage, mid-growth stage, late growth stage as a business?

Yeah, it's, well, it's profitability and certainly going public are both kind of output metrics, but we mostly focus on input metrics. While checking the output metrics every day. It's sort of a good way to manage a company. And some things, not everything can be measured. Not everything that's important in this world can be measured. That's right. The first area, and you can measure it, but it's often a feeling, is the culture of the company. Because fundamentally, that's the competitive advantage that we drive, is having better people who are better organized, doing better work at higher velocity. than the competition. And that is the goal of a culture. It's not this like wishy-washy thing. It's like we gotta drive velocity towards our mission, towards our purpose, which is making global commerce easier for everybody. We want to make global commerce so easy that there's more of it. We think that would be very good for the world. It's not, it doesn't have to be some high-minded, oh, we're saving the world kind of thing. It's like just help businesses run better. It's obvious how that will Help our customers' businesses run better. That's obvious how that would make the world a better place because they'll hire more people, they'll make more money, we'll create opportunity. Like, so it's pretty easy to get people pumped about the mission at Flexport. And then yeah, the goal of the culture is to drive velocity, which if you remember from your physics class, velocity is different from speed. Velocity has a vector, so it's speed in the right direction. Sometimes you're changing directions.

Yeah.

In fact, in supply chain and logistics, you're changing directions a lot. Like, you know, it's like week to week there's new kind of chaotic circumstances happening, especially on the global side with like missiles hitting ships and trade wars and what, you never know what's going to happen. It's a fun industry. So the culture piece, how do you drive velocity in the culture? Well, it's having the right people in an org structure that lets them do a lot of work without too much bureaucracy. From there, it's really about trust, like building trust between leadership and the frontline and between all the other teams that have to work together and that should work together. So building trust, doing the things that you say you're going to do is a really simple way to talk to that. Getting back to the way that we often did things, which is that we did things earlier in our life stage, which is being really transparent, as transparent as we can be. Around where we are as a business. What are our goals? What are people working on? Try to break down silos. So that's the trust piece. Build a culture of learning. Like, we really want to get people excited about the very fact supply chain is the best industry to learn about. If you like business, like, you should work in a supply chain company because you get to see how all the other businesses are run. You're a backstage pass to the world economy and see what's going on and how you can, and then you get to get in the game. Love that. I'm like, Unlike fans, I'm a fan of business, but unlike a fan, I get to get in the game sometimes and help people with their business. So it's pretty, it's an awesome industry. So we get, we try to get people really excited about the learning opportunities that that creates. We want to be good at making decisions. I think a lot of cultures suffer from decision, either going too slow or making bad decisions. And so balancing that, Sometimes you got to go really fast and sometimes you got to delegate decisions. Sometimes the CEO needs to make a handful of decisions or leadership. The ones that can't be reversed should be made by senior people with a lot of judgment and experience. So you got to be good at making decisions. You got to have focus and direction. So that was one of your early things, what are you focused on? For us, focus right now is definitely, it is the culture itself, but it's a culture of quality. Delivering quality results for the customer. And that can be measured pretty precisely. It's on-time performance. It's your response time when people ask, escalate something. So case resolution times, inquiries, quote response times, how fast can I get back to people with a good rate? Win rates is a good indicator that your rates are competitive or that you're offering good service. It's things like net promoter score. It's not my favorite metric, but it's better than most of the other ones. And yeah, so having good focus and direction on what are we trying to do, not taking on too many things. Hard for me because I like to do new things and take on a lot, but focus and direction is definitely a really important part of the culture. Then the next piece is quality. Focusing on that. It's part of the culture, but it also has to be like, okay, what are the initiatives? Like, what are we doing specifically on that? There, I've said enough what those metrics are, and then like long lists of important and well-delegated initiatives on that. The third real focus area, so it's culture, it's quality of our operations and our business execution. Third is really cost discipline. We had Accrued some bad habits. Many of those over the last decade of venture capital came easily to Flexport. I've never seen a company raise— we've raised more money than almost any company, venture-backed company, and probably more than any in supply chain. And I've never seen a company that managed to raise even any amount of money without pretty quickly increasing their spend and losing a little bit of the financial discipline.

Yep.

That's not— And then that got worse over the last years. We really went very aggressively, Dave and I, during our time working together, hired a lot of engineers, kind of went big, put pedal to the metal, increased our expense rate quite a bit. And we're having to unwind that and take things back down to a level where we're like, hey, let's live in reality. We don't sell Gucci handbags. We sell global logistics services. And people are going to, you know, people are focused on, our customers are focused on price. It's got to be affordable. They're willing to pay for service and quality, but like ultimately you got to be in range on the price. And so that means any extra costs that we bear, we are going to have to increase our prices and be less competitive. So we've been going pretty hard at looking at every dollar we spend and what's that return on invested capital? How can I show that this adds value to the customers? Or is going to generate high returns from an investment standpoint in the years to come. So that's kind of the 3 focus areas for me, and I'm spending all my time on that and almost no time on anything that isn't tied to one of those 3 things of like delivering quality, upgrading that culture of the company, and yeah, cutting costs.

Got it. So I'm going to step ahead here and Over the last several years, we've observed startups, some in our portfolio here at Dynamo, where they started with a transactional component, but they've also layered in SaaS. And now Flexport, you've championed software for a long time, and I'd be curious, how do you explain the value of your software alongside the transactional business? That you lead with, whether that's in kind of container logistics, the omnichannel, the to-be-rebooted digital freight brokerage. How do you explain that to customers when you're interacting and engaging with them?

Oh, well, SaaS can mean a lot of things. We don't have a subscription SaaS offering where you just like pay a monthly fee per seat or some other thing. We use SaaS, we use software to do a few things. One, we have, we enhance our core offerings in logistics, and there's a million ways that software makes the logistics offerings better. More visibility, more control. Global supply chains and domestic ones as well, but especially global ones are fundamentally a communication challenge. You have the 3 kind of core flows of a supply chain. We usually just think of the goods flowing, but the data has to flow in parallel and be made available securely with the right permissions to the right people in that chain to make decisions. And they have to be able to put data back in about what they have done, what they are going to do. And ideally it's not just people, but the assets, right? And therefore software talking to software to lower transaction costs and get more timely and accurate data out of those assets. Where is this thing? What's happening? Right now. And so it's the goods, it's the data, and then it's the cash flow. The cash has to flow around. So you've got these 3 core flows of the supply chain, and it's on a global basis, 24/7, 365, all over the planet with thousands of companies. I mean, on the Flexport network, there's probably close to, I don't want to overstate this. Well, now that you've got 400,000 truck drivers on mobile apps, I would say there's, yeah, you're probably approaching, you're over half a million different entities that we're connected to Flexport. It's pretty crazy to think about, you know, and, and of course that's stretching because you're like, okay, every single truck, but yeah, every single truck that we've got. And then all the ships that we connect with, all the importers and exporters, the, the way Flexport's business model works is we'll We tend to sign up the importer, although we're increasingly signing up exporters around the world. But we tend to historically have signed up like an importer in the United States, and we've got a really fast-growing European business. And we tend to sign up those companies there historically, and it's changing pretty fast with the rise of e-commerce and especially Chinese e-commerce. But historically, since I, you know, I don't want to go deep history, but let's say, you know, the last couple of decades, The importer of freight forward on freight has made about 90% of the purchase decisions about who the freight forwarder is. I think it's coming down. I hear it. I've seen stats that it's more like 80. I don't even know where that stat comes from. That's a guttural butt in the industry. That's our understanding of things. And so, yeah, we tend to focus on the importer, but now at Flexport, when you sign up an importer, on average, we're getting 18 of their factories to come on board and become users, right? And then if I get those users to import, to sign up more importers, et cetera, And so, and then the way the business model works is the importer plates orders to those factories through Flexport. The factory is a user, hey, come on. All right, goods are ready. Come pick it up. Go and execute the transaction. And so making that full end-to-end loop run smoother, more successfully. Your question is about SaaS. So like, yeah, there's a million things that software can do there. We do have elements of software that we charge for, but we don't charge on a monthly subscription. It's a transaction-based service fee.

Mm-hmm.

It looks, in that case, you know, there's some good, great software companies that are based like that. I think Stripe is a good classic example. Twilio is another good one. For that, it's sort of like our order management software, pay per transaction. Even customs brokerage, although there's definitely a human element, a lot of the way that we run our customs brokerage is with With great automation and software. And so we're charging per transaction. But most of the way to think about Flexport is sort of managed services where software makes things a lot better and gives us a competitive advantage. I think we're good at building software for global logistics and trade and financial services. Now, the reason we build good software that solves users' problems is because we're living it ourselves. And I think that's what makes us different than a traditional software provider coming into this industry, is that we feel that we're in these transactions so deep that we know better what that user experience should look like. We're better at building freight forwarding software than a freight forwarding tech company would be because we're talking to the end customer every day about what they need.

Get the dog food effectively.

We're dogfooding it and we're building for the importers and exporters and not just for the freight forwarders. And those are our customers. So it lets us create different classes of user experience than someone who's just like, oh, I make tools for freight forwarders.

That's super interesting. So, you know, through this year, being a supply chain-focused investor, that's all we do here. Getting questions about, hey, like whether supply chain startups that are like solely transactional, and when I say transactional, I don't mean the elements of software that you just outlined. I mean kind of digital brokerage, forwarding, the like. Are these really good business models for VCs to ultimately fund? And those familiar with the cyclical nature of logistics are not necessarily shocked by the current market. Like these things kind of come in cycles. You have the macro cycle, you have freight cycles, but how well suited are VCs and the general audience for this? And are they right to kind of question, hey, like, are these even VC-fundable businesses? Because I have a particular bias, but I'd be curious your take and your rationalization of such.

Well, investors are looking for return on invested capital. They're looking for low risk. And they're looking for a long duration, which is not immediately obvious, but good investors want long duration because, you know, you could double my money in one year and then next, and then it's over. That's a short duration investment. Now I'm like, okay, great. What am I going to— now I have to find a new investment. Much rather something that doubles my money every year for 20 years, right? I actually want a long duration as long as I'm getting that return on invested capital and I have low risk. Like those are your kind of 3 things that you want. As an investor, return on invested capital. Look, you don't want to be in an asset-heavy business because assets have too low of a return on investment. So if you can avoid that, people focus a lot on margins, but you want to talk about free cash flow. And, you know, so it's a return on your dollar of investment in terms of free cash flow, not in terms— Jeff Bezos says it the best. You can't eat a percentage point of margin, but you can eat from a dollar. And so, you know, you can have a 1% margin business on a trillion dollars is a Who cares about the percentage point of margin? What you're looking for is fundamentally free cash that generates— look, asset-heavy businesses tend not to generate the kinds of returns you need, or they do so with a lot of risk because you're putting leverage on these assets and you get boom and bust cycles. So I would avoid things that require lots of assets, but a percentage point of margin is not an issue because even another way to think about margin, I mentioned the dollars, but I like to think of margin as how many dollars of free cash flow do you generate per hour of labor in your team? Because if your technology and your business model allows you to take a new grad out of college or even someone with a high school degree and have them make $1,000 an hour worth of labor, I don't care what your percentage point of margin is. I don't consider that a low margin business. You're taking new grads and turning them into high-end attorneys effectively for a business model. So that can be a very good business. Straightforwarding in particular, not to encourage more entrants into the space, but it's one of the highest return on invested capitals on Wall Street. You can ask financial investors. It is a cyclical business, but not necessarily way less so than the asset-owning sides of the business. And it's highly fragmented. 20,000 freight forwarders in the world. No one has more than a few percent. Everyone's in single-digit percent. I think the number one forwarder has less than 5% of the global market. It's something where technology can create a ton of value and competitive advantage. It's hard to build the tech, which is good. You make money doing hard things. And I speak to freight forwarding because that's the thing I know the best, but In general, I think as long as you're focused on return on invested capital, now low risk means, does mean avoiding assets. It means avoiding debt. It means making sure you're not in a fad industry, that this thing actually works. From that standpoint, supply chain is, well, it's been around with us since the dawn of time. People tell me it's the second oldest profession, is moving goods around the world. And from a step, you want to look at a hockey stick growth curve, pull up the growth of global trade. And I don't mean just since World War II, you've had 4% annual growth, which doesn't sound a lot to investors, but you've had 4% annual growth for 800 years since the Mongol invasions. And that when you compound on a long duration like that, you go, oh wow, like this is something you can just count on. Now you got to grow faster than 4%, of course, but if your market's growing 4%, There's some ups and downs and people are worried about the cyclicality. Oh, you know what's happening right now? The prices are low or something, and that's fair. But I mean, go look at the graph of trade and what's happened. You will not even notice, you will barely notice World War II when everybody was killing each other and tens of millions of people died and trade I mean, it's a little blip and you just continue onward. And it was, you barely noticed World War I. And don't tell me, oh, we're going through the worst time in all human history, like right now. We're not. Like, we've literally had world wars and the brink of the apocalypse and nuclear bombs were dropped.

Like, trade is enduring.

It endures. So like, you'll have cycles, ups and downs. There's blips. Like 2008 was pretty rough for the sector. Financial crisis, pandemics, and all kinds of gyrations through the cycle. There will be more in the future, I'm sure. I'm not saying it's all up into the right, but human beings fundamentally want to trade with each other. That's like, yeah, on a lot of sense what makes us human, that no other species is able to trade with each other and benefit from each other from specialization. We're the only species that does that. And that you could count on that trend. That's just something innate. Now, are governments and politics and things going to get in the way? Yes, but humans will find a way. We always find a way to route around and find loopholes and opportunities. And they change this rule so you can't trade with that country. Okay, you're going to trade with the other country and trade domestically or whatever. But it's certainly a great industry and one where tech can add a lot of value. So In fact, to the extent that people don't like it as an investor is the reason to double down on it. You don't make money just chasing the herd in investing, as you know.

Yep. Yep. No, I think that's all well said. And I guess some of my takeaways are, listen, kind of this is enduring and hyper-durable, but if you're not willing to have a truly long-term mindset, maybe this isn't for you because there will be some gyrations between here and there. But it's an amazing place to compound capital otherwise.

Yeah. Oh, and it's definitely not a get-rich-quick scheme.

No.

You know what I mean? You're doing real-world stuff. You have to actually add value. Like I tweeted yesterday about the role of blockchains in logistics and like, I saw that. Well, I'm a big fan of like new technologies and cryptocurrency and I'm not saying that Bitcoin's bad. I think you should probably hold some Bitcoin in your portfolio. Like that's fine. But like, From a logistics standpoint, any, I think there might be use cases for a shared database that's neutral that we all put data into so we can get track and trace visibility and like not have it owned by one company. So that if that encouraged more companies to put their assets on the platform and let everybody track each other's assets, it's not held by one company and therefore People don't feel like, hey, I'm giving my data away. And those people, that might be a use case. However, when you go and try to say, okay, what would be the value of the tokens that then pay out to the database providers?

Right.

Which is all of us participating. It would be so low that it takes away all the appeal of cryptocurrency, of get rich quick. And then you're like, okay, who wants that? Like the whole point of, for a lot of people, the point of cryptocurrency and It's like, I'm going to get rich quick. And you're like, okay, sorry. That's not going to happen in global logistics. You're going to have to do something that's really hard. You're going to have to stay at it for a long time. You're going to have to have a good return on capital, which means you're not going to hire 500 engineers off venture capital and just pray that someday it works. You got to have a working business model pretty early in your life cycle. And be very capital efficient, et cetera. So that'll probably discourage a lot of people, which again is a good reason to go into the space, like, because there'll be less competition.

No, I totally agree. Totally agree. I do want to touch on kind of pulling a prior thread. You just mentioned business models. You made the very intentional choice in, while you deliver value through software, You price that and allow the exchange of value to be transactional. And Ryan, I've sat in boardrooms with portfolio companies that are doing really well, have been able to solve a customer pain point through software, but get badgered as to why they're not growing SaaS or have priced their SaaS in a transactional manner. And I'd be curious, kind of how did you deduce and come to the point that Hey, charging a flat fee per month or per year is not where we're best aligned. Actually, what's best is to do it in this more transactional nature, even if it is software-oriented value.

It's not a universal recommendation by any means, right? Every business is different.

Sure.

I don't think people should change their decision, something they learn on this podcast or any other. It's like, you got to know your business way better than Anybody else? By the way, you should know your business better than your investors know your business. That's your job as your CEO. So, but listen to everybody, but you don't have to do what they say. The reason that I, for Flexport, believe that transaction-based and even logistics service-based business model's better than selling seats, and we had this debate, there are smart people at Flexport who believe we should just charge a seat license for everybody as an additional revenue stream that drops straight to the bottom line. I get the attractiveness from the investor standpoint. Like if you do that, Does not cost money. Every dollar goes to the profit line. The reason I think it's wrong for our business is because we're in a network effect business that every importer adds 18 exporters. And a lot of those exporters are now adding other importers. Something currently it's only 8% of the importers that come. If I get an importer, they invite a 18 exporters, only 8% of those exporters add another importer. But guess what? That's more than a viral coefficient of 1.1.

Yep.

So therefore you have viral growth. And especially if those importers then invite 18 more exporters, oh my goodness, you're going to go crazy, right?

Yeah.

So I don't think, I think that if I charged a subscription fee and it turned off, some percentage of users who didn't want to come on to the platform because they don't want to pay that fee, then I'm not getting the other 18 factories on board and it dampens the network effects. So that's a simple way to think about it mathematically. Another way is that, hey, we're really a land and expand business. We're very different than an SAP or an Oracle, which I guess those are also land and expand businesses, but The first landing piece is like a many-year implementation to transform your entire company and change everything about how you do business. And in our business, like, hey, the transaction-based nature of the thing is like, look, software's free. Come on, I'll get you an account right now. Ship something, see that it goes well. Ideally, I get you to ship enough things that we can have some statistically valid hypothesis testing about our, whether it's our transit time, our price against your index, the responsiveness, some of our quality metrics, some of the things we're trying to achieve for you, the way the data flows that you can trust it all. And so, and as we know in statistics, you need at least 30 in a sample, 30 items in your sample to have statistically valid output. So that's a nice sales pitch. I got to get 30 shipments, not one. But ideally we're getting some test here, and if it goes well, then we'll do more. And we have 20 different products, maybe more depending on how you want to define a product. But if you look at the top navigation at flexport.com, you'll see at least 20 offerings. Different things we can do. I'm trying to get you to do a bundle. The more of those things you do, they all connect together and you add more value. You can buy any one of them modularly. In fact, even yesterday we launched our carbon calculator that you can now track emissions on any global logistics shipment that is shipped with anybody else in a really simple form. And you can, it's free. You can upload a CSV file of your shipments and get back Here's our best estimate of the carbon emissions for that. So some of the stuff is just like, you don't even have to ship something.

Yeah.

But ideally we're getting you in the door, finding some hypothesis for how we can create value. We're testing that statistically and proving that we do a good job. Then you're doing more and more. And if I just charge you a seat license, you're like, I don't think I'll just say no to that seat. I don't know. I get very frustrated by software companies with like very expensive price tags. who won't let you try it out, who make you do a Zoom call, and then they want $6,000 a year before I've tried anything. As a buyer of that, I'm like, how about I just don't buy the $6,000? I'd rather test something and use it out.

Yeah. So you mentioned crypto, which was the last kind of tech hype cycle that we've lived through. We're currently in a period of AI LLM mania. How do you see this?

I'm not anti-crypto. I want to be clear. I just think for when crypto has to tie to the real world, you're then trusting real-world people to enter data correctly into your database. And so if I'm going to trust a forklift driver that they loaded that container with the stuff they said they were loading it with, I might as well trust the database provider. Like the point of crypto is that it's trustless and you can do business with people you don't know. Guess what? That's illegal in global trade. And you will be risking your life because if you import something from someone you don't know, yes, it's illegal too. If it's cocaine, you're going to jail for the rest of your, you know, for a very long period of time. And you really, you want to trust, you want to just do business with people you don't know? No, you do not. I promise you.

Yeah. No, I think your observation, criticism of blockchain is well placed and kind of how we view the world. And there are rebuttals to it, all that. But thinking about kind of, you have LLMs that have really taken the world by storm in the last 12 months. How do you see that improving? As you mentioned before at the start of the conversation, software helps us deliver value and become more efficient for what our customers need us to do. How do you see that playing out? Because at Dynamo, we think like we're in this very novel and really interesting period where a lot can actually be delivered in a more automated and high-quality way. There are caveats, things like hallucinations and such, but how are you thinking through it?

We're finding, I would say we've had more progress, certainly more promise, let's say. Like more promising real-world actual results in automation as a result of large language models and other— yeah, it's large language models. There's some, we do a lot of other types of machine learning, but we've been doing that for a long time. But the large language model breakthroughs that we've had in the last 3 months have completely changed my view about the possibility of automating the core operations of a global logistics provider completely changed my view. I was pretty jaded. I thought you'd make incremental improvements, a few percentage points a year, which is pretty good. If you, again, you do that for 800 years, it looks like a Silicon Valley hype cycle. You do that for a few decades, you'd be very good. And I, so I believed that deeply on just like kind of core software-based automation, but With LLMs, you're seeing that you can just do stuff and implement it in a week where it was taking a 6-month project. That, and the reason that that's working really well for Flexport, and I think we have competitive advantage doing this is, well, we have data at the level of like an incumbent because we are now, we're the 7th largest freight forwarder on the world's largest trade lane. We're the 3rd largest freight forwarder in the United States of the American freight forwarders. And so that gives you kind of the data scale to train models to make decisions. The specific types of models that you can do with this is routing. What route should we do? Which specific ship should I put this container on to achieve the transit time that my customer's trying to achieve in a very dynamic world? And that's one example. Fundamentally where LLMs really work, and you mentioned the hallucinations where they sort of, where they don't work. What we're finding is that if you can give it a very simple and discrete task, it does a really good job with very high accuracy. And this is where I think we not only, we have competitive advantage around data, but we've spent 10 years making the tasks of freight forwarding incredibly discrete, simple breakdown.

Yep.

Instead of moving a container from door to door and assigning that to a smart human, which is how freight forwarding has always worked, or a couple of humans on different sides of the planet, ours breaks it into a very simple task with nice custom workflow for each task. It's like 104 on ocean freight, typical container. I want to say there's 146 tasks that have to get completed. In our system. And each one has a really nice web form and some degree of automation that's been there for a long time of extracting data, like OCR data out of a customs document or something. But now we're just assigning week after week. They're like, oh, we automated this task now. And every week in the product announcements channel, we're putting out stuff. It's like, oh wow, they automated that. They did that. They did that. So we'll see. You know, it's still, it hasn't made a meaningful impact yet financially because it's just a few months. of work. We were early users of the OpenAI, friends with the OpenAI team on a number of, including CEOs. So we were able to get early access to the APIs there. Also working with a few other companies that are doing large language model, either a layer on top or make their own foundational models. And they're, yeah, the results are really powerful. So I'm very optimistic on that part. To give you that ship routing. In fact, that's, that's machine learning we've built in-house. And I don't even think that's large language model. It's more classical machine learning.

Optimization work, so to speak.

Yeah. What we do there, to give you an example of what, it's still a form of AI. I mean, we used to call that AI before large language models came around. What we do there, why you need tech for this. So let's say I've got a container that's going from Shenzhen to Long Beach. Simple problem. Or send Zen to Southern California. Sounds like a simple problem. It is basically. I've got 2 ports to choose from on the destination side, LA and Long Beach. I then have a number of contracts. I think we have like 8 different carriers that we work with on that core trade lane. We then, each one has multiple strings. That's the kind of series of ports that the ship will stop at along the way. Each of those strings has an expected transit time, which may, which is almost certainly different than the one that they publish from the carrier. So we have to build database, not, hey, look, it's hard. It's hard because there's weather events, there's port calls, there's issues, like it's life. But you, so we build our machine learning models that predict what that transit time will actually be based on historicals. And based on things that are happening, the weather and everything else. And so then you have an expected transit time from that sailing. Now there's this dynamic part, all of that's dynamic, but relatively easy to model. Then there's this very dynamic thing of, well, how much space do you have? You've got contracts with these carriers. How much allocation do you have on each string each day, each week? And then your factories, your customers are pushing the cargo ready date constantly.

Yep.

We, current run rate, actually this data is a month old. I haven't looked at it this week, but it's 1,600 times a week we get a change to the cargo ready date for one of the containers we're moving. The fact, you know, production's hard.

Okay.

Not ready this week. It'll be ready next week. When humans are doing this, there's 2 ways to do it. One is you cancel that booking and you place a new booking. In fact, that's what every freight forwarder does. They place multiple bookings with different carriers and they're canceling them left and right to make sure that, but when the, when the book finally does show up, When it does show up, they have it. And there's no charges for cancellations in ocean freight. Non-enforceable contracts. And on a week-to-week basis, they just, 30% of all containers that are booked with an ocean carrier get canceled. It's crazy. What Flexport does is we don't do that. We place fewer bookings. We're not double booking with carriers. We're placing the right amount of bookings based on our expected cancellation rates from our customers. And then when we get a cancellation, instead of canceling that booking, we look across our network and our software machine learning does this 10 times a day to find another container from a future sailing that I can pull forward. If that cargo's ready, I pull it forward. This machine learning has improved our transit time by 20%, reduced our cancellation rate with the carriers, down to the industry lowest. So it's around 10%, depends on the strength, but we're canceling like 1/3 of the containers that our competition cancels, which makes the carriers think we're the most reliable. We are the most reliable freight forwarder from a carrier perspective. Then they reward us because if you're an ocean carrier, put yourself in their shoes. 30% of your containers are getting canceled. What are you going to do? You got to overbook. No choice. You have to overbook your ship.

And then you end up rolling when kind of times are tight. Yep.

It's exactly, you know, but people who saw this, the airlines, like when we were kids, like they were constantly making announcements. I need, I'm looking for volunteers. I got to bump passengers, but they use data science and really just statistics to correctly over— they overbook airplanes too, but they're pretty good at predicting who's not, how many people are going to show up and make sure. And then they use it with the rewards program so that if first-class passengers book, they move an economy class person up there into that seat. And so it's kind of a similar idea, right? Like if I've got faster services, there are fast boats across the Pacific that don't make as many port calls. They're not faster ships, they just don't make all the calls along the way. Yeah. So they're faster and somebody canceled on that. I'm rewarding somebody else with a really fast shipment and they get thrilled, right? And it's like getting upgraded to first class. So These are things that machine learning can do in this industry that, but you've got to be building your own software to do this. If you're just relying on a third-party SaaS operating system for your freight forward, if you're a freight forwarding business running someone else's software, like it's nice, it's cheap. You don't have to hire all the engineers that we do, but you innovate at the pace of your software provider instead of going, hey, let's go tackle this problem right here and now and make it work.

Yep.

Yep.

So I'm going to kind of land us here to the finish line. I have 2 things I want to pick your brain on before we part ways. First, nearshoring. Everybody wants to talk about it. They love talking about it. What's the data actually saying from the position you're in? Is this actually happening? Where is it happening? What type of stakeholders are actually doing this?

It's absolutely happening. America's having a manufacturing boom right now. I've seen stats showing it's as big as the manufacturing boom we had in World War II. So there's absolutely much more manufacturing to be done in the United States, Mexico as well. Doesn't mean that trade's not happening because all the inputs from the manufacturing process are still coming from overseas. Not all, but many. But America, by the way, is the least dependent country in the world on global trade. We're number one in global trade, but we also do, as a percentage of our economy, it's the lowest of any country in the world. We're very fortunate to, we have the greatest geography on planet Earth. And this is not just because I'm American, it's indisputable. With the level of natural resources, the number one producer of food, we have more navigable rivers than all the other countries in the world combined.

Yeah.

We have more oil production. I think we have more oil than Russia and Saudi combined.

Yeah.

I mean, we're like number one by a lot in oil production. We don't export that much because we consume so much, but we're pretty autonomous and depend on trade way less than other economies. We're still the number one importer of goods because we're the richest country in the world and we can afford it. Like, yeah, man, the American consumer is just really special. We consume 70% of our GDP is consumption. China's like 40, I want to say. Europe is like 55. So we're, the American consumer is the backbone. We like to buy stuff. It's kind of awesome as a business person, you can find people to buy your stuff. But trade is not going down. Trade, you had a boom in the pandemic where trade went crazy. It's back. It is down from that level, but it's back down to kind of historical. I think you'll see it rise on historical levels. A lot of it is that companies are diversifying manufacturing away from one place, and that one place had been just China. A lot of that's reaction to tariffs. People will, you know, supply chains are set up to be low cost. That's the reason you put it in China in the first place. So tariffs definitely have a big impact. But frankly, Chinese labor costs have gone way up. Labor costs in Mexico are now about half of what they are in China. And that is, that was not the case 5, 10 years ago. Well, I lived in China 18 years ago for a few years running supply chain. And since 2000, the labor cost in China has gone up by 14x since the year 2000. And that's good for China. You know, their people are getting paid more. I don't, people are like looking at this like, they're gross consumers. Perhaps. Yeah. But I mean, that's good for the world. People should make more money. But there are cheaper, much cheaper places. Mexico is a cheaper place now. It's not necessarily a better quality manufacturing. Chinese has incredible manufacturing capabilities and quality. So people, many factories are still finding that's the best place to make their product because their manufacturing capabilities are really good. But people are diversifying a lot. Vietnam is going through a manufacturing boom. The infrastructure is just not there to support it. That sounds like a generic term, but there's the 2 biggest cities in Vietnam, Ho Chi Minh and Hanoi, are not connected by a freeway. You're talking about a 2-lane road with stoplights the whole way. The ports are not connected. You've got to, they're, Vietnam is fortunate to have a river network like, like the U.S. does so that you can, a lot of FlexPorts container shipping In Vietnam, it's our second biggest origin is Vietnam. A lot of it goes by river barge to the seaport because the trucking network's not there. India's got a, uh, is growing really fast in manufacturing. So you're seeing a diversification as well as nearshoring. I don't think the nearshoring is that much to save on supply chain costs.

Like, to be honest, shipping sovereignty, things like that.

Maybe it's quality. Certain areas, it's, there are incentives, CHIPS Act, making semiconductors in the US. There's other things that are government incentivized, but it's not about logistics costs or even transit time, probably because it's cheaper to ship a container from Vietnam to California than it is to drive a truck from Mexico to California. Much cheaper. It's cheaper to ship something across the ocean. Ocean freight's so cheap. I mean, it's about half the price to ship a container if you're going to one of the coasts. And even inland rail, if you're going from Asia to Chicago, it'd be about half the price of shipping a container from Monterey to Chicago. I mean, it varies week to week, but so, you know, it looks nice on the map. Look how close it is, but trucking, it's just not that efficient. And rail networks are kind of difficult to work with if you're just shipping a single container. So yeah, I think the supply chains are going to be dynamic. They're going to respond to labor price tariffs and everything else, but people, they're just going to, that's always the case, right? Like it used to be, we bought cheap stuff from Japan and then it went to Korea and then it went to China and now it's Vietnam, but it'll keep moving all the time. And hopefully the result is that people in these countries are making more money. I mean, if a country is suddenly not competitive because their labor cost is too high, good for them. Like their labor costs went up. That means human beings are getting paid more.

Well, you know, in thinking about those local economies, they're kind of going through an industrial revolution of sorts. So as their kind of economies emerge to saying, Hey, we're no longer kind of the cheapest place to make a certain item, that equally triggers different types of opportunities. Because to your point, the population is earning more, so maybe they're also spending more, and that unlocks different things for places like Vietnam, India, so on.

I mean, you hope so. That— I think that's the bet those— the politicians in these countries are making. It's been hard to get certain countries to spend more on consumption. The American consumer is just a special They're just a different kind of people apparently. Like we get, we spend more than we make and that fuels the economy. And in Asia they save more than they make, you know what I mean? They save more than they spend. There's a different culture there and that, yeah. And good for them, but it, and that savings has to generate investment and that can lead to overinvestment. We talked about asset bubbles, but maybe you build too much infrastructure. And right now in Vietnam though, return on infrastructure investment is very high. We're working, I mean, I love the fact that we do all this barge shipping, one, because it's cool, but it's like, hey, you know what? That's a very creative solution to the lack of infrastructure we got. We're working in the river ports and shipping in the Mekong Delta and stuff. It's pretty cool.

Yeah. Yeah. I mean, and in the US we can certainly learn from that because while we have an extensive waterway, I personally think that we, Still don't utilize it to the best of our capability.

It's utilized. The Mississippi River network is incredible. It's utilized a ton for ag exports, for energy commodities, bulk kind of bulk commodities flowing out that down out there, out through New Orleans. It's way underutilized for container shipping. There's basically almost none. There's like 1 or 2 services with a couple hundred TEUs at a time. And it's not used to my knowledge for, to kind of replace domestic trucking shipments, but it absolutely could be. If you go look at that map, you can cover, you could build huge lane networks, you know, right from Ohio, the Ohio River all the way out and up the Missouri River, up the Mississippi River, through the Great Lakes, down to New Orleans. I mean, it's an incredible network, very carbon efficient, very cheap. It'll be a little slower, but it'd be a lot cheaper. I think that's a pretty exciting opportunity, but it's a little bit asset heavy. I'm not sure it's venture. You want someone who's got a bit of a lower return, but lower risk. I mean, it would make money for sure. It'd be very interesting to build. Yeah. It'd be a fun company to build.

So my final question for you, we're sitting here at the start of the year. We have 12 months ahead of us. Any hot takes for 2024?

Yeah, well, in logistics world, It's always the case that you got to be ready for stuff that I will not be predicting on this podcast. It's a crazy world. 2 days ago, or maybe it was 3, I can't keep track, the Yemeni rebels, rebel army, these are not like just like ragtag guerrillas. They're like a well-funded army unit, civil war, fired missile and hit a ship. Most of these targets have been Israeli-owned or Israeli-affiliated ships, but this was just a random ship owned in Malaysia or something. I forgot. Not, definitely not. It was just targeting shipping. That is a very big deal. If you lose the, if the insurance companies pull out or dramatically raise rates for going through the Suez, you'll have to go around the Cape. For these big ships to go around the Cape, First off, just doing that takes much longer, therefore reduces capacity of the network. Also, they, I, my understanding is the ocean is so heavy, the seas are so heavy down there that you can't load the ships nearly as full.

It's one of those dangerous routes to kind of circumvent.

Containers start falling off. I've never been down to that part of the, well, I have, but not on a ship. And So that would, it only takes a handful of containers, 7 containers fill a 747, and there's about 600 cargo 747s in the world if my memory serves me. So it wouldn't take a lot of containers. You know, a container ship holds over 10,000 containers, 40-footers, the big ones. So it wouldn't take a lot of cargo diverted from the ocean because it's less reliable or more expensive to just jam the air cargo. network completely and drive prices to the moon. We've got a drought in the Panama Canal that's already shifted tons of East Coast volume to the West Coast. West Coast, September, I didn't see the stats for October, November, to be honest. Right. They're not coming to my mind. I'm sure I saw them, but September was the biggest month in the history of LA Long Beach. It's a big narrative violation if you're like, oh, trade's down.

Yeah.

But like, nope. All the East Coast volume is being diverted to the West Coast because of the Panama Canal. East Coast, you've got the ILA, which is the labor union representing the work, the longshoremen on the East Coast. Contract comes up and they're saying they will not work without a contract. I think it was September of next year that it's getting negotiated. That's a big deal, obviously. And I'm not sure what the carrier's view of it is. They might be willing to call the bluff. So that's all oceanside air related. These markets are related because when the ocean breaks down, people switch to air and there's not a lot of air capacity in the world relatively. The stat that I learned recently that kind of blew my mind the biggest was that between 30 and 50%, depending on the airline, 30 to 50% of the cargo getting moved from Asia to North America is e-commerce parcels. Right now. I mean, that's up a lot. And the leading company in that, Temu, is projecting, I heard, I think they published this, like over triple-digit growth, 100% growth. Some of these companies are public, TikTok. There's dozens of these companies, not to mention just the FedEx, UPSs of the world that 30% going is with that kind of growth, air freight prices, you know, if you're in that business, and a lot of that's peak season because these are kind of like e-com, therefore very much tend to be tied to Christmas buying season. You see very high air prices at the moment. And I predict next year, if you need air capacity at peak season, you're going to want to sign some longer-term contracts.

Contracts.

And you probably overpay during the slack season to make sure that you have that space. at peak because the e-com boom is crazy right now. So yeah, like, but hey, I bet you anything if we come back and come on, I'm happy to come back on the podcast whenever, but I bet if I come back 12 months from now, we'll be like, oh yeah, Ryan, you didn't even predict that thing. Because I don't know what's going to happen then.

With that, man, I certainly appreciate you making your first appearance. Hopefully not the last here on The Future of Supply Chain. We covered a lot of ground and I think ultimately for people, there's probably no more exciting spot than supply chain to be spending your time. To your point, if you're interested in business, there's a backstage pass to that.

It's a free MBA. You get to go, I mean, it's not free because you got to work your butt off, but you're going to learn more working in a supply chain company about how businesses are run than you would at any business school. You just will because you're seeing everything. First of all, Within your company, you're learning the sales and marketing and hopefully the finance and accounting functions. And then you're learning about everybody else's business too. And how are they making money? What the cost of inventory, you can see it real time. Oh man, this is expensive. And if it sits there unsold, it's really expensive. And I'm financing that inventory. And inventory is nothing more than dollars that have been converted to a different form. You can start to think about interest rates, success, like It'll teach you everything about business. And I don't think there's— the world is a super interesting place, but there's not a lot of disciplines that are more interesting than business. All righty, man. Cheers.

Thanks for listening. If you enjoyed this episode, leave us a 5-star review and tell us what you liked. And be sure to head over to podcast.dynamo.vc to keep up to date with our latest content or subscribe on the podcast platform of your choice. Until next time.