E182: Lessons from 17 Years at Menlo Ventures and Accel w/Tyler Sosin

0:00So everybody comments on your name, villain. What is the origins for your name and how did that come about? Well, the origin for villain comes from a quote from Batman. Harvey Dent said to Batman over a dinner, quote, you either die a hero or see yourself live long enough to become a villain. And just as like a cinephile, I always love that quote. But when I reflected upon my investing career, you know, it truly resonated with me as sort of a North Star of what success can be. And that is, you know, finding companies, you know, that basically survive and grow and continue to survive and endure through many decades often and become, you know, essentially the dominant players in their category.

0:50And obviously these companies, when they're young, they start off as heroes. But if they do manage to get to a 20, 30-year mark, they're often villains at that point. They express really interesting tendencies that rational capitalists should love. They have got amazing customer lock-in, really dominant relative market share, pricing power, and they're really hard to dislodge. And so that's why I like them as my North Star. What are some examples of some now villains that you invested at Menlo? Well, I'd say more generally, like, hey, the universe of tech companies out there, like who would be, you know, in my opinion, a villain.

1:29And this is coming at it from a compliment, like Oracle would be a villain. They've been around for a long time. In the vertical software categories, you know, that I love so much, you know, Jack Henry and FIS and core banking, Vertifor and applied and insurance. you know, Epic and Cerner in medical health records. You know, these are companies that have been around for multiple decades. They're very large. They are extremely dominant in their categories. And often if you were to ask, you know, people who use the products or people are trying to disrupt them, they are the targets because they're no longer innovating.

2:08They capture a lot of economic rent, but it's in some ways almost impossible to displace them. It's almost like this full circle of the innovators dilemma. It's a startup incumbent now disrupted by a new startup, but on its way to being a hundred or$500 billion company. Yeah. And I should say, I don't think villains necessarily have to be enormous to be that way. I think they just need to be very dominant in the categories they compete in. As Peter Thiel would say, he hates competition. He likes monopolies. Competition is for losers. Competition is for losers. And I feel the same way. I'd rather be in a business where, for a variety of reasons, at some scale, it's very hard to compete with them.

2:54And so then they can start to display some villainous tendencies and basically have amazing shareholder returns as a result. So villainous tendencies are a side effect of having so much market power. It's also the capacity to have a relationship with a customer where you can extract economic rents. And, you know, that's very hard for most businesses to do. Most businesses compete in commoditized markets. They have to basically run on that treadmill or sprint on that treadmill forever. But there are a few companies out there that don't have to do that. And those are the best investments. I know you didn't define a market cap, but let's define that now as a$100 billion company.

3:37How many of those companies have this kind of extractive relationship with their customers? And how many are their customers just love to give money like a Starbucks? There's different sorts of customer lock-in that you can appreciate. You know, like I sort of see myself as an anthropologist. And so the question is, what are the circumstances and the behavioral tendencies that create lock-in with a customer? And on the human side, habit formation is a huge one. It's why the nicotine companies have been so successful over time. It's why the Starbucks of the world have been so successful. In fact, if you look at, I believe it's either the profit margins of addictive businesses, there's a strong correlation between how addictive they are and how profitable they are over time.

4:25And then there's some secondary effects, reinforcement effects that really matter. But, you know, for, I'd say, consumer products, addiction is a major point. And then the other major point is things around, you know, habituation and just becoming like really familiar with something and therefore not wanting to give it up because it's what you know. And there's a friction to changing to something new. You have a bank account. That bank account is connected to all these different bill play partners. Your entire financial life is set up on it. It's a hassle to move. And it's that cognitive load and perceived friction that keeps people from switching.

5:06regardless of whether it's easy to do or actually really difficult to do. So you invest at the earlier stage into these companies that you want to be a villain, to be mature and be a villain at some point. What characteristics are you looking for? When I come back to what I'm doing at Villain, I'm very focused on vertical software and technologies. I'd say vertical software, vertical AI, vertical payments. These are the sort of substrates that I will be dealing with at this firm. The very best companies have very, very sticky relationships with their customers over time. In a more dry sense, it's basically they develop annuities with these customers.

5:48And the best businesses build a widget or whatever it is, and that widget creates an annuity by selling it to that customer who basically consumes it for many, many years. And so I want to see the earliest evidence of that annuity and the earliest evidence of a founder who knows how to sell that annuity with some amount of efficiency. And if I find that in a vertical software market, I can get really excited. That's evidenced by a low burn rate. So soft customers continue to purchase the product month after month. Is there any other metrics and leading indicators that you're looking that a customer would has this annuity type of relationship with the company?

6:31You'd have to see what their behavior and usage are. You'd have to test pricing and their willingness to churn if you were to suddenly raise pricing. I'd like to know that a founder with an initial wedge in the market that's efficiently developing, generating these annuities has a roadmap for basically building new products and features that they can sell to that existing customer to deepen the relationship significantly over time. And you would argue that you're creating more of a feedback loop from the customer when they're buying multiple things from the same company, then they're more likely to become a long-time customer.

7:05Yeah. And there should be two, maybe three reasons for that. One, in cloud software, and this is like an important tenant of this business, like you have this data plane that you're creating, maybe with that initial wedge product. And, you know, in the best of cases, that data can be used in a second or third product to give you the right, like the permission to build a product that would be better than what someone else building that product DeNovo would build, because you're already using, you already have the data that's relevant. So it's a matter of a different workflow system or a different set of computations, but you have an advantage by having the core data set that's used in a variety of different tasks.

7:43So that's number one. Number two is, you know, as you develop a deeper relationship with a customer and have multiple products to sell them, there's a very powerful phenomenon, cross-subsidization that can occur. You are pricing and packaging multiple products together. Often these products, you know, work better together in terms of a synergistic, you know, outcome. but also in terms of pricing, you might be able to say, hey, if you buy this product and this product, they both work together and we can give you a discount where if you bought these products separately, it would be more expensive.

8:19And over time, that's a very compelling way for more incumbent companies to compete against startups. You know, it's the bundling effect. Give me an easy to understand example of a company that started with one product and then successfully sold multiple products that most people would be aware of. Microsoft would be a villain and would be very well known for having multiple products in a bundle where over time they have a lot of success bringing out that new product, bundling it into their existing suite, and ultimately killing competitors or really driving competitors into a state of commoditization as a result.

8:56Going back to these companies that you look for that you think could be villains or market leaders. What's the first kernel in a company that you see that this might be one of these companies? When you think about competition and how to find markets where competition is less of a threat, there are probably a few ways of doing it. One is network effects. I mean, that's a wonderful business model if you can attain it. The second is something like a Carvana where it's like the amount of infrastructure and capex spend required to get to a scale where there's this virtuous cycle that is just so hard to disrupt is enormous, right?

9:36Amazon would have, you know, similar characteristics working for it. I'd say the third that's lesser discussed is market size. And I read a book probably about 10 years ago that just opened my eye to this. And it was a fascinating book called Competition Demystified. I encourage everyone to read it. But it was basically this sort of anthropological survey of monopolies and how they fail, right? So you think of monopolies as like, hey, these things really, like they shouldn't fail, but every once in a while they get disrupted. And the question is why? And the answer, which is compelling to me, is it's an exogenous effect or something happens that actually increases the size of that market.

10:16And as a result, it creates this sort of pocket of oxygen that allows a new entrant in to get to a certain amount of scale where then they're able to basically compete with the incumbent. And that it comes back to one of the big areas where competition can be controlled is, you know, one of the big factors is market size. And actually, it's smaller markets that I tend to like or sort of midsize markets because you have there's just less revenue to go around and is results in behavior where because the market size is perceived smaller or is smaller, there is less capital funding new companies in that market because the perceived outcome is smaller.

10:58Often these markets, when they're smaller, they're just like less understood. They're more niche. They're more technical. It requires someone who like has exposure to that market and a build mentality that's hard to find. And so there's fewer, I'd say, credible entrepreneurs who go into these markets. And as a result, you have fewer competitors. And when you have fewer competitors competing with each other, maybe a couple can grow up to be at scale, and that's fine. And then there's usually some sort of rational consolidation that occurs. And so in smaller and medium-sized markets, as a result, you end up with these market share constructs that are often quite lopsided, where it's like, hey, there's three players who control 85 % of the market, or two players that control the market.

11:39Look at a lot of the businesses I mentioned earlier, Epic and Cerner, Vertifor Applied. I mean, these are relatively large markets, but they're dominated by two people. And it's larger because they were small markets to begin with that grew over time. But these two players captured those markets. And then once you are a large player in a small market with a lot of the locking that I described earlier, even if someone wants to disrupt you, the fixed cost of doing so can feel prohibitively high. And so that keeps a lot of people out. So again, like my thesis for vertical software, it's not to say that vertical software can't build really big businesses, but I think a lot of great vertical software is great because it is actually going after a smaller market where you can have less competition from day one, less competition entering over time.

12:26And therefore, it gives you a lot more freedom, like a lot less pressure to have to grow at any certain rate or to have to have like a build velocity that's keeping up with seven other competitors. and it affords you, therefore, a chance to be much more capital efficient and take a much longer-term view. Put some numbers on it. What is considered a small market in your playbook? I would say anything below like$100 million of SAM or TAM, total addressable market, is considered a— So$100 million of customers spending money in that market. Yeah, we're like the perceived, hey, if we acquire this entire market, we could build a$100 million business.

13:08I think that is relatively small. And for me, I would be interested in companies pursuing markets of that size up to$500 million. Let's just go in the middle of that$250 million total market per year. A bunch of venture capitalists are looking around the table at different firms, and they all see the same $250 million market. And they say, this can't return the fund. Double click on the rationale and why it's not as competitive. Exactly that. I think you would find a business. I mean, my ideal villain investment, you find a company that is expressing that early efficiency, building an annuity with the wedge product that they're selling and therefore is growing efficiently.

13:51And when you look around, you see very little competition, right? Like maybe there's one legacy incumbent and then that's it. That's great. And then the venture guy looking at that says, well, geez, you know, it's a nice business. but my opportunity cost is enormous. And so if I put a$5 million investment in this company, I might make a 10X, but that doesn't move the needle for me. That doesn't make my fund. Yeah, I need to believe that I can make 150, 200X on this investment. And hopefully with the funds the way they're organized now, is I need to believe I can not only invest that$5 million, but to really make it worth my time, I need to invest another 50 million behind that or a hundred million behind.

14:37There's a minimum viable check size or check size per company. Yeah, there's a minimum check size. The way most funds work is like, hey, they've got a certain amount of slots, a minimum check size. A fixed number of board seats. There's a fixed number of board seats. And then they think about the amount that a investment can return as a percentage of a fund. and they're looking to basically with every investment they make believe that they have a fund returner on their hand um funds as they've gotten a lot larger um it's just much harder for them to get excited about a business where it's like wow that's that's great that's a 400 million dollar tam company's growing really nicely like whoop-dee-doo it doesn't move the needle because i'm looking for the 50 billion dollar company that i can write that first check and then i can write a much larger check from my growth fund, and then I can do something even larger from an SPV.

15:31And that's a very different mode of thinking and capital allocation. How do you make your mouth work? Three things matter. One is fund size. I think a fund size of$150 million or less is suitable for this strategy because you could have a bunch of$3 million,$5 million investments where you're making a 5 to 7x, And I think that those could have like a real, like they could be needle movers for that fund. Unsized matters. Concentration matters. Like I think we would be looking for a portfolio that's a bit unusual in the venture industry where it's 15 to 20 investments versus I'd say 25 to 30 or 25 to 40.

16:15And nominal pre-money matters. It's most liquidity in venture outside of the mega IPOs are acquisitions in the$100 million to$500 million range. And so building an investment philosophy and a strategy where we make a lot of money on those outcomes. And then if you were to reverse engineer into that, let's just say the average outcome you're thinking about is a 200 million EV once you sell it. Therefore, what does the pre-money have to be in order to make a 5 to 15x return? And my guess is that for us, it's going to be somewhere between 30 million pre, depending on some of the underlying characteristics of the company.

16:54And so then you might say, hey, well, Tyler, how do you find companies that have that early product market fit that you're looking for where you can sort of get them at pre-monies that sort of make sense, you know, that are anomaly on the lower end? And the answer is, well, you're looking for businesses where there's a fly in the ointment. There has to be something that is countervailing, that is uninteresting to the mainstream VC. And for me, the two things that I'm willing to take risks on are, one, the market size, and two, is the growth rate, right? And so market size, we discussed earlier.

17:29Growth rate, it's a longer conversation. But, you know, when you look at the venture industry today, I think there was a guy, like a benchmark partner interviewed on a podcast recently where he said, hey, like these new AI companies, like, are growing at, like, 4X year over year. Like the new, it's now. The new 2X is 4X. I venture guys always speak in like multiples of growth rates for the first three or four years. I never understood why exactly. But it's like now it's like you got to be 4x, 4x and then 3x or like you mean nothing to me. Right. And you look at these companies and it is incredible that can grow from like zero to 100 million in 18 months.

18:03Like it is phenomenal. Right. But like I'm delighted when they say this because it means like there's this like wide swath of companies out there who are only growing at like, you know, 120 percent. who are like completely uninteresting as a result. And so when you get those factors, like, hey, they might be growing a little bit slower. And 120 % is like exceptional growth, in my opinion. They might be growing at like 70 to 80%. That would be fine with me too. They might be going after smaller teams. That's fine too. As long as the pre-money valuation makes sense and the companies are efficient and the founders and I lock arm in the ethos of compounding the business as opposed to just growing the business at the highest rate possible at all costs, we can have a very, very productive and lucrative partnership over time.

18:55Is there a different psychology with these founders that are looking to build a$200,$300,$400 million company? Are they older, more experienced, less kind of in their 20s looking to become billionaires? And is there something different about these type of founders? I often think the founders who are willing to grow, go a little bit more slowly have been burned. Like I was actually speaking to a founder today who has this ethos of growing a bit more slowly, but also just like efficiently. And, you know, his last company ran out of money. It was just cash zero and died. The other phenotype of founder is one that's like been in the woods for a while.

19:35I've done a few angel investments as we've gotten villain off the ground. And some of these companies have been, you know, in the woods for five or six years. And they started with something and they had to pivot and they had to pivot again. So by the time they get to this product market fit that I love, like, I don't know, they've been kind of burned by the venture industry. It's just very hard for them to raise capital. The instincts of like survival and frugality are like in their DNA at that point. In what ways is being burned like that an asset? And what ways is it liability? I think it's an asset.

20:10To me, it's much scarier. Because you could argue the opposite, which is it keeps them from making bold bets and compounding and all these other benefits. But you would argue it's an asset. Generally, I would say it's an asset. In what ways? Because I think the vast majority of even great businesses that fail, fail because they grow too fast. There's this urgency to keep up with the Joneses, which again is that 4x growth rate. it's this feeling like hey if we don't achieve this perceived set of numbers like no one's going to fund us we're going to die our competitors are going to overtake us I mean there's like this panic that sort of you see as they start you know walking through their mental models and I think that can lead to some really really shitty capital allocation decisions right like most companies fail in my opinion in the venture landscape because they try to grow too quickly they try to build too much product too quickly.

21:03And they basically parallel process too many things that need to be done in a more sequential manner. And some of this also is just like certain markets require just more time to sort of like break. And there's a learning curve to any business being built that just requires time. And so I actually get quite like nervous about companies that are just growing fast, whether they're permitted to because they have great economics or not, there like there is something that like frightens me a little bit because um i think you know like organizational capacity can get strained to the point where things can break in a in a in a way that it's hard to put back to pieces and you've seen that in the venture landscape you know we've got all these assets out there that are basically hung right like from 2021 all these mega funding rounds, chasing a lot of growth.

21:58And you could say there's probably an alternative history where if these founders had a different mentality of just like growing, I'd say taking a compounding mentality where it's like an endurance compounding mentality where it's not triple, triple, double, double or 4X, 4X, 3X, but it's more, hey, I'm going to grow this business between 60 and 80 % for the next seven years or the next 12 years. and I'm just going to do it on my time frame, but we're going to continue making progress. We're going to do it with precision and with capital efficiency. I suspect a lot of the founders who now have hung businesses would be in much better places.

22:37It seems like there's like a dialectic here, like two opposing philosophies. One is about founder market fit. So you have like a Facebook that's literally plowing and burning billions of dollars trying to build their network effect before they even knew whether they had a business model. That was like the question for many years. And then there's a question in 2012, whether they could port their business model into mobile. They had to deal with all that. Sam Altman is another example, OpenAI just burning all this. People still don't know whether LLMs will have a sustainable model, but they're just going out there and doing market share.

23:10And then there's the businesses like Qualtrics that grew and compounded over 20 years, and oftentimes not in New York or San Francisco and these second-tier cities where they're compounding, and maybe it's one of these industries that starts out as a$500 million TAM, and it also compounds 12 % per year, and suddenly in year 20 it's a$5 billion industry, and they have this kind of monopoly position. So both models could work, but they're certainly different personality types. You make a great point with Qualtrics. You look at Procore. I think that was a 20-year overnight success. I think Service Titan was sort of under the radar building for a long period of time before it really broke out as an asset that VCs and growth equity firms liked.

23:59But I think there's something about that journey that makes these businesses amazing. And for me, hyper growth is much less exciting than seeing a business that was a kind of a slower grower that actually over time starts to see their growth accelerate. And often it's the case because the flywheel is working. The multi-product strategy is working. The familiarity of the market with the product is working. They're becoming the standard. The ecosystem is converging around them. And these create these flywheels of operational leverage where a business that was kind of sleepier growing at 70%, 80 % maybe suddenly is growing at 120 % at a much larger scale.

24:41And when you see those things, you've got to be like, wow, that's going to be an incredible business. Sometimes the different arms of the business aren't individually that spectacular. When you put them into one system, they achieve product market fit. It's interesting you bring up like Google or Facebook or OpenAI. There's a podcast I eventually want to maybe do myself called like The First Five Years, which would be going back and trying to get to like the first five years of financial data of like these, you know, really amazing companies. Because I actually think that, you know, Facebook was quite profitable early on.

25:19If I look back at like their S1, like business grew like an enormous amount through the early years. I don't know what the quality of revenue was. Obviously, like Facebook had these network effects that were just incredible. And so it's a business I really understand and appreciate. And same with Google. I think Google actually was really quite profitable out of the gate, growing at several hundred percent a year. And so there are special businesses like that. You know, OpenAI is a very different business. It hemorrhages cash to the extent that these other businesses at their scale did not. As a sort of someone interested in company history or economic history, it reminds me a lot more of the memory business where like the need to reinvest in the next generation model is feels like it's important for surviving and continuing to be the best LLM out there.

26:11And I personally just think that's a really hard place to be. I do think it's an interesting study today of these businesses that are being funded that have enormous burn rates in their early years and they're growing fast. And the question is like, when they get to be more mature, how valuable are they and how defensive are they? Startup history is also something that I'm really interested in, just how these things came about. And one of the most interesting things is this kind of three-person club that Reid Hoffman, Mark Pincus, and Peter Thiel had talking about social networks before Facebook.

26:43Talk about having a prepared mind. They would just talk about this. Obviously, Reid Hoffman also started LinkedIn. I think Mark Pincus started a social network that ended up not getting off the ground. But they were kind of developing this thesis both in real time individually and also as a group. So they had this really powerful, prepared mind for when Facebook landed on their lap. They were almost waiting for the Facebook to come about versus to the rest of the world. The rest of the 7 billion people, it just seemed like a totally novel, totally idiosyncratic business. They were really ready for it.

27:17For Facebook, it was probably some combination of cloud computing and modern software engineering and behavioral psychology, understanding addiction and how to get users engaged, that all of these things needed to come together to make a network effect that we understand it as now viable. And before, Villain, you were at Menlo Ventures, storied venture capital franchise, best known for Uber and I'm sure many other Deca unicorns. When you were inside Menlo, how much more powerful is it to be around a group of really smart people kind of workshopping these ideas versus I have this thesis? How much did having a group help you formulate your thinking?