Why Flexport CEO Ryan Petersen took his company back

Today, I’m talking with Ryan Petersen. He’s the founder and CEO of Flexport, which makes software to optimize shipping everything from huge containers to e-commerce deliveries. It’s a fascinating company.

We actually had Ryan on last year, but this episode is a little different: since that last episode, Ryan handed off the CEO role to 20-year Amazon veteran Dave Clark, decided it wasn’t working out, and then fired Dave and returned as CEO just a few months ago. This all happened inside of a year. I always joke that Decoder is fundamentally a show about org charts, and Ryan and I got deep into why he made and unmade the biggest org chart decision there is.

The last time Ryan was on Decoder, he was excited about the transition. He told us Dave knew the business “like the back of his hand,” and that appointing him would let the rest of the company become more “innovative and focused.” 

That didn’t happen. Flexport laid off 20 percent of its employees in January, right before Ryan left and Dave fully took over. Clark successfully oversaw Flexport’s acquisition of Shopify Logistics in June, but things unraveled quickly after that. By the first week of September, the board asked Clark to step down and Ryan came back.

And it hasn’t all been smooth sailing for Ryan since his second time at the helm began. He’d only been back for a few weeks when the company laid off 20 percent of its workforce — again. He’ll be the first to tell you that the shipping market is volatile and carving out a niche in it is hard work. You’ll hear him explain that even though Flexport is something of a tech startup, he’s sure the way to success is through old-fashioned customer relationships.

You’ll also hear him explain why the market is so volatile, and it’s really interesting stuff. It all comes down to interlocking systems that have to work together in order to get things around the world: Rainfall in Panama, unrest in Yemen, and changes in Russian airspace all affect how the goods you order online can get to your door and how much it costs everyone along the way. 

Flexport, at its core, makes software, and you’ll hear Ryan explain that work. But moving stuff around the world is a crazy patchwork challenge of overlapping systems, and Ryan tends to look at things very systematically. He’s got a fascinating perspective on the world’s economy, and you can see how that shapes his view on what it takes to run Flexport itself.

Okay. Ryan Petersen, CEO — again — of Flexport. Here we go.

This transcript has been lightly edited for length and clarity.

Ryan Petersen, you are the founder and CEO — again — of Flexport. Welcome back to Decoder.

Thanks for having me back.

I’m really excited to talk to you. A little over a year ago you were on the show. A lot has happened in that year, including a global supply chain crisis coming to some kind of resolution, which I want to talk to you about, but then a lot has happened at Flexport itself. I’m just going to go through the list. You were on last September. You had hired Dave Clark as CEO, he was your co-CEO. You were about to step down and become executive chairman.

That happened, so then Dave Clark was the CEO. Then Dave left the company — we should talk about the circumstances of that — and now you are the CEO again. In the meantime, you’ve made some acquisitions. There have been some public statements back and forth with Dave about him coming and going, and you’ve had two rounds of layoffs, and the world still needs to move containers around. That’s quite a lot. Did I miss anything?

Man, that last part, the world needs to move containers around. There’s been a crazy volatility in the freight market, so it’s definitely interesting for companies out there to understand what is happening, what to expect over the next year or two.

Let’s get into all of that, but start at the very beginning. Give people the one-sentence explanation of what Flexport does.

Flexport’s a global logistics provider. So we make it easy for companies to ship anything anywhere, in any quantity really, any product, whether it’s a small parcel to multiple containers or hundreds of containers, thousands of containers. Some of our customers from factory floor all the way into customers’ doors or into retail stores as well.

And just to be clear, Flexport makes the software that makes that go?

We make it happen, and the software is — the reason it works with high quality is that we build software for every leg of that chain. Fundamentally, our business model is customers pay us to move product, but what differentiates us from every other logistics provider is that we are a technology company. So we’re building software to make it more seamless so that the data can flow in parallel to the goods, and now increasingly so that capital, the money, can flow as well. So supply chains are about three core flows: it’s the flow of the product, the flow of the data, and the flow of the cash, and building tech to make that seamless for everybody. That’s what Flexport does.

I encourage people to go listen to the previous episode with you, because it is one of my favorite episodes of Decoder. But you don’t own trucks or ships, you connect to those providers through software?

Yeah. And the reason you need a company like us is that the people who own the assets fundamentally, a ship goes from port to port, but the goods have to go from door to door. The planes stop at the airports, but the goods have to keep moving. And the freight forwarding layer is responsible for clearing customs with the governments, coordinating that truck to come pick it up, maybe putting it on the rail.

We do a lot of barge shipping in various parts of the world — China, Vietnam, and Europe especially — moving containers on the barge to and from the ocean port. So it’s a complex coordination problem, at the end of the day, and this is where software really helps because you’re getting many parties. A typical transaction, moving a container or an air freight shipment, you’ll have at least a dozen companies do something in that chain.

And then more once you start counting the governments that are involved, customs agencies, and others. A bank is usually providing some kind of capital for the transaction. An insurance company is underwriting the cargo risk. So it’s just a lot of coordination that has to happen. I mean, if it was just buy low, sell high, put you on a container ship, you’re just not adding enough value there to justify your margin as a freight forwarder.

One of the things we talked about last time, and I just want to continue to set the stage before we get into it, was that a lot of the process had been very manual. It was a lot of forwarding emails, filling out spreadsheets, making phone calls. Your software automated that, made it clearer and easier for all these parties to coordinate. But getting uptake on the software, going to all those different parties and saying, “Use our software to coordinate,” was one of the big challenges. Is that still one of your big challenges, or has that receded? Are you more established in the market now, how’s that working?

I think the reason that we’ve been successful where many technology companies have failed in this space — or failed or struggled to get traction or struggled to take off as you’re implying — is that we are first and foremost a customer solutions company. Like, “We’re going to solve this problem.” And if that means picking up the phone to call the trucking company or the port or emailing somebody, or go meet them in person and take them to lunch and figure out what needs to happen to make this better, we’ll do it. And then we come through over time and build the software to solve their problem, that asset owner’s problem.

So trucking is a great example. When we started, on every container we moved the stat was 42 minutes of labor of coordination with each truck, with each trucking company on every container. So we’ve now built software, a mobile app for the drivers. We’ve got a dispatcher app for our trucking companies that serve us in the US and in Europe, and we’re bringing this to more locations. We don’t pick up phones and email them unless there’s an exception.

They use our software, they see loads coming, they assign it to their drivers, and you build that automation over time. The same is true with our warehouses where we have an amazing warehouse management software system deploying into partners’ warehouses, not just our own. So we don’t even have to run warehouses all over the world, but you get the visibility, the control over those assets, even if you don’t own them. And that’s the ultimate goal in owning assets is you don’t get the return on capital that technology investors are looking for from owning assets. So we try to avoid that wherever we can and build software instead.

You have two big acquisitions this last year: Shopify Logistics, or Deliverr, and then a digital freight network called Convoy. How do those integrate? How do they extend the vision of what you’re doing?

Shopify Logistics actually is a major acquisition we did. Over a billion dollars in equity value; it was a stock deal. We love that deal because it extends us all the way to the door. The way to think about Flexport previously was we would take cargo from factories around the world and deliver it out to warehouses and distribution sites, fulfillment centers. This adds those warehouses and distribution centers so that now we go all the way to customers’ doors and into the retail network.

So we have integrations with the 15 biggest, most important big-box retailers in the US. A nice roadmap to keep adding more companies that we can deliver to. So that really makes you enterprise-grade, where you can go end to end for a customer. For me, it was a part of the vision that I had sort of been like, “Well, yeah, of course we’ll get there in 10 years.” There was enough to focus on in the freight forwarding world. But the reason it’s very exciting is that you now have data for those customers that are using the full end-to-end and, increasingly, we’re cross-selling this thing across both customer bases.

You’re able to see customers place orders to their factories through Flexport. So now I know, alright, how long is it taking to produce those goods? Place the order at the factory and the factory becomes a user of our system, then places a booking to come pick it up. So I know the lead time from when you order goods to when they’re produced. We are, of course, responsible for transit time into your warehouse network, so that’s your whole supply side. You can call it supply-side latency. How long does it take to get goods made and delivered? Now you have demand-side data: what’s selling, what’s going out, and where are those end customers?

So you’re now in a position to go, “Wait, we could start making suggestions to customers.” Not just about routing of goods, but about how many goods to order, when to order them, which fulfillment centers they want to be in, in order to optimize their business. And by optimize the business, I mean how much inventory you keep on stock relative to the orders that are coming in. Because you have a fundamental tradeoff that you’re making as a brand. If you have a lot of inventory, you can have really fast ship promises, and you can deliver next-day, same-day delivery. If you want to do same-day delivery like Gopuff style, you’ve got to have goods in every neighborhood all the time. Very expensive to have that much inventory there. If nobody buys anything, it’d kill your business.

If you have way less inventory, you’re going to have very unreliable ship times. In fact, it might be three weeks, a month, you might have to go produce the goods on demand, like furniture companies sometimes take months to deliver your stuff. So you’re now in a position to help companies optimize that tradeoff. And even what we’re working toward is going, “Hey, if you’re a Shopify merchant, you should be never out of stock again, as long as you’ve placed the order back to the factories.” Instead of saying out of stock, it’ll go, “Hey, ships in three weeks,” because I can see it on the ocean. I can see it coming in.

And so you start to be able to be a growth engine for these companies. Nobody wants to be a cost center. We’re a cost center today. If I could be a growth engine, I’m talking to your marketing people, I’m talking to your salespeople like, “Hey, how can I help you grow your business?” Well, selfishly, you’ll put your nicer employees on it. What’s the personality profile I’m going to hire and put in charge of talking to my freight provider? It’s not like if I’m running a business, I’m not going to find the nicest guy in my company to do that job, I want the bulldog who’s going to go beat people up on rates. So it lets us talk to the friendly guys who are trying to grow the business.

That’s a great way to optimize hiring and structure, “What kinds of personalities do we need? Here’s the cast of Cheers. Who’s the grumpiest one? Who’s Ted Danson? How are we going to assign these roles?”

I love our customers, and they are supposed to be discontent. That is what they’re hired to do, right?

That’s a big vision, right? “We’re going to not only optimize how people use freight to move goods around the world, we’re going to take the data back out of that and help them optimize their businesses and help them have better relationships with their customers.” How does Convoy fit into that vision?

Convoy was a different type of acquisition, and I probably couldn’t have afforded it, but they ran into some trouble, ended up being a distressed asset, and so we got a great deal. We didn’t acquire the company, we acquired the technology and a core group of about 50 of the top engineers that came over. They had built incredible technology for truck brokerage. This is full truck load domestic within the United States, trucking. Something that Flexport did, we have about a $270 million top line revenue business in trucking, if you count all the types of trucking that we do.

We did, last year, about $3.8 billion in revenue. So for us it’s a pretty small piece of the overall puzzle. That said, every container we bring in eventually becomes a truck move somewhere down the line. It’s got to move by truck. Trucking is as big of a market as international forwarding. Domestic US trucking is just a massive market. Convoy built incredible tech. We have now 400,000 drivers on a mobile app, which I hadn’t thought that much about this space because it’s kind of adjacent to what we do.

What I was realizing as I got to know their technology over the weeks before the acquisition is there’s an incredible amount of tech that has to go into doing that compliantly and safely and securely. It’s like Uber for trucking, but you’ve got to do background checks on these people to make sure they’re not going to steal your cargo. So it’s like doing six different background checks before every single load. They have to have the app on while they’re driving. If they veer off the suggested route, it sends an alert to the physical security team being like, “Hey, is this cargo getting stolen? What’s happening right here?”

So they have best-in-class tech for that kind of stuff. Customers don’t necessarily see that, but they appreciate it once they learn about it. And all these routing algorithms, the margins were really good on the business. They just didn’t get it to the scale required to cover the fixed costs. We only took about 10 percent of those fixed costs over. So I think it’s going to turn out to be an amazing acquisition. We’re working with all the old customers they had to go, “Hey, come back. You can trust us. We’re going to light this thing back up.” So I’m pretty optimistic.

And from a vision standpoint, what that looks like for Flexport is sort of like, “Hey, one-stop shop. Ship anything anywhere, any quantity, any mode.” Today, both of those businesses — the Shop Logistics, which we’ve rebranded as Flexport Omnichannel, and Convoy, we haven’t rebranded that yet — but both of those are US focused. But I think you’ll see us bring those products globally.

Especially on the Shop Logistics thing, I want to be able to go to a brand in the US and say, “90 percent of your sales are in the US, but if you look at your Instagram followers, it’s 50 / 50 global and US. Why aren’t you selling on Flipkart, Rakuten, Coupang, Mercado Libre in South America? Didn’t you know I can light you up? We’ve got the international piece, we’ve got the customs, we forward deploy the inventory into fulfillment sites down there, and just help you grow your revenue.”

Again, I’m talking to the nicer people in your company. If I’m growing your revenue, that’s like a different conversation. We don’t do that yet, today it’s just US. But vision-wise you’ll see that from us over the coming four or five years, start to take that service global.

This brings me to some of the drama, right? You were talking about vision. In particular, the last time you were on the show you had Dave Clark as your co-CEO, you’re about to step up to executive chairman, let him run the whole thing. And what you said to me was that you wanted to be focused on the customers. You’re a customer guy.

You wanted to go work with the customers, identify their needs, and be in that role and let Dave operationalize the company. That happened. He is the one who drove forward the Shopify Logistics / Deliverr acquisition, which it sounds like you think is pretty transformational for the company. But then it stopped working out, and now you’re the CEO again. What happened there?

The first piece is the most important one of talking to customers, and I think the fundamental disconnect is e-commerce logistics versus consumer, which is a consumer proposition. Think of a major e-commerce marketplace called Amazon. You’re not talking to your customers. If you talk to the customer that’s a major defect, something went wrong. You’re supposed to just fill out a web form and the thing gets delivered. B2B, freight forwarding, global logistics, and just generalized B2B, you’ve got to talk to your customers constantly because every business is unique.

As human beings, of course we’re all unique, but when it comes to shopping, we’re not that different. You want to click a button and have it show up. There are not that many flavors to it. And therefore you can just kind of make a web form and not talk to the customer. B2B, the business needs are so diverse based on is it a wholesaler, is it a retailer, is it a [direct-to-consumer] commerce versus selling through other people’s retail stores? Are they just a manufacturer and importing raw material?

The number of permutations is really, really large, and there’s a lot of SOP, standard operating procedures, that go in there very differently. And frankly, relationships matter so much more. You’ve got to be talking to customers nonstop, and that was just not happening at the level that’s required to grow the business, in our assessment, my assessment. So we weren’t growing, and it was pretty obvious that that was the reason why, because that leadership needs to get out there and talk to customers constantly. And so that’s the main thing. It starts at the leadership: you lead from the front, get out there, talk to dozens of customers every week, get the feedback, solve the problems that you’re hearing about.

So I think that was the biggest disconnect there. And we increased our cross bases too much. I think probably some bad habits accrued from a world where capital was infinite. We have a great balance sheet, but we don’t have infinite capital. And so I think some big company behaviors, we were just spending too much money. So those are kind of the two big course corrections I’ve had to make over the last… I’ve been back in the seat now for three months, cut costs like crazy, letting go of a lot of nice-to-have things. We let go of people, which always sucks. We let go 650 people, and a lot of awesome people, a lot of my friends even. But you’ve got to live in reality with the P&L and build a profitable business.

So we probably over-hired a little bit, hired some kind of nice-to-have big company type stuff that we’re just like, “We got to be a scrappier, leaner startup.” And our customers want that. We’re not selling Gucci handbags, we’re selling logistics services. You’ve got to be affordable. So if you build up a cost basis, your prices become too expensive for your customers. So we’ve been really focused on lowering our costs so that we can be affordable for customers, and at the same time engage more deeply with the customers. We’re not going to live in a world where you just tell a big enterprise customer, “Hey, just fill out this web form.” Talk to them and solve their problems.

It feels like the cranky guys that you’re talking about don’t want to see you have nicer offices than they do.

No, oh 100 percent. Yeah. By the way, not all of this is by any means on the Amazon team that was here. Well, actually a lot of folks from Amazon still work in Flexport. They’re awesome. But a lot of this is 10 years of Flexport, kind of like zero interest rate capital came very easily for the company. We’ve got fancy offices, I’m in one of them right now. And you’ve got to be much more cost conscious over time, and you’ll see us roll off leases into more affordable office space, et cetera.

I want to get to that, the cost-cutting and how you’re structured now, the Decoder stuff, but I just want to stay on the CEO transition piece for a minute. I always joke that this is a show about org charts, and this is some deep org chart stuff. So you hired Dave from Amazon to come operationalize the company. And this is the thing that you told me that you wanted to have happen, that you wanted his operational excellence from Amazon to come in, that Flexport was established, that you were growing.

You want to run the company faster and you’re going to go talk to the customers. This is the big discrepancy: at some point you say, “Dave’s not talking to customers enough, we’re not growing as fast as we should.” And Dave is on the record in some of the coverage of this saying basically the same thing that you’re saying, but at an earlier point in the history. The problems were “much more extensive than I thought they would be when I agreed to join.” And the company had missed cost margin and revenue forecast for multiple quarters before his arrival, and it needed process and financial discipline.

So his view of this, from these quotes, are that first he needed to get the house in order before he could start growing again, and your view was he wasn’t talking to customers and you weren’t growing. There’s overlap there and then there’s differences. What do you think the overlap and the differences were?

I don’t think what Dave said is wrong. Yeah, of course we’re a startup, you’ve got problems, you’ve got to solve things. The financial discipline, for sure, I mentioned that some of this is our own 10 years of zero interest rate, well-funded. We’ve raised $2.3 billion over our life cycle. So we’re well-funded. And I’ve said this in other places, like companies that raise too much money end up spending too much money. And I don’t think we raised too much money, but we spent too much. It’s like how do you create that? I’ve never seen a company figure that out.

One thing I’ll say about that, there has been some drama, that’s undeniable, but it’s not actually real to a large extent. A lot of this is, well the demand for scandal in the world is much higher than the supply of scandal. And you just get kind of like, “Oh, you get a couple journalists in here and they’re just going to play it up to a big degree.” And in fact even fed the cycle where they were saying things to Dave as though I said them. It was someone else who I don’t even know who it was.

Even the journalist told me it’s an ex-employee, which I assume means somebody who got fired by Dave, talking crap, saying things about Dave, but then playing that off as someone close to Flexport said this. Then Dave thinks that we’re trashing him when we’re not at all, and Dave’s overreacting or reacting to that. I don’t even think it’s overreacting. If I thought my ex-employer was talking crap about me, I’d be like, “Hey, screw those guys.” So I totally see where he’s coming from. But the reality is there’s actually not hard feelings there. 

Dave’s points are accurate, but I’m not going to say anything bad about him. I just think, “Hey, you’re a CEO of a company, you got a lot of problems. Cool, let’s go fix them.” You know what I mean? And that’s got to be the attitude. And we definitely overspent. And I approved these plans, by the way, I’m not blaming Dave for this, I was all-in on it. We took our tech engineering org from 450 people at the beginning of this year, 2023, to when I came back there were 1,300 software developers in the company.

I mean you kind of add 900 people to an engineering org in one year. I think any engineering leader out there would tell you, “Wow, that’s very hard to maintain culture and actually get productivity out of a team when you grow it that fast.” And not to mention it’s just incredibly expensive. So a big part of the cuts that we did was I cut that team back in half, but still more engineers than we’ve ever had in our history as a company, until June of this year or something. I mean we’ve just very aggressively ramped it, and that’s bad discipline. It’s bad for culture to over-hire and then rapidly course correct.

And I don’t blame Dave for that because I was the chairman of the board, I had to approve those hiring plans. I’m very excited about our technology roadmap and I think I’m like, “Yeah, let’s go build it now.” Living in reality with the financials. And what has happened, which is out of our control, is the price of freight is down 80 percent year over year, and our revenue is kind of…  price times quantity is our revenue. So our revenue is only down probably half year over year. So a non-freight person will look at that and go, “Oh man, this company’s in distress. The revenue’s down by half.”

But actually volumes in the industry are down 20 percent. Flexport volumes year over year are up about 12 percent. A non-freight person looks at a technology company and goes, “You grew revenue 12 percent, that’s pretty weak.” But if you take market share, if your volume’s down in the market 20 and you’re up 12, you’re definitely the leading company in taking market share. But the financial reality is, “Hey, you probably can’t afford to 3x your engineering team in a world where market price is down, your volume’s only growing 12 percent. Your revenue’s under pressure.” But no, there’s way less bad blood than the media would let... The media wants to make it look like Dave and I are in some kind of war. It’s not like that.

Well, you fired the guy. I mean that’s the part I want to ask about. The mechanics of it are, you hired the CEO, you evaluated his performance for a fairly short amount of time, and then you asked him to leave and became the CEO again.

Yeah.

I wouldn’t say that’s OpenAI levels of drama, and it doesn’t seem like your entire company revolted and you needed to get rid of the board of directors. It’s still a level of drama in a world where, you are correct, where there’s high demand for drama. Walk me through the mechanics of your decision. When did it click in your head, “Oh, I need to make a change here”?

There’s not one moment — and by the way, it wasn’t a decision that I made. This was a unanimous vote of our board, which has a lot of great established leaders on it with a lot of experience. A couple of bad quarterly results in a row where we’re blaming macro. When you dig in, talk to customers, it’s clear like, “Oh, you have to fix some quality issues.” Quality in a logistics business about on-time performance. It’s about quality of data, responsiveness when they ask for a question or request a quote. Some of our quality metrics weren’t there.

We’ve made massive improvements just over the last three months as we focus on them. But you’re sort of making excuses on the one hand, and then just cost basis was too high. It was clear to everyone, including Dave, “Hey, we’re going to have to make some big cuts.” And I felt, and the board felt, that if you’re going to make some cuts to your headcount, and also that your culture is a little shaky — we’ve had a lot of new leadership, a lot of core Flexport folks were feeling underappreciated perhaps. There was some shakiness in the culture, let’s just say that.

And the culture for Flexport is all about, “Hey, are we out here engaging with our customers, solving their problems constantly?” Some of that wasn’t happening. And so if you’re the CEO cutting people and saying that the culture is the number one thing we’ve got to improve, it’s like a really hard message for a hired CEO to deliver. It’s a way easier message for a founder because people know I care deeply about the culture and the people here. It’s a lot of my friends, we built this company together over the decade. It’s still a very hard message for me to deliver, and it’s been a challenge to go, “Yeah, I care deeply about the culture. It’s my number one priority. Oh, and 650 people are losing their jobs.”

That’s pretty hard. It’s easy to come off as a fake BS person in that, and we felt pretty strongly at the board level that it was a better job for the founder to come do. Also, when it comes to growth, we’ve got to grow faster. Over my decade leading this company, we were one of the fastest growing companies in the world. Not just freight companies, like companies, period. So I feel very deeply I know how to grow this business in ways that over the last year we just didn’t. We grew 12 percent, but I can grow this thing. We’re going to grow way, way faster than that in the years to come. So just a track record of like, “Hey, what does the company need right now?” A stronger culture of engaging with customers, faster growth. These are things I’m good at.

You mentioned two things here that I just want to put together and ask a question about. One, obviously the zero interest rate phenomenon of many, many startups and overspending, that’s real. I think we see that throughout a lot of companies. But then two, you mentioned that Dave was from Amazon, he brought a team from Amazon. Some of those people still work there. Amazon’s culture at the beginning was very much a zero interest rate culture.

This is a company that famously was telling its own investors, “We’re not going to make a profit. We’re going to keep reinvesting in the company. Go pick another stock if you want easy profits. This company is built to dominate.” And they spent a lot of money and now they’re a dominant company. To the point where you could say the Amazon playbook is to spend your way to a monopoly position.

Whether or not you think Amazon is a monopoly, you spend your way to a dominant position. Do you think Dave was running that playbook and then the change away from the zero interest rate world undid that playbook and you had to pivot? Was that the right playbook? Do you think he was doing something else?

To be honest, Dave’s vision was really compelling. Dave’s strategy was mostly my strategy, I approved a lot of it. I think if you had an extra couple of billion dollars on the balance sheet, it’d be more reasonable. And so perhaps, I think, wish there was a Vision Fund 3 with $100 billion so we could see that world and live in that reality.

Taking that Vision Fund money is kind of a kiss of death for a bunch of those companies.

Well, it served us well. They’ve been great partners for us. We only get one timeline, so we don’t get to live in the alternative world where we let Dave play out his strategy. I would love to see that play out. I would love to go for it. He was going big. We’re still going big, but he was going bigger and bigger. It would be epic if all of Dave’s plans were right, and he’s a very credible guy, and he could be right. You know what I mean?

It’s possible that reality was the thing. It would’ve been truly epic, and would love to live in that world because we’d be just like a dominant, dominant... Probably nobody’s going to build a monopoly in global logistics, but we would be a very big and powerful force for the world. But there’s also a probability in that that you just fail, and that’s not a probability I’m willing to live with. Flexport’s never going to die.

We’re going to be a company that’s a generational company, but part of that is you’re not betting the company. And I think the levels of spending that we were going and we’re still in cost-cutting mode, we’re still looking at every expense, reducing our OpEx, growing our profitability every month. I think as the founder of this company, this is my life’s work and I plan to do this for the rest of my life. There’s not a world where we’re going to bet the company on any kind of timeframe.

Let me ask you about the Vision Fund, actually. They are an investor of yours. The joke I was making is Vision Fund’s famous investments are Uber and WeWork, where they did spend a lot of money to try to get to monopoly positions, like there won’t be taxis, there will just be Ubers. There won’t be landlords in New York, there will just be WeWork. That was the playbook, right? You’re going to spend your way to a dominant position in the market. Was that ever a part of your playbook when you were talking to Vision Fund? We called it Dave’s playbook, it would be epic. Was that a pivot?

Well, at the end of the day, Vision Fund’s a major investor and a great partner for us, but we make the decisions in the company at the executive team level, not Vision Fund, and they’ve always been super supportive of the decisions we want to make. My read on it was like global freight forwarding, it’s very dangerous to just go try to subsidize, let’s say.

I don’t know WeWork’s business model that well. I love the product, by the way, I like renting short-term offices that are flexible. We have six WeWorks around the world in some of our smaller cities. But Uber, they were famously subsidizing rides to get to monopoly share type strategy. Very, very dangerous in logistics. People are pretty mercenary. If you’re cheaper, they’ll switch to you and the moment you’re not, they’ll switch off of you.

So you don’t want to win on price, you’ve got to win on the quality of the service you provide, the data that you can provide, the ability to get lock-in through compliance services and customs. It’s a way stickier business if you’re selling on value. But if you’re coming in and just discounting on price, you’ll find as soon as you’re not, you switch. By the way, Uber may find that, right? That might be the case if people just switch over to Lyft as soon as… a lot of people have both apps and just switch to whoever’s cheaper.

I race Uber and Lyft all the time. I do it unrepentantly. I open both apps and I see who’ll send me a car first.

That’s admirable behavior. I’m lazy, I just use Uber. 

What’s interesting is, it’s marginal cost. It’s time. It’s who will send me that car first.

Yep. And logistics it’s kind of famously like, “Hey, you don’t want to just come in and discount on price.” It’s a dangerous... And by the way, you’ll often end up in a rat race to the bottom because what a lot of freight providers do is discount heavily, and then kind of boil the frog. Because rates are very volatile, and so they’ll tell you about when the rate goes up, of course, the cost goes up, but they don’t call you when the costs go down. So they find ways to make you think you’re getting a deal. It’s a more nuanced industry, I think, you’ve got to be careful with it.

The last few on the company stuff, then I want to talk about the shipping industry at large and what you’re seeing right now as well. Take me into the meeting room. I mean, I think of the audience for Decoder as every business school student in America, everyone wants to run a company. These are the moments for you as a founder, as an executive, as the chairman of the board. These are the moments that no one gets to see really. You make the decision, you’ve talked to the board, “I’ve got to make a change. I’m the founder. I’ve got to step back in here and preserve the culture.” What’s that conversation like with Dave?

Oh, it was pretty straightforward.

It was on the phone, is it on Zoom?

We weren’t in the same city at the same time, so we did it on a Zoom call basically.

He’s a real professional. He just wanted to know, “Okay, what’s the path forward? How do we land this news? How do we communicate this effectively, externally,” et cetera. It’s low drama. He understood, it’s a professional business, so there’s not really a story there.

Literally, as I think about business school students, which is a huge part of our audience. Did you make a calendar invite? Did you have a heads-up? No one knows how to do this stuff. You just have to do it, right?

I mean, you fire people. If you run a company, this is pretty much you have to do it. It sucks. It’s the worst part of every manager’s job, but you just have to do it. Personally, I’m a believer in just being extremely clear about what’s happening.

So then after that there’s the back and forth, whatever’s going on in the press. How are you holding things together inside the company as those reports come up?

Well, actually internally the vibe was incredibly positive from day one. Of course I’m going to get a biased sample; people who are happy about me coming back are going to be the ones I hear from more often. But you could see it on Blind, you could see it in a bunch of different channels. I’ve got scuttlebutt, I got people telling me what’s the real word on the street. People are very positive because they saw some of the changes that were needed. They were pleased about a lot of the decisions that were made because one of the first things I did was come back and reorg the company’s operations.

Through our growth years, we were organized where a customer got a dedicated team of operators. That’s who’s moving the cargo for them, coordinating the shipments, the same people every time. So you get to know these businesses, get to know their SOPs. Of course, we digitize and document the processes for each customer and they’re unique. But when it’s the same people every time running it, you just get much higher-quality results, relationships, the culture’s stronger. We call it a squad. They become much more of a gelled team with a little mini culture within its own.

They take pride in their book of customers. We work with some of the coolest brands in the world. They start buying their products. They enjoy working for those customers. We had moved away from that in pursuit of efficiency. Ironically, you become less efficient in logistics when your quality suffers because bad quality mistakes cost you all the efficiency gains you could gain in the week or in the month. You forget to pick a container up at the port, you’re going to spend hours and hours negotiating and paying fees and stuff, and if you just were Johnny on the spot you won’t have that problem.

So I’m not a genius by any means, but I had been talking to enough people pushing to go, “Hey, we got to go back to the old model.” So on day one, I pushed us back to that. That created incredibly positive momentum in the culture. So when the media’s reporting negative things about us, and internally we know that’s not the case, it kind of creates this rallying cry internally, “Us against the world,” that I think is really positive for us. And then what I did in my first two and a half months on the job was I did five customer calls a day, five to eight. I think on average it was around five.

So all day, every day I’m just talking to customers. And I managed to talk to something like 60 percent of our revenue in that time, get feedback, document it, start solving the problems that I hear about, build that momentum. I can’t talk to all the customers, we have like 5,000 customers, but talked to the biggest ones and got a lot of momentum around the culture piece that I’m talking about. And so people start to see that and go, “Okay, this is awesome. This is what we signed up for.” They want us to get back to our startup scrappy roots.

They don’t want to see us waste money. They want to see us get profitable, our employees. So the vibe’s been super strong. Part of why I want to do podcasts like this is I get to tell a more detailed and nuanced story than what gets reported. I know you also do longform journalism, but it’s harder to package that in a story than it is, “Hey, we have an hour conversation,” and people can listen to a more nuanced telling of what’s really happening at the company.

But if we don’t do the journalism, then you won’t be able to create the “Us against the world” mentality. We’re here for you, man.

Oh, yeah, exactly. Exactly. Yes.

It’s a critical part of the ecosystem. 

The thing that I’m interested in is the mechanics of change. That’s what our show is about: how do you make decisions? That’s the key question on our show. And it seems like here you made this big decision, you executed it, you came back as the founder. But I’m curious, has this changed how you make decisions? Notably, the last time you were on the show, I asked you that question and you said, “I don’t like making decisions. I want to hand off as many decisions as I can.” And now it’s like I’m hearing you, you want your hand on the wheel.

Yeah, it’s interesting. Yeah, I have to be way more top-down than I have been in the past. Not that I don’t like making decisions, I like to empower other people to make decisions, but I think I did that a little too much. We want to empower lots of people to make the kind of reversible decisions so they can move really fast. One of the things I got for six months… and in about half of that time I was a full-time investor, and I’m still a venture partner at Founders Fund.

It was very clarifying to have that time to reflect and think about businesses from the perspective of an investor, and then come back, even through that entire time, start thinking about Flexport from the mindset of an investor. And the main thing investors are looking for is return on invested capital, of course. Within a company it’s kind of like, “Okay, your marginal dollar spend, how much profit is that going to generate in the years to come?” That’s your return on invested capital at the margin, which is the fundamental thing that the most valuable companies in the world, their next marginal dollar reinvested in the business has a very high rate of compounding, and now you’ve got a compounding machine.

I don’t think that that’s necessarily the lens that founders have all day. They think a lot more about competitive advantage with customers, the culture of their team. Those things should lead to a higher return on invested capital, but it’s not the first and foremost lens. And I don’t think it needs to be, but you want to be able to tie the stuff that you do to return on invested capital. So coming back into this seat with a lot more focus on, “These are the three things that matter.” And if you’re in a job for 10 years, if you’re not taking that time to reflect and go, “What are the two, three, maximum four things that really, really matter,” and spend all your time on that. It’s very clarifying.

And so for us, it’s really that culture piece, and especially the culture of talking to customers every day and solving problems for customers. Coming out of that, it’s quality, on-time performance, data quality, that’s what people pay us for. The response times when people are requesting things from us, how fast are we able to solve their problems? And then cost discipline, we’ve got to be cheap for our customers, therefore we’ve got to not waste money internally. And that’s kind of it. Those are the three things that really matter for this company. And so I’ve spent all my time obsessing on those three things.

So how do you make decisions? Is that the framework? How does it line up with those three things?

We have a bunch of initiatives that line to that, and I spend all my time on the initiatives that are lined up to one of those three things. And a lot of it right now is on the cost discipline side, making sure we’re running really good unit economics and not wasting money on fixed cost stuff. It doesn’t add value. And how to make decisions is pretty clear. I mean, you get smart people in the room and listen to them and then make a call. I don’t think that’s rocket science. I think Amazon has some pretty good principles for this around deciding if this is a reversible decision or not.

What’s really interesting here is even the customer focus, that’s Amazon. A lot of what you’re describing are pretty big Amazon values: focus on the customer, bring the cost down. Then you had an Amazon guy, and it seems like you got away from those values. That seems strange.

Amazon still, to my view, is obviously the best obviously top company in the world. And I do think they’re really customer accessible. It’s just a really different type of customer. Consumers versus businesses are super different. And I mean, get some Amazon merchants on your podcast, I don’t think they’re going to tell you that Amazon is super customer obsessed when it comes to people who are selling goods on Amazon.

Fair enough.

Their customer is different. And so that’s actually a really fundamental difference. Our customer’s the merchant. One of the big differences between us and Amazon is that Flexport is a neutral provider of logistics and Amazon is a provider of logistics, but their customer is the end consumer, and you’re often going to be at conflict. It’s almost inherently, the merchant and that consumer want different things.

You’ve mentioned cuts, you’re smaller. You mentioned how you changed some of how you are talking to customers and how you’re doing the work. How is Flexport structured now, have you changed it?

We changed the operations piece. We’ve made some tweaks. We put the operations into those customer-organized squads, so we merged a few groups. That work had been broken off across ops, the account management team, and we just kind of merged those two teams, and then tightly paired a sales person to that group. We call that a squad. It’s like this cross-functional team. That’s the biggest change we made.

We moved some leadership pieces around, moved some people around to different roles, elevated a bunch of internal people but haven’t made dramatic org structure changes. I think actually I kept a lot of what Dave did where he had way more direct reports than I ever would have in the past, and I’ve now done that, too. I have a lot of direct reports. I think 12 right now, I kind of like it. I don’t like making decisions with all 12 people in the room.

I’m pretty clear about that with the group. I’m like, “We’ll get a smaller group together based on what the decision is that needs to get made.” And that smaller group usually involves someone or several people from down in the org who are more experts on that particular topic, more in the weeds on it, and make decisions in smaller groups. But for CEO-level visibility into everything that’s happening, I kind of like having more directs, and it works really well.

Let’s talk about shipping, which is what you do. The last time you were on, it was the middle of the supply chain crisis, maybe the height of it. You had just come to literally worldwide fame by asking the port of Long Beach to stack containers higher. You were riding high, but the supply chain was not, and rates were really high. It seems like the crisis has eased and rates have come way back down. Is that what you’re seeing, that the crisis part of it is over?

The main cause of the crisis, let’s say, was actually just a huge surge in demand for goods. I mean, you had a 20 percent increase in container volumes flowing through the ports of the United States and ports couldn’t keep up, but we ship more containers than ever before. So in many cases it’s like a demand crunch rather than a supply problem. The supply side of the industry shipped more containers than ever in human history, so it’s hard to say, “Yo, you guys are failing.” But there was more demand than ever.

That subsided, so demand is way back down to pre-pandemic levels. So just the delays went away. Interestingly, though, the carriers, the ocean carriers, the people that owned the ships made a lot of money, hundreds of billions of dollars across 10 companies. So these were incredibly profitable years for them. They ordered, and I think they’ll tell you themselves and have said it publicly, they ordered too many ships. They went out, took all that cash reinvested into more ships. I think they felt like the boom times were...

This is a classic human psychology. In any asset-based businesses, it’s boom and bust because in the good times you go buy more assets, then all of a sudden you have oversupply, the price crashes and the assets, now you have too many assets. Yeah, the boom and bust cycle. They ordered about 25 percent more container shipping capacity coming online in the next two years. So you’re going to see lower prices, it’s going to put them under a lot of pressure to make money. It’s good for consumers, the price is going to come way back down.

The really interesting thing that’s happening in container shipping, well there’s a couple. One is the Panama Canal has a drought, and there’s a huge blockage down there that depends on rain water. It’s rain fed, the canal. Unlike the Suez Canal. The Suez Canal is just ocean, the Mediterranean connects to the Red Sea at sea level, there’s no locks. Panama Canal, there’s locks. You go uphill and then back downhill. So it’s fed by rain water. So the drought has meant the levels are too low, can’t pass ships through. That’s very interesting because it’s leading to delays. And then Suez, you had just two days ago, the rebels in Yemen shot a missile and hit a ship.

There’ve been a number of these attacks on ships. That’s something to really watch for. If we can’t ship through the Suez it’s going to be a real disaster for global logistics. And then the other super interesting phenomenon, I just got back from two weeks in our Asia offices. I visited six of our Asia offices in 10 days, and I met with a bunch of major cargo airlines that were based over there. And they told me between 30 and 50 percent of all of the volume that’s moving in the air cargo is parcels for e-commerce. And that’s up majorly.

And some of these companies are projecting 100 percent year-over-year growth, which is just crazy to think about. So if the Suez suddenly becomes unreliable for shipping from Europe to Asia, or Asia to Europe, it only takes a small amount of volume diverting to air freight, because a 747, as big as it is, will only hold seven containers of cargo. And these big ships hold 10,000 plus. So if just a handful of them switch over, all the cargo airlines, the price of air freight is going to go nuts. It’s going to be very unreliable for fast-moving consumer goods to fly on air. So it’s never a dull day in logistics. It’s like the most fun industry in the world. It’s like every day something new.

You mentioned obviously the Suez, there’s militants there, there are two wars going on right now. Has that affected your business?

The Russia-Ukraine war’s had a big impact. First off, some of the biggest cargo airlines in the world were based in Russia and Ukraine, and that capacity has been effectively pulled offline. I think some of those planes, well, some of them got destroyed. You’ve probably seen the Antonov 225, the world’s biggest cargo plane, blown up. A lot of the Russian ones are just grounded, and these planes if you don’t fly them and maintain them, they just can never be brought back online is what I’m told. So you lost about 5 percent, I think is the number, of the world’s cargo planes taken offline permanently.

Then Russia overflight, you have to go around Russia right now. Many airlines won’t fly over Russia. I noticed this on my own flight home from Singapore to the US. Normally you would fly over Kamchatka Peninsula, end up over Alaska. I was looking at the mountains up there. And now you don’t do that, you go over the Pacific the whole way. You avoid the Russia overflight. It adds an hour or two, it adds fuel costs. Going to Europe, the same thing, a lot more is flying south of Russia. It’s costing you a couple of hours. Sometimes you have to refuel in Dubai or something like this.

That costs money, it slows down transit times. I mean I hate to talk about money and transit times in the context of a war, it makes me sound like I’m insensitive, obviously. But if you ask about the impact on logistics, that’s pretty real. The Israel one is, I think, more of a risk at this point. Something that everybody needs to keep an eye on, everybody in logistics. If the Suez suddenly becomes under threat and the insurance providers won’t provide insurance, then you have to go around the Cape of Good Hope, or you have to go around Africa.

And those seas are very rough down there I’m told, it’s very difficult for some of these container ships to go there. They’ll have to have less containers. If you have a super overloaded container ship in heavy seas, containers will fall off. So that can be very, very disruptive. And it’s a much longer journey so that if a longer journey reduces the supply, increases price. So yeah, again, it’s not the important thing about war, obviously, but when it comes to logistics, it has an impact.

You’re describing this as every day is a new challenge. You’ve got to grow your business inside all of this. What do you see 12 months out, 24 months out? Are you thinking on those kinds of timelines or are you at one, three, five?

You always have to be a long-term thinker and short-term, that’s the nature of running a high-growth company. We are laser focused on quality right now for our customers. I don’t worry about growth very much at Flexport because we know if we just do a good job, that growth comes, and we know how to grow, too. We have a great differentiated product. When it comes to technology, user experience, there’s nothing like Flexport in the global logistics market, which is a multitrillion-dollar industry. When you count all the ancillary services we offer — trucking and Flexport Capital is our inventory lending business — we just have trillions of dollars of market available.

So it’s just do a great job, get in front of customers, show them your product, show them what you do. You always win business. Now it’s all about quality. And I’m not even worried about on a one- to two-year timeframe. For us to get profitable again, we need to grow about 35 percent. I actually think we can do that in one year. I have set that goal internally. I think it’s fun to have ambitious goals and I think it’s doable. But if you do it over two, three years, that’s a layup in my opinion. So we’re just mostly focused on quality and let the scoreboard take care of itself.

You mentioned the supply chain crisis such as it was, it was really a demand crunch because people wanted more stuff. It is the holiday season. Is all that extra capacity that came online, is that going to be useful this holiday season?

It’s been relatively smooth sailing as far as getting deliveries in for holiday season. Peak season is over as far as ocean freight. That freight usually arrives in October. So air freight, even now, it should already be flown in. It is the middle of peak for our fulfillment business, Flexport Omnichannel, where we’re delivering to consumers’ doors.

Things are flowing pretty smoothly, at least when it comes to our operation, 99 percent on time performance for shipping out on time. So we feel pretty good about it. I haven’t gotten the latest media hits from UPS or FedEx. They often put out these good reports. I haven’t paid that much attention. I should go see what they’re saying.

The last time you were on it was because you pointed out, “Here’s one big fix to help this port.” Is there something else in your mind where, “If I could wave a magic wand and fix one thing in the supply chain, I’d fix it”?

Yes. The thing that’s really interesting right now that we are fixing is what’s happening with AI. We’ve had incredible results working with OpenAI, as well as with a smaller company called Adept. It’s my friend’s company. He used to be the head of engineering for OpenAI, actually. He left to start his own company. What we’re finding is that because we’ve spent a decade breaking the work of freight forwarding — of moving a container around the world — into small, discrete, atomic tasks that have to be completed, those tasks become very susceptible to AI knocking it out.

Whereas if you just told AI, “Hey, ship this container from Ho Chi Minh [City] to St. Louis,” it would just hallucinate some weird answer. But if you’re like, “Hey, move this data off this port terminal website into this database,” it does it instantly and accurately. And so if I could fix one thing, it would be like, I don’t know, maybe it’s a fast-forward button on our AI roadmap because every couple of weeks it’s like, “We automated this thing that used to take a thousand hours a month of Flexport labor, now it’s been chipped away.”

So it’s not like overnight AI just solves the problem. But methodically going through one task at a time within our operations… the cost of logistics, the cost of shipping a container door to door, about 10 percent of that cost that the customers pay is going toward labor in the coordination layer. Labor in companies like Flexport. So if you made the cost of shipping everything 10 percent cheaper, I think that’d be a huge boost for the world economy. So yeah, I’m pretty excited about that, but there’s no gods of AI that we have to pray to to make this happen.

It’s a lot of hard work by just implementing tech, and a lot of it’s off the shelf. Some of it’s open source, some of it’s from these companies I mentioned, some of it’s in-house developing models on our own. But yeah, it’s very, very exciting times. Frankly, I’ve been talking about automation and logistics: “This is what we’re going to do, automate this 10 percent of labor down, down, down.” It’s been very challenging to do, much more so than I would’ve told you five years ago or 10 years ago. Obviously, we haven’t made the progress that I wanted to. Suddenly in the last three, six months, OpenAI and other companies have made this, suddenly it’s possible.

That looks like really enhanced robotic process automation? You’re going to log into some bad system, pull a PDF down, put it into a good system and move on, or that looks like AI is actually doing it?

It’s like RPA that doesn’t break. That’s probably the best way to describe it. And we build a lot of scrapers. Flexport gets data, like “Where is this container?” On a given container, we’ll probably have 16 to 18 different data sources on where a given container is. So a lot of what we have to do is reconcile, “Who do we believe?” We call it the belief network.

“Which data are we going to present to the operator or to the customer, based on the reliability that we know of these data sources?” Some of those data sources, even with that many, it’ll fail and you get no data, and you have a human log into some website or make some phone call. You’ve got to, otherwise I don’t know where the container is, I’m not going to be able to deliver it. Or we have bots like scrapers, RPA that goes out and gets the data.

Those bots and scrapers are often failing because the carrier slightly changes the website, now all of a sudden, the thing is kind of brittle. The AI just does it. Or take an ocean carrier contract, it comes in these Excel files. They have all the different port pairs. It’s thousands of rows, many tabs. If-then statements, subject to, like, “Oh, if it’s out of gauge, it goes up by this price, or if it weighs too much, this price.” That was taking several days to get that into our databases. We try to have a 24-hour turnaround time, but you don’t always hit it.

And these contracts are only valid for a few weeks in many cases. So if I lose three days on a two-week contract, it’s less about saving my labor than it is getting those three days back, so I can give the right rates to the customer quickly. The AI can do it in less than five minutes and be more accurate than the humans. So things like that are just improving the customer experience and lowering costs.

How fast until you roll AI out into the actual logistics of shipping?

We’re doing it every day. Most of it’s behind the scenes today. But another great example is, we’ve got a container booked from a customer, “Which ship should it go on? I’m trying to hit this transit time, the customer needs it to arrive by this date.” Let’s just say from Yantian, which is the port in Shenzhen, China. I think it’s the second-biggest port in China after Shanghai. I’m going from there to Southern California. I’ve got two options in Southern California: LA and Long Beach. And you go, “Alright, it sounds kind of simple.”

But actually I’ll have contracts with six different ocean carriers, and that’ll give me 12 different strings. In a given week there’s 12 different ships that go on that combo. And then I’ve got multiple weeks, obviously. So which ship is this container going to end up on? That, until this summer, was a problem that humans at Flexport would make the decision. Now it’s a problem that AI decides, and the reason you really need AI for this is because it’s not about saving.

There’s not that many people required to make these decisions, right? It’s pretty cheap. It’s not about saving labor. It’s because every week, 1,600 containers have their dates changed by the customer. The manufacturer is delayed, and the cargo is not ready when we thought it would be. So in the past, I would end up having to cancel that booking for the ocean carrier, which makes me a bad customer, if I’m canceling my booking. And that’s how it is. In ocean freight, 30 percent of all container bookings get canceled for that reason.

So what we do now with AI is we replan every container in the network 10 times a day, and we’ll pull another container that’s ready forward a week. We’ll reshuffle everything. So now when someone cancels, someone else is getting a better transit time, and I’m not canceling on the carrier. So the net result of this was a 20 percent increase in our on-time performance. It’s dramatic. A reduction of our carrier cancellation so that we’re best in the industry for cancellations, and save ourselves 2 percent of our costs because finding the cheapest routing that’ll hit the transit time.

I mean, it has to be the best use of machine learning in shipping. Maybe the best use of technology in shipping. And we just launched that in July of this year. So a huge potential, and you can do this at every single job. A lot of it is like, “Hey, they’re doing a better job than a human.” It’s not like the humans are that expensive, but it does improve your margin and let you lower your cost, too.

I feel like I could talk to you about that alone for another hour, so you’re going to have to come back.

Alright, I’m happy to.

Thank you for being so candid about all the changes, I really appreciate it. Ryan Petersen, thank you so much for being on Decoder.

My pleasure. Thanks for having me.