The Exchanges

Q So when you've had this series of successes in building businesses, what led you to start investing?

A So the funny thing is I started investing essentially at the same time I started building companies. So my very first investment came in 98 at the same time I built the first company because by virtue of being a consumer facing internet CEO that was very visible to the public, a lot of other entrepreneurs approached me for capital. Now in the early days I had very little capital, so I was putting more sweat equity and advisory shares and a little bit of capital. And I thought actually long and hard, should I be investing? Is investing alongside running global multinational corporations a distraction from my core underlying mission? And realize that if I could articulate lessons learned to others, it actually meant I internalized them and made me a better founder if I had like these heuristics and theses that I could articulate. And number two, by meeting all these founders, it kept my fingers on the pulse of the market. I really understood the best practices And we're the bleeding edge in tech in every category. And that also may be a better founder. And so from the get go, started investing. Now, because I was mostly building marketplaces, except for the zingy interlude, I decided I couldn't allocate too much time to being an investor. And so I decided I was only going to focus on marketplaces because I had pattern recognition. I could decide very quickly. And I created a set …

AI assessment note: “I started investing essentially at the same time I started building companies.”

Q That's 2013. What's gone on between then and now?

A So now we're a team of 33. We are four partners, two associates and one analyst. And then in addition to that, we have two entrepreneurs in residence that are looking to build companies. But in the meantime, they take investment calls as well. Two apprentices or future entrepreneurs in residence. So we have 11 people filtering deals. And then we have an entire back office team of 22 people, including the platform team, the finance team, et cetera. Every week right now we get 150 deals inbound. And in fact, we're extraordinarily privileged. I mean, because of our positioning and our brands, the deals come to us, which is an amazing quality as well for the people on the team, because if you look at the associates or analysts and most other venture firms, they spend all their time deal sourcing and cold calling, analyzing LinkedIn. We have the opposite problem. We're drowning under the fire hose of deals and we have to filter through the deals. So right now we get a 150 deals in about a week. They get randomly assigned to one of the team members. The team member reviews whether or not we should take a call. We ended up taking a call with about 50 of these. It's a one hour call, which actually follows the same deal memo and the same four heuristics we defined before. Every Tuesday from 10 to 12, we have an investment committee meeting where we review the deal memos from the week be…

AI assessment note: “So now we're a team of 33. We are four partners, two associates”

Q I'm curious to turn a little bit to The dynamics of marketplace businesses. What have you learned about how these businesses work and succeed?

A So of course, when you have a marketplace business, you have a chicken and egg problem where you need to figure out if you sort of the supply or the demands. And of course the answer is it depends, but most often you should start with the supply because the supply is financially motivated to be in the marketplace. And so when I launched OLX, I went to the car dealers and the real estate brokers And when I launched Auckland, I went to the coin collectors and the SEM collectors and collectibles collectors to have them put their supply online. And I told them, look, we're small, we're early. We may not sell, but it doesn't cost you anything. Why not? It's an incremental distribution channel for you for free. And as a result, it's actually reasonably easy to get supply. Now in modern marketplaces, especially local services, one, it's actually often better to go hyperlocal. You curate your supply to the very best level. And then you focus on bringing them demands. And what I've learned is on a product marketplace, once you have a 15 to 20% sell-through rate, ideally 25%, where you can start at 15, 20%, or in a labor marketplace, where you can start representing 15, 20% of the revenues of the supply, you've basically reached liquidity. And so once you reach that level, you re-increase the supply again, and then rematch demand. And you keep doing that forever. There are a lot of pitfa…

AI assessment note: “once you have a 15 to 20% sell-through rate... you've basically reached liquidity”

Q And so when you got out of school then, how did you decide how to participate in this world?

A I went to Princeton and I studied economics rather than computer science because it explained to me the way the world worked. But I went to Princeton having no money coming from France. And so I actually built my first reasonably small company there, exporting high-end computer from the U.S. to Europe to pay for it. When I graduated, I knew I wanted to be in tech, and there were three choices. And to answer the question, choice number one is I could build a startup. Choice number two is I could join a startup. In both of these, I felt that if I built a startup at 21, not connected, I was shy and introverted, I'd never really managed a team, even though I'd built companies before, they were all like sole proprietorships, I probably would fail. And if I joined one, being 21, I would Not necessarily be taken seriously. So I'm like, you know what? I'm going to go to McKinsey. It's a business school, except they pay you. And it was actually useful. I learned oral written communication skills, learned to work in teams, and learned what I felt I had to learn. And then at the age of 23, felt it was time I left McKinsey. Then, uh, time had come to go and build my first startup.

AI assessment note: “When I graduated, I knew I wanted to be in tech, and there were three choices.”

Q So what is it about marketplaces that's attracted your attention as an investment strategy?

A What I love about marketplaces is you actually create an extraordinary amount of value for humanity. You take these markets that are opaque, fragmented, where transactions are happening through in-person, through email, through phone, and by virtue of bringing them online, you're unleashing efficiency and growing the market dramatically. And there's nothing as undervalued as a marketplace that has found part of market fit. What we need to do is identify when that happens. And sometimes it's early, and sometimes it's late. The company we invested rather late in that has been an extraordinary success for us has been Coupang. Coupang is now the Amazon of Korea. They're worth, I don't know, fifty billion dollars or so. We invested a four billion valuation. We'd seen them early, but first they were a Groupon, and the union economics were underwater. Then they pivoted, and they were one of many Amazon players. But finally, there was a moment at a four billion valuation where they hit it, and that's when we invested. So even in the public markets, when these companies are there, people underestimate how quickly they can grow. But the network effects are so powerful that it leads to exponential growth, and humans don't really think very effectively in terms of exponential terms, and they don't realize that when you have 50% compounding year over year, it ends up to lead to very big num…

AI assessment note: “What I love about marketplaces is you actually create an extraordinary amount of value”

Q You mentioned qualifying this deal flow that comes in. So why don't you walk me through what you're looking for stage, geography, marketplace business?

A We look at every stage in every geography in every industry, but with a specific focus on marketplaces, of course, only online businesses. And so the reason we go from one 50 to 50 in terms of like taking calls, so one 50 deals come inbound and we only take 50 of these As many of these are just out of scope for us. They're in biotech, they're in hardware, they're in space tech, and we'll do things in healthcare or in space if it's a marketplace where the dynamics are the same and we can recognize the pattern recognition is the same. I mean, it's pretty insane how similar the marketplaces play out, whether you're in petrochemicals or oil sources workers or an Uber type marketplace. And so the business model Pattern recognition is extraordinary, and then it applies in every geography, in every industry, in every stage. Now, if you look at our portfolio to date on the 773 companies, we're 55% U.S. and Canada, 25% Europe, 10% Brazil and India, 10% the rest of the world. And by rest of the world, I mean like Vietnam, India, Nigeria, Kenya. In capital allocation, we're 50% pre-seed to seed, 50% beyond words. Now, we are obviously from a number of deals, there are many more C than A, many more A than B, many more B than C, given the way the venture funnel flows. In terms of business model, we're 70% marketplaces, 30% other. Now, the other falls in one of three buckets. Bucket number o…

AI assessment note: “we're 55% U.S. and Canada, 25% Europe... In capital allocation, we're 50% pre-seed to seed”

Q And so how did it evolve from just you to FJ Labs today?

A In 2009, I met a team that was building eBay of Latin America, and so I gave them the business plan technology for my company. In fact, we launched them from my servers in Paris in record time, and they ultimately merged with MercadoLibre, and they became the leading e-commerce company in Latin America. They were a bunch of co-founders. One of the co-founders became my co-founder and co-CEO at OLX, And then one of the other co-founders was building his own startups in Latin America, was investing, and we were reintroduced in oh eight or oh nine. He had a lot of deal flow in real estate and in travel, and he was doing more due diligence and looking at legal documents that I ever had patience for. And so my business partner at OLX said, hey, you should work with them. He can help you on that side. So we started pooling our investments. And so by 2013, when I'd left OLX, we'd already made over a hundred investments together. We'd already had like dozens of exits. It was doing extremely well. And we created FJ labs, frankly, not with the idea of creating venture funds. And in fact, at that point in time, I evaluated what do I want to do next? And I had a whole slew of ideas. And the one that stuck and that was working was, um, investing my own capital and either building companies and I'm still building one or two companies a year or investing in startups. And so that started, took…

AI assessment note: “So we started pooling our investments... And we created FJ labs”

Q So once you've done all these deals and you've got this big portfolio of companies, how do you spend your time trying to help them out?

A So we have one superpower, which is to help them fundraise. What I didn't describe is the 150 deals we get every week, we get them from three sources. One third of those deals comes from other VCs because we don't really compete with other VCs. We don't lead, we don't price, we don't take board seats, we write small checks, we're not competing for allocation. In fact, we're not even ownership sensitive per se. If we want to write a hundred K check, there's only a hundred K. We'll take the hundred K. We'd rather be in than not in. As a result, every 12 weeks, we sit down with the top hundred VCs in the world. We'll sit down with everyone at every stage, and I mean, Andreessen, Greylock, Vessemer, General Catalyst, Tiger, First Round, I mean, we'll really cover everything. It's a really symbiotic relationship. We bring them all of our best deals, such that they can invest in them, And of course we have such a large portfolio and then we have so many deals from the earlier stages coming to be funding that we have extraordinary number of deals to send them. Number two, they send us their best deals in marketplaces because they want a perspective and they only send us one or two deals a year and we'll send them maybe 20 or 30, but it actually works. And number three, of course, the founders love it because they get funded by the best VCs. You can imagine that if General Catalyst did…

AI assessment note: “we have one superpower, which is to help them fundraise.”

Q Yeah. And the other is timing, which is you saw that ringtones weren't in the US, but then there's a dearth of capital and timing. How do you think about bringing together ideas and marketplace readiness?

A It'd be no sense in a way that wouldn't be big in the US if it was big in every other country. In fact, one of the key lessons I've learned over the years is that humans are frankly identical around the world. Individual humans want to have a sense of meaning in their lives. They want to be entertained. They want to communicate. And ideas that work in one place have a tendency to work in another because the underlying needs that we have are the same. And so When you see the slag, there's usually a market structural reason for it. And I realized that that market structure was going to go away. We had different networks that were not interconnected for messaging and there were no payment systems, but given how popular and successful and what the margins were of these types of products in the rest of the market and how they were benefiting the operators in general, I knew it was going to come to the US sooner or later. And the lack of capital in a way was a blessing because there were a few companies that could have done this. Few that were well capitalized, but they just didn't care, and this was not their mission, so that was perfect for me.

AI assessment note: “When you see the slag, there's usually a market structural reason for it.”

Q How do you think about funding one of the businesses in your portfolio across stages?

A So we are a cross-stage fund kind of for that reason. We want to be able to back them across time. Now, of course, we never lead them, so we need to have these relationships with the other VCs to help them. The way we try to have our founders think about it is if the capital is there and we don't think it'll necessarily be there in the future forever, given the frothiness of the world we're in today, You might as well take it. The way we think about dilution is if you're going to create more value than the dilution that you're incurring by raising money, you should do the deal. It's more about how much value you're creating the capital you're raising that matters.

AI assessment note: “We want to be able to back them across time. Now, of course, we never lead”

Q How do you think about the risks of effectively misreading people in such a short window of time?

A So what's interesting is there is a 15 year period of my life where I signed every single legal doc sent to me without even reading it. And as far as I can tell, I was never screwed over. The meeting, when the exit happened, I got my fair share and I invested the valuation that they said I would invest it. So most people I find are actually well-intentioned. And the amount of work that it takes to eliminate the few bad apples, I actually think an average is not worth it. And in fact, it's not worth it because it actually decreases the user experience for all the good actors to be trying to identify the bad actors. So how do I identify people? Well, the first thing is I look for passion, genius, things that come across extremely quickly. One thing humans are good at, regardless of category, we're great at judging other humans. If you meet someone, you know within a minute or not, if you like the person or not, right? If you're going on a date, I think it's like within one minute, you know. And the rest is pretense. And it's the same here at this point. Like, I know very quickly if I like the person, if I think they're good, if they're eloquent. Now, because I push back on their assumptions, I will question how they came up to their numbers. And so I don't just take their story as is. I push back along the way, and I see how they deal with the pressure of my pushback. Of course, …

AI assessment note: “the amount of work that it takes to eliminate the few bad apples, I actually think an average is not worth it”

Q What have you found is the best path to learning something new?

A Well, the twenty-twenty-one, it's a lot easier than ever before, because you can go to YouTube, you can go to Reddit, and you can go to Substack, and there's, like, infinite amount of variation in anything. The best way to learn is often building something. In my life, often, I didn't know what I didn't know, and so I didn't know, A, what was impossible when I had to did it, but also didn't know, like, when I launched the classifieds company, I didn't even know that ginormous publicly traded European company existed. I didn't know the importance of SEO, and yet, Within two years, SEO was 70% of the traffic, and then eventually we had to pivot. Sometimes being naive and not having the exact answers is actually helpful coming in a category, but reading a lot and frankly, executing. Right now, as I'm getting scored in crypto, I'm building a crypto project for fun. I'm learning to code in Solidity and Rust, and it's hard, but it's interesting.

AI assessment note: “The best way to learn is often building something.”