How to Build A Competitive Edge with Ryan Petersen of Flexport

Thank you, it's great to be here.

Yeah, it actually goes even back further than that. So in the early 2000s, I worked for my older brother. He's running an import-export company. We were entrepreneurs buying scooters, dirt bikes, ATVs, like dune buggies from Asia mostly, and selling them on the internet. And it was kind of like, it was pretty early days before there was no Shopify or Stripe. A lot of the problems that would be pretty out of the box, we had to build a lot of that tech ourselves. And then obviously solved the problem of sourcing products internationally and selling them. So we ran that business for many years. My brother was my boss, actually my first boss out of college, working for him and his business partner Michael Kanko. And that's when we first discovered this problem of like freight forwarding, international logistics being kind of a nightmare. It just shouldn't be that hard to ship stuff. You know, you solve the hard problem with like finding good quality products and finding customers for those products, like you would think getting it from that factory to those customers would be simple, but it really wasn't. The big pattern that we saw with everything we were shipping was they were all big, like motorcycles are large. And so we were really in the freight world, not this small parcel world. You couldn't ship these things with FedEx or UPS, you're shipping freight. And we just kept like suffering through that. And eventually I realized like I should start a company to be online customs brokerage. The first idea was sort of like TurboTax for customs, you might call it. Customs is tax, you have to pay your customs taxes, so make that super simple, like a wizard basically walk you through the process, help you with the documents, the clearance. And then we very quickly realized we'd have to sell freight too. So the early, early customers when we were just customs brokerage in the very beginning were kind of pain in the butt. Like if you're separating your customers from your freight as a really small business, there's something weird going on. So like I think the first customer we ever imported was this Berkeley professor who bought a tuk-tuk during his vacation to Cambodia, like one of those motorcycle taxis. I know exactly what a tuk-tuk is. That's amazing. Cambodia or Thailand, I can't remember. And like somebody shipped it for him but like didn't want to handle the customs. And we're like, I was like, well we know a lot about motorcycle customs, like we'll go for it. But it was never going to be a repeat customer in a million years. It was sort of a silly, but you know, you have to start somewhere. And we quickly realized like, hey, real companies that are doing repeat transactions want to buy freight and customs together, especially small business. Like big companies, they'll sometimes separate it for compliance reasons, but small business wants to buy those two things together. So that set us off on like a now 11-year journey to become the best freight provider in the world.

Well, at its core, you know, it's interesting because I was a user of these services, so I didn't even appreciate all of them. Like I just wanted to know where my stuff was, how much it was going to cost, what documents and paperwork and data I would need to get it into the country, be compliant with all the various laws and regulatory agencies. So that was my view, was like, well why can't anyone tell me where my stuff is, when it's going to arrive, how much it's going to cost, and what data I need to provide? Like, and that's what we set out to do without knowing that the underlying cause of that problem is that there are many disparate systems. Because if you want to move a container from door to door, nobody has the assets to do that all as one company. Wow. Like nobody owns trucks in every country plus warehouses, customs, big ship, door to door. Like now some companies are maybe trying to do that, but at the end of the day, like it's such an unwieldy giant company that would need to exist that they wouldn't be good at it if they could, and they don't really exist. So what happens instead is you have a layer on top of that called freight forwarding. And I joke often it should be called freight email forwarding because it's kind of like just passing documents as PDF attachments, you know, to email between all the different parties. And it just leads to a lot of holes. There's no visibility because, you know, there was no tag to track these things. It's kind of a people layer on top of the asset owners of the world. And so that was the status quo when we got into this industry. And we have been just steadily connecting all the parties via integrations. In some cases we build our own tech for them. Like we have tens of thousands of trucks on our mobile app to haul the freight around. We've got warehouse management software. If the providers don't have their own, we'll integrate with theirs, but if they don't have a warehouse management system, they can use ours. And then we of course integrate with all the various parties out there in the world that do customs, ocean freight, all these things. And then try to give the customer a single pane of glass to overview their whole supply chain. But it's, you know, it's like a 10-year slog and we're still, I feel like, just getting started with what's possible to bring automation into the game.

Yeah, I haven't thought about it in such a concise way, but it probably wouldn't be that hard to think through. I mean, the very first chapter is learning about the problem, you know, and that's what I did in the import business and going through the pain and being like, oh, this is actually a thing. And that took a long time because it wasn't obvious that we could solve this problem. It was just obvious that there was a problem. Two was, okay, I want to do this. It actually took me about four years to get the license from US Customs. And I probably could have gone faster if I knew more of what we, you know, if I'd been an experienced customs broker and had the relationships. But it took about four years to get licensed. So that would be the chapter two. Chapter three, we got the license, but I don't know how to do freight. And that's what I was describing earlier. We just cleared customs for weird mom and pops, like I said, the tuk-tuk example, there are many others I could point to. Chapter three started to professionalize this. Did Y Combinator, brought in much better engineering leadership. I'd been doing like the original product management and a lot of coding, you know, myself or with teams that I'd hired in the Philippines in the early, early days. So brought in a CTO who had been the co-founder at Yammer, and he started to really professionalize like the tech team. I brought in one of my business school classmates who emailed me. I did YC, I raised money from, I raised the seed round with Google Ventures and First Round Capital and a few other good investors. And I had my business school classmate reach out. He was like, hey Ryan, I read this article about you. Did you know that I'm Boston Consulting Group's like main partner on logistics and freight forwarding? I was like, oh no, I didn't know that. Like, and he'd been there for 12 years, I want to say. And they do this thing where you get like a three-month sabbatical. And he came to me and worked for free for three months, like a second path. So I knew at that point, like, I got to try, anyone who works on their three-month sabbatical, you know, for a startup is like, got to try to hire this guy. Assembling that kind of core team, it's a little different than a lot of startups where you like have the four buddies that are founders in the beginning. Like I was this more or less solo founder. My brother was a co-founder, but he was like running other companies as well, never was like full-time at Flexport. So I kind of got the seed round, got some traction, and then got great people, which I think is a good approach that more people should try, more entrepreneurs. And so that was, yeah, I would say that's the next chapter, assembling that team. And then it was really kind of trying to keep up with the demand. And that was probably like a six or eight year just like, wow, we really hit real product-market fit. We were growing like 10x or more in a lot of years. And in the physical world, that's like really hard, you know, it's not pure software. We're managing 10 times more containers one year than the year prior. Wow. So building the process to scale that, making sure the tech could scale was a long journey. And then maybe the next chapter is all the way up through, you know, raising giant, we raised some really big rounds of venture capital, $2.6 billion, you know, to date, lifetime. So somewhere along that line is just like venture capital scale, fund the journey, scale it like crazy, mega hyperscale. Then of course that brought in a new CEO and I became chairman of the board for about six months. And then that was a whole chapter, which was not my favorite chapter of the book. It's sort of the hero's journey, you got to go through the down part along the way. And then the comeback story, the last year where we've dramatically, you know, just gotten back to our old roots and gotten back to the growth and the culture, that entrepreneurialism that built the company. So yeah, it's like a lot of, really thought of it that way, but quite a number of different periods in the company's journey.

Well, first off, I came at this from the customer point of view, right? Like I was a customer. The reason we started the company, I felt like these freight forwarders were always taking advantage of me. One, they had no tech that we knew we could build better software. It was like, you know, 2013, SaaS proven, like every single industry could be made better off-the-shelf tech that you could implement for making UX better, for managing documents and data and all this stuff. Like we knew we could do a better job there. And then it was a culture of, tech alone wasn't solve the problems. It was like a culture of actually caring about your customers and playing the long game and being a values-based company that, yeah, we want to make a lot of money, but we're going to do that by earning the trust of our customers so that they give us more and more business and we scale with them. And this industry is super transactional. It's very frustrating still the degree to which freight companies and their customers engage in the sort of mercenary behavior of like short-term memory negotiation. Like it's, and I don't just mean like us, and I don't even mean us and our customers, but there's carriers that own the ships and the planes and the customers. There's a lot of kind of unnecessary adversarial relationship there. And one of the things that we've tried to bring as a newcomer to this game is this view that all the parties need to win. Like the ocean carrier and the customer should win and Flexport. And that's actually incredibly challenging to create, to shift the dynamic. We approached it from a simple framework which was like, probably inspired by some reading I'd been doing in biology actually around like, can we create like autonomous cellular automata? It's not even biology, it's like the merging of computer science and biology. I don't know if I'm pronouncing that word correctly, I've only read it. Automata or automata. It's these like self-replicating organisms that can be autonomously like grow and have a life of their own. And we wanted to build like, okay, so you have this problem of one, the law of large numbers, you're trying to build something, most things get worse when they get bigger. Yep. The most customer service experiences, especially like big companies, like 800 number call center, nobody cares about you. You like, I don't like big companies because they treat their customers like they don't matter. And yet I want to be a big company. So I wrestled a lot with that. And the answer that we came up with was like, let's have a team of teams, like lots of small companies, company within the company, and a little small cross-functional, we called a squad. I almost called it a cell, but Paul Graham told me that sounded like a prison cell. So but like the biological cell, like okay, this is like a team that can scale and it's cross-functional. So there's a sales person, operations people that are responsible for the customer success, customs brokerage that does compliance, and they would be able to really own that customer relationship, get to know them. And as you grow, you split the squads, and you know, and each takes half the book and they both grow from there. And so we kind of were inspired by Darwinian evolution on some level or some kind of biological emergence property. And that worked really, really well for us of like just allowing you to scale with, and you don't need a huge amount of top-down management in such a structure because the customer's the boss. They're telling you what they need. They'll be way more demanding than the boss would be. Like, and if they're really tied to the KPIs and the outcome for that customer, it's worked remarkably well.

In general, the big question I think that was going to affect this industry and the wider economy is just, is globalization, is that going to continue? Was this just a post-World War II 70-year run and now we're going to go back towards people buying things domestically or in their local region? Or is it still like ever since World War II, the US Navy has guaranteed freedom of the seas and anybody can sail a ship anywhere in the world and buy and sell stuff. That's the nature of globalization. And that's not just challenged by politics, but by geopolitics, or you know, with the rebels in Yemen firing missiles and taking out container ships and other kinds of ships that are going through the Red Sea. The most vital maritime choke point in the world is now cut off effectively from container shipping. And that was like, the US Navy tried to go down there and show them who's boss and like wasn't able to stop it. So that's a major development. There doesn't seem to be political will in the United States or elsewhere in Europe and other places affected by this to go solve that problem. So it's sort of related, like it's an extra tax on trade. Unfortunately, it's a tax where nobody is not going to some government to benefit, it's just a tax. That's the biggest one, is what's going to happen with globalization. Tariffs are maybe related, but a close second. And then obviously what's happening with AI is just incredible. And we think we're going to probably be the industry that's going to be the most affected, either positive or negatively, depending on if you're the ones applying it or not. But you know, I was describing that freight email forwarding, there's a lot of manual workflows that go into coordinating these transactions end to end. And it turns out where AI just like does a lot of this stuff once you structure the workflow, which we've spent the last decade building structured workflow tools. So we're currently automating about 1% per week in our freight forwarding business. Wow. Yeah, it's pretty crazy. So we're, by end of next year, we feel like most of the manual workflow, the majority of the manual workflow will be done in an automated fashion by either AI or just well-written scripts. And that will lower the cost of shipping anything by ocean by about 5%. Wow. Or we make 5% more margin, which 5% more margin in a 5% EBIT margin industry is pretty crazy. So that's going to be a huge trend. It also is allowing, AI is going to be much, much better at solving some of the classic problems of route optimization, sort of traveling salesman type problems, but inventory management. How much inventory should you order in the face of uncertain demand? Where should you put the inventory? We're finding that AI models are better at this than certainly than like the random person doing Excel planning at a customer would be. Totally. And we acquired Shopify Logistics last year as well as Convoy, which is a trucking technology company. And that gives us the true end-to-end, like from factory floor all the way to customers' doors. If you order from a lot of Shopify sites, the product is delivered by Flexport. Wow. Like fulfilled by Flexport. And so now you're saying, okay, how do I apply AI across the whole spectrum? And there's a lot of amazing AI companies out here, but like they don't have our advantage of just use the AI, we have all this proprietary data from a decade of doing this, we build our own code, so we can just apply it and then go tomorrow and go live. I don't have to beg anybody for permission to use the AI. You're going to see amazing stuff on the UX here. It's going to bring down costs. Will it bring down costs as fast as tariffs and other stuff increases? Maybe not, but it's still going to be a pretty important trend.

Yeah, I mean, I think I give my mom a lot of credit. My mom's a very successful entrepreneur. She built and sold two companies. She's a biochemist first and foremost. I'm not even sure if she calls herself an entrepreneur. She's a biochemist, but she sold two companies in the food safety space. Not like, you know, Silicon Valley exits, but she did very well. And she would make me and my brother too, we would earn our allowance by delivering sodas to her office. We would stock the fridge at her company and then we had to write software that would generate the invoices. We were like eight, nine. Each of me and my brother both did this like at different ages. So I had to learn a bunch of skills about, I mean, it wasn't like hard work, but you're like kind of like generating an invoice. I have tax returns from when I was like eight years old. I don't think I had to pay taxes, I think my income was too low, but I had to file the return. And then she was getting a tax write-off for paying my allowance, you know, company paying my allowance that way. So that was creative. But I would give more credit for my entrepreneurialism to my older brother, who is just like a supreme entrepreneur. Like, he may be unemployable, I think, like even more so than me. Like most of us are. Yeah, yeah, probably. But he has extreme self-confidence, like believes that he's better than everybody else at business. Not like in life, but like doesn't believe that there's some big company that's better than we are at doing stuff. Like he's always kind of challenging. He was running a business with real hustle. My first job was just working for, I mean, I worked for Domino's when I was in high school making pizza, but my real first job was working for my brother. And he was the one that was taking the risk there. Like he had been a computer programmer at Intel and just used his salary to buy motorcycles in China and start selling them on eBay. Like that was, and he's a programmer, so he built all this software for e-com checkout, payments, inventory management, all this other stuff. But so I kind of learned the ropes from him.

Yeah, you know, it's been a journey. And one thing I've learned is you can't have a pure bottoms-up or where everybody just self-organizing does whatever they. I tried that, it's a nightmare and disaster, and you get a lot of bureaucracy from like not knowing, you know, you need tops-down guidance. The biological analogy here is like you have a nervous system and an immune system. Yep. The immune system is like bottoms-up, self-organizing, it sees a virus and attacks. But like you need a nervous system telling everyone where to go and what to do. So I think I was a little, I was too light on that in the early days, too much like, ah, the customers will tell everyone what to do, just like let the thing organically go. I've become much more of a tops-down, tell people what to do or at least where we're going. Much more comfortable making a lot of decisions, but micromanagement in some cases, I don't think that's a bad word. I think being in the details, as we call it. But that quote you said is just copied from Jeff Bezos, by the way. The type one, type two decision of like, is it reversible or not? If it's reversible, you can definitely want to move fast. Having that framework is important. Having a good decision-making framework for any large company or growing company, like who can make what decision. We use RAPID. Like there's someone who's recommending the decision, the A is someone who has to agree, you're not allowed to put anyone there, the P is who's going to perform, I is who's providing input, let's get input from lots of people, and then the D is the decision maker. And the D, the key question for everybody, the first question is who is the D, who is the decision maker? If you can't figure it out, escalate. And if two people say they're the decision maker, find where in the org chart those two people meet and they're the decision maker. Like, and that, you know, if you can get a bit of helping people, and it's sad that you need these kind of things in big companies, but sort of the nature of it. One of the things that we've started to do is this, where AI is going to be really powerful, is getting summaries. What's going on with every customer, for example, and having the account manager who's responsible for that customer just write up a report about every customer and put them all in a RAG and have the LLM summarize it and tell me what's going on with the customer base. Of course we have managers and stuff, but just summarize it and tell me what should I be focused on, what's growing, what's not growing. Like that works beautifully. So I think there's more you can do with this stuff than just build automation and customer features.