Invest Like the Best with Patrick O'Shaughnessy - Bessemer Venture Partners - Building a VC Firm that Lasts Centuries - [Invest Like the Best, EP.327]

Starting point is 00:00:02 Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO and founding partner of Positive Sum and the CEO. of O'Shaughnessy asset management. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of positive sum or O'Shaunacy asset management. This podcast is for informational purposes only and should not be relied upon as a basis

Starting point is 00:00:48 for investment decisions. Clients of positive sum or O'Shaunacy asset management may maintain positions in the securities discussed in this podcast. My guests this week are Jeremy Levine, Kent Bennett, and Brian Feinstein. They are partners at one of the oldest and most storied venture firms in the world, Bessemer Venture Partners. Our conversation is split into two parts. First, we explore Bessemer itself. It's over 100 years old and has a unique operating model with lessons for every investment firm in the market.

Starting point is 00:01:19 We then discussed Jeremy, Kent, and Brian's investing styles and outlook. What they look for in businesses, their thoughts on various sectors like vertical market software, and we close to the discussion about AI and defensibility. Please enjoy this great conversation. The first voice you'll hear is Jeremy, then Kent, then Brian. And Jeremy, Brian, thank you so much for doing this together with me. I don't think I've ever done three people from the same firm at the same time. So this is kind of a fun experiment. And when we talked last time, we figured we'll give this a whirl and see how it goes. And I think

Starting point is 00:01:49 in this case, it's going to be really fun because Bessemer is such a story firm. It has such a long history. It also seems to be among the firms, the one that is most willing to evolve with the times, change its investing style, change its stripes, change, how it thinks about. about presenting itself to the world really easily through time. And so I'd love to spend the beginning chunk of our conversation on the firm itself, its history, its unique history, and why that allows you to do interesting things as a partnership. Jeremy, maybe since you've been there so long, you could give the brief overview of Bessemer, maybe up to the point that you joined, and then we can all talk about the period since.

Starting point is 00:02:29 The truth is a bit like an oral history, I suppose, because none of us was around or even born when what became Bessemer Venture Partner started. But as it's been told to me, and therefore probably been massaged and edited over the years like all good oral histories, and therefore maybe 10% of it is still truly true, when Andrew Carnegie, Henry Phipps and Henry Frick started Carnegie Steel together, they were three Scottish immigrants and they commercialized the Bessemer process, which was invented by another Scottish individual. And if you went to grade school in the United States, you probably studied the Bessemer process in eighth grade U.S. history, which was essentially a way of making steel in high volumes really cost effectively. And these three entrepreneurs figure

Starting point is 00:03:13 how to actually build it. And it worked really well. And then eventually, I think by 1907, as Carnegie Steel was humming really well, J.P. Morgan got the bright idea to buy up all the steel companies, aggregate them into a single company called U.S. Steel. And essentially, create a monopoly, which of course led to the Sherman Antitrust Act, but that's a story for another show. And I think Carnegie ended up making $400 million in 1907, which is an ungodly sum of money, even in today's terms, but 116 years ago, it was unfathomable. And Phipps and Frick both did really well as well. Fips is essentially the patriarch of the firm that we all work for today. He took some of that money and put it in a special trust for his descendants with the goal of creating a really

Starting point is 00:03:58 long-term family office. And he was a private individual in great contrast with Carnegie, who put his name on everything he touched. Phipps even named his own family company, Bessemer, in a nod to the brilliant inventor who created the process that generated all of his wealth. And Phipps was a pretty good investor, too. And he ended up investing in what he called high-risk, high-growth opportunities in the early 1900s. And some of the first ones were companies like Ingersoll Rand, International Paper, and W.R. Grace. And eventually, over literally 60 years, the companies migrated from industrial companies to technology companies as the IT industry was born in the 70s. And so it was essentially the high-risk, high-growth department of a family office became Bessemer Venture Partners

Starting point is 00:04:44 formerly in the 1970s or maybe even early 1980s. But there were basically a group of people, none of whom were family members who were working for this family office trying to generate really good long-term compounded returns investing in technology projects with enormous patients. And the family, who's a very close partner of ours today, invested in or with a really long time horizon, which suited them really well and ended up driving fantastic performance because not many people have that sort of patience to be investing in five, 10, and 20-year projects regularly. And it created this firm, which we're all part of today, with a very long-term time horizon

Starting point is 00:05:21 and a really stable capital base because the family understood that, yes, performance goes up and down through the years and through the decades, but in the long run, and they now have 116 years of performance history to understand it really well if your patient as an investor.

Starting point is 00:05:37 So that's the short background. Today, we are about 200 people. We are about 20 investment partners who continue to work in the same way that our predecessors did for many decades for this one family, although I'd say we have many causes that we support. They are essentially our LPs, other family offices, foundations,

Starting point is 00:05:56 charities, hospitals, universities, the usual suspects. And we work in a highly disaggregated, empowered manner where each partner or a team of folks can go make investment decisions as long as they're putting up a meaningful amount of their own capital and our LPs capital at the same time to hopefully invest in the next great technological invention or trend or business. So that's sort of best in a nutshell. maybe Cantor Brian can supplement that with more interesting tidbits.

Starting point is 00:06:24 What I might chip in to your starting question is there is a ton of historical irreverence. While we really benefit from that history and this strong capital base, we're also feel constantly free to rethink the entire model because nobody here owns it. There's no founder who we look at and think, well, you don't want to offend so-and-so in the corner office by questioning the way we did that thing. Our sort of historian-in-chief when I joined was a partner named Phila Hardiman, and he was actually a professor at the business school at Harvard, and he's just a hoot. And he was the first to make fun of the traditions that we had and just constantly question everything.

Starting point is 00:07:03 So I think there's a long tradition of us saying, this is how we've done it for years and years. It's probably somewhat wrong. Let's rethink it all the time. And in the 15 years I've been here, our operating system has changed in several major ways. We spend a lot of time studying the, I guess I'll call them like the heritage brands, the Hermezes and the Rolexes of the world. And I've always been interested by the degree to which time cannot be accelerated. Like you need a certain amount of history and time to really build these incredible brands. Maybe someone can do it faster.

Starting point is 00:07:35 But for the most part, these brands are extremely old. I wonder if there's something like that and whether or not you think there's something like that at play here with Bessemer. Venture as an industry isn't that old. And there just really aren't that many investment firms that are as old as this one. How does that heritage impact the organizational structure, the partnership, how things work, who's attracted to come work there? Does it have a big day-to-day year-to-day impact or had it been started 20 years ago would be roughly the same place?

Starting point is 00:08:05 I think the important part is that there's no owner here who is the founder. I don't think people sign up to work here because Bessemer is 116 years. old. In some ways, that can be like a bit of a marketing challenge because cool young entrepreneurs don't necessarily want to emulate what was done in the steel industry days 100 years ago. But I think the much more important thing is that we have an operating system here that's just set up to sort of endure past any one individual. And that leads to lots of healthy characteristics of how we recruit people, how we develop them over time, how we promote them. And all of that, I think, is really stabilizing stuff that makes the machine work. We just feel incredible.

Starting point is 00:08:45 fortunate to have been given this platform by the prior generation. When someone hands something over to you, you're the beneficiary of it and you're able to build a great career and do incredible things. You just feel enormous responsibility to hand it over to the next generation. And I think that's how the longevity comes into play. Do you describe literally how the firm works then? I'm curious about without a founder, really even in the visible history of the business, a leader, right, that started the thing.

Starting point is 00:09:13 how does then a group of 20 partners organize themselves and get things done? You read about like Valve software or something like there's interesting examples of different organizational structures and how they work. What's Bessemer's equivalent? I'll start. I think Kent's point that no one owns it is matched in importance by the fact that no one runs it. And by it, I mean something really specific, which is the investment decision-making process. And so we essentially split everything into one of two buckets.

Starting point is 00:09:43 the core of investing, actually making investment decisions, and everything else. The everything else part looks much more like a what I'll call a traditional company. We call it our infrastructure team. It's run by a person. Her name is Sandy Grippo. She's fantastic. And it has the functions of a more typical company. There's a finance function and a marketing function and so forth. On the investing side, it's pretty unique in that nobody is in charge. And so one of the things I like to mention to entrepreneurs when I talk about this is that in the 22 years I've been at Bessemer Venture Partners, I've seen one investment, one actually get shot down. We have this investment committee. And by the way every partner is a member of the investment committee and every partner has an

Starting point is 00:10:27 equal vote. The way we vote on things is actually on a scale of one to 10, where as long as your average vote is 5.5 or higher, you have approval. So we're not really asking anyone for permission to do something, we're asking each other for feedback on what we individually want to go do. And so we feel the responsibility and the authority to make the decision. And we want to tell each other like, gee, Brian, or G. G. G. Gemera, that doesn't look like a very good decision you're making. I might vote that a six or a five. I'm not trying to tell you you can't do it. I could. I could try to vote at a zero, but frankly, even with a zero vote from one partner, the average which can still be well above 5.5, so there's formal approval. But I want to tell you,

Starting point is 00:11:12 honestly, I wouldn't do it if I were you. And so we want the person who's closest to the decision who has the most information to make the actual decision. And then we just hold each other to accounts. We write down, who made the decision. And sometimes we'll collaborate in pairs or more. We'll share the responsibility and the credit and the blame. But we write down at the time, who made the decision who gets the credit and the blame. And as a young person growing up that's wildly exciting and empowering because you don't have to sort of work for the man or the woman in charge forever. You're essentially establishing your own track record and reputation. And at the same time, when it comes to working with entrepreneurs and the decisions related to their company,

Starting point is 00:11:52 you're empowered to help them too. You'd have to run back to home base and ask the boss, hey boss, what should we do in this situation? You decide because you're responsible. We've been able to maintain that idea of individual responsibility with collective support for a really long period of time. It's one of my favorite parts of my week is when you meet with an entrepreneur and they say, you have a great conversation. It's the beginning of potentially a long relationship. And they say, okay, so what is your process from here?

Starting point is 00:12:20 And what they mean is, I understand. I'm going to have to go jump through some series of 58 hoops and go shake a bunch of hands. And we get to say, like, there is no process. This is the process. We're going to meet. I'm going to decide if we want to invest and say yes or no and we'll go from there. they feel very relaxed at that point. And then when you're sitting on a board of a company and the company is facing a tough

Starting point is 00:12:41 decision and co-investors around the table say things like, I have to go ask my partners what I can do here. It's just like an insane statement. It's like, well, you're in the room. You know the situation. It's really empowering a customer to be able to say, okay, here's the right thing to do and make sure we can do it. What about on the incentive side?

Starting point is 00:12:59 So I understand the decision making process and the individual autonomy and the power that that brings and the sort of partner that might attract, what about the reward on the back end? So if I'm the one that brings Shopify in the door or something or Twilio or one of these great investments that Bessemer's made over the years, how are the rewards shared amongst the partners? And what have you learned about tuning that incentive structure the right way over time? It has evolved quite a bit over time. And honestly, none of us is privy to what happened 85 years ago, but the core principle is the following. Almost everybody at Bessemer has grown up at Bessemer. That's not entirely true, but I'd say three quarters of the partners. And that means we started like

Starting point is 00:13:40 Brian did as an analyst when he was maybe 22 years old, or I think I started in my mid-20s, I suspect Kent was in the same zone, and we grew up at the firm. And so we generally give people the authority, the checkbook after we've known them for five, seven, nine years. So there's a high level of trust and confidence in people's individual investment judgment by the time they're actually writing a check. And then what we essentially say implicitly is there's this great platform that has been created by generations of people who've come before you and you're now standing on it with the ability to screw it up or make it better, hopefully make it better. What we're going to do is we're going to reward you largely based on the performance of your investments for the first

Starting point is 00:14:25 number of years in your career. And there's some shared common ownership, if you, of the whole enchilada as well. But your thinking is going to be dominated by your own investments. And we think that's really clarifying for a young investor because they can focus, proving they can do it, if you will. And then as they perform, if those investors turn out to be the next, say, Shopify, they'll do really well financially. But the firm will also do really well. And they'll have essentially earned their share, if you will, to a co-equal ownership, which is what everyone can get. to over time. Once you've done that and you've demonstrated really meaningful individual success, you become a true owner of something, not like a fake owner where there's really someone else

Starting point is 00:15:05 who's still in charge, who still you work for, but a true co-equal owner. And that's what I think keeps people at Bessemer for a really long time because you're not thinking like, oh, I can keep working for so-and-so who started this place or who got here before I did, or I can leave and start my own firm. That doesn't happen very often because there's not much incentive to do it when you feel like you can be a true, co-equal owner. That's the system, and that's essentially how we both try to balance giving people autonomy responsibility, but not necessarily saying, here, you've inherited the earth on day one. There's a little bit of go prove yourself, and those who are up for proving themselves

Starting point is 00:15:39 tend to love it. How far does the autonomy go on the style of the investment itself? Because one of the things that seems true about Bessemer is that the sorts of investments that it will make and has made, has evolved a lot of retirement, constantly change. If I came as a new partner and I brought a late stage buyout type deal or something that really clearly wasn't like a venture style investment or something like that, what reaction would that be met with? How much are you pushing people to really pursue their own unique individual investment strategy versus conform to what the charity that is an LP of Bessemer thinks that you guys do?

Starting point is 00:16:17 I'm always interested in style drift and how much of that is good versus bad and how much the firm encourages versus doesn't. So how does Bessemer manage that sort of thing? Patrick, I would say we encourage more autonomy than almost any other venture firm, both with respect to decision making and style. That could be a speculative seed investment in a team, back of a napkin, all the way through a late stage growth buyout in a mature business that's generating profit across every area of high technology. I mean, the breadth is truly extraordinary. A lot of folks associate us with software.

Starting point is 00:16:59 That's the bulk of the portfolio. That's the bread and butter. But we have investments in life sciences, in fintech, in frontier tech, runs the gamut. The autonomy is important because conventional wisdom says you can't scale venture capital. And that's true in a small partnership with a consensus-driven. decision-making model, because if you try to scale that and try to arrive at consensus, you either end up with mediocre decisions or the decision-making process breaks down because not everyone can meet the company.

Starting point is 00:17:29 We like to think we've cracked the code on how to scale venture capital by enabling each individual partner to have that autonomy to pursue whatever investment areas they want to pursue and make the best decisions they possibly can. There's an important part of the operating system we haven't talked about yet, which we call the roadmap internally. So it would be very odd behavior if I showed up to a partnership meeting and said, hey, everyone, I'm investing in a biotech company because the group knows I don't know a thing about biotech, and they would think I'd lost my mind.

Starting point is 00:18:01 The behavior we have here is roadmap-based, which means every partner will pick up over time a number of essentially themes they're interested in. We call them roadmaps, and we go off and do a bunch of homework on those themes, typically, before we make our first investment, and we start talking to each other about them. At our off-sides and other periodic meetings, we'll present a roadmap. Outside of the context of any investment, we'll say, gosh, I'm interested in vertical software. Here's what I'm seeing there, et cetera. And then when you bring in a company that's squarely in that roadmap, your partners know

Starting point is 00:18:30 that you're within a lane that you've thought a lot about, certainly somebody could branch into like a new style of investing. So let's say you wanted to do a ton of seed stage investing with like a very a high frequency, but you would typically present that to the partnership as a strategy first and have a fierce discussion on that and that strategy and then begin to execute on it. So it's not as if you're just showing up and saying, trust me, I'm doing something you've never seen before. It's going to work out. There's a little bit of preconditioning. What's shared in common amongst the best roadmaps that you've seen presented across your

Starting point is 00:19:05 collective histories at Bessemer? Like, what is it about the roadmaps that are really effective? I thought a lot about this because we did a retro analysis of all of our best roadmaps, where they came from and what worked about them. Typically, I would say the best, but not all the best roadmaps, but most of them are positioned around some big change in the technology landscape. So cloud investing being probably most productive roadmap starting 20 years ago. So they start with the tech change as opposed to starting with a problem space. So if you start with aging as a problem and you say, gosh, aging is really bad and mortality, that's bummer. I'm going to go look for technologies to solve that problem. It's like, well, that problem's been around for a long time.

Starting point is 00:19:46 Nothing has really changed to make that better. Whereas if you start with large language models and, gosh, what could they do to make baby boomers lives better right now? You're starting with a tech change and bringing that solution and pointing into a problem. That's really successful. The other characteristic of successful roadmaps tends to be that we didn't, come up with the idea without any real-world exposure. And so our best roadmaps, it's a little bit of a dirty secret here, but have come from an investment we made where it started going much better than we expected. And we sit back and say, like, I thought this was going to work, but not this well.

Starting point is 00:20:21 This is going amazing. Why? And it tends to point us to some go-to-market strategy that's part of that investment that's like driving the sort of outsized success. So a lot of our roadmaps, we get inspiration from the portfolio where we accidentally stumble in. into some new tech trend that we didn't maybe even see up front. As part of that, there's often a core insight that's a little bit contrarian. When Jeremy first started investing in user-generated content, he had this insight that content will generate traffic, which will in turn generate more content and create this viral loop

Starting point is 00:20:55 that you can ultimately monetize through advertising. And that led to his investments in Pinterest and Yelp and LinkedIn and all these great user-generated content companies. When we start investing in vertical software, we have this small core insight. Yes, the markets are small, but you can often trade market size for market share and build these industry-defining companies that end up capturing 40, 50, 60 percent in their markets and ultimately become really valuable. When our partners, Byron and Ethan and David started investing in developer platforms,

Starting point is 00:21:27 conventional wisdom said, wait a second, there's no screen. There are no buttons. It's just an API. You can't build a business this way. And so there's often a bit of a contrarian point of view or a little insight that we exploit and mine and pursue, and that drives the roadmap activity. One of the really interesting thing, which comes back to some of the firm governance and structure that is connected to what Brian just said. None of us own any of the intellectual property or ideas, not the stages, not the geographies, not the roadmaps. we share them with each other to get feedback. And then when someone says, wow, this is working.

Starting point is 00:22:07 And a great example, I think Brian was an analyst at Bessemer when he discovered MindBody, and I was a partner at the time, and we ended up investing in the company. And we then realized, wait a minute, mind body started to offer payments as a product. This is an unlock that will drive really significant revenue growth. Let's go find every other company that might do something similar in another category. MindBody is software for yoga studios and at fitness clubs and like, shortly thereafter, I think, nine different partners at Bessemer were all like, that's a good idea. And we end up investing in ProCore, Toast, X-Time, Clio. I think literally we have Devondy vertical software companies invested in by nine different partners.

Starting point is 00:22:48 If one partner owned it was able to call it off limits, we would have been maybe 10% as productive. But because it's an organized free-for-all is maybe the best way to put it, we tend to end up helping each other by sharing the ideas. and then benefiting from each other's execution against the best ideas. The challenge is that sometimes I think we're all relatively smart. We get attracted to the single best ideas, which may end up with a slightly more concentration of the portfolio in one spot. But honestly, I think we'd rather have that than be forced into these categories or groups, trying to find that best investment in crappy area is a much less fun and less productive endeavor

Starting point is 00:23:25 than shifting the resources of the firm automagically, if you will, to the most promising opportunities. If there ever is major conflict then inside the partnership, what is the nature of that conflict under this incentive structure? Is there already? We certainly have had conflict. How can you not? If you have more than one human in a room for decades,

Starting point is 00:23:48 you're going to have conflicts. What we actually do to help resolve that is, one, having no one in charge and having each of us able to talk to one another to kind of get through problems is productive. but we've also had in most of our internal and significant meetings for years, more or less since I've been at Basmer, so at least 20 years. We've had an independent psychologist or coach in the meetings. This interview has just changed over time, but they tend to have some longevity, so they understand us, they get to know us. They have no ownership in the firm whatsoever. They have

Starting point is 00:24:17 nothing at stake. They're just there as an independent to help referee anytime there is some issue or attention. And sometimes having an outsider who knows you call you out on something is really effective. It's one of many things we've adapted or adopted over the years in order to deal with an occasional strife. We're not perfect like any family we have our own dysfunction. The other thing I'd point to, and hopefully this doesn't sound to Kumbaya, but we did our values exercise several years ago. The number one value, like if you pulled everybody at Bessemer and said, what's a top value at Bessemer? Everyone's number one, essentially, was intellectual honesty. It was like the first thing they said. And it's something you notice very quickly if you join Bessmer

Starting point is 00:24:55 is that people just say what they are thinking. There's zero politics. There's zero salesmanship. And that as a core part of the operating system is critical for really avoiding conflict because if you're just saying what you think, you're not advocating for your own portfolio, your deals, et cetera. Like the conversations just get direct to the issue and the resolution with much less friction than some of these organizations we hear about that are much more political where people

Starting point is 00:25:20 have favorites. And we just don't have any of that, which is such a massive relief for all. all of us and I think makes life much more pleasant. I'm loving these answers because the ideas are so generalizable to any kind of investment partnership and firm. Like it doesn't need to be early stage. So I have a few more. I'm interested in the same sort of question that I asked on roadmaps for the concept of the sort of apprentice model. I remember talking to Sarah Tavill about this when I first met her many years ago and kind of learning about the way that Bessemer, like you said, you've all started there in your 20s, tends to get talented people very young and groom them for a long period of time before

Starting point is 00:25:55 even giving them the checkbook, in that five to nine year period, what works? What have you found to be the most effective way to run this program that you think maybe others could borrow or steal? It's just a lot of time in the same room together. I sat next to Jeremy on two boards before I was a partner. I sat next to Bob Goodman, one of our senior partners for 200 hours of board meetings and other conversations and then Felda Hardiman, my first mentor, I spent several years in the same room with him, just listening to him, talk to companies, and occasionally chipping in at first, and there were time building confidence. So I think that's the core part. There's no shortcut to that, but boy, you learn a ton over time and you just take for granted how much you've

Starting point is 00:26:38 picked up along the way. I think there's a second elements to it. And I'll tell you an anecdote I got from, at the time, a young associate. And she had just joined the firm. And I was doing like a half an hour, welcome to Bessemer. I hope you found where the bathroom is located, introductory session. I asked her, like, is there anything different relative to what you expected now that you've been here like a month? And she said to me this thing that's shocking. And I've told this story many times. And she said to me, yeah, there is. I can't believe how honest the investment memos are at Bessemer. And I was like, what do you mean? Honest. And she had been at another investment firm before she joined us. And I asked her, I'm like, did they lie in the investment memos at a journal firm? She said, no, no,

Starting point is 00:27:19 no, no. It's just that she realized within just a few weeks, this, I think, really critical insight, which is that when we write these memos, we're not trying to convince somebody what to do. At our old firm, there's really like two people who kind of ran the investment committee, and the goal was to convince them to give you permission to make the investment. So, like, yeah, you talk about some weaknesses or risks in your memos, but it was kind of like answering that silly interview question that you might get when you're interviewing for a job, and someone says, tell me about your weaknesses. You're like, oh, I work too hard.

Starting point is 00:27:46 It's like a sort of a bullshit weakness that you're saying because you have to have some weaknesses or risk, but it's not real. And in contrast, she said at Bessemer, it was very clear to her, people were writing in the level what they really think because they're not asking someone else to make the decision for them. They're explaining their decision. And the teams that work on investments at Bessemer are typically a partner and an associate and sometimes an analyst. So it's two or three people.

Starting point is 00:28:09 So it's not like there's nine people that you can hide behind. It's a really small group of people who's saying like, this is what we think the right answer is. And you feel this responsibility. Again, you're not trying to convince someone. And so you end up learning much better when you think you're making the decision for yourself than when you're trying to convince someone else to do something, but it's really their decision because you feel this responsibility.

Starting point is 00:28:29 It's empowering, but it's also scary. And most people, when they feel responsible, end up doing better work and end up learning faster. And we should also say that everybody's, then all the investors are invited into our Monday meeting. And so they all join that meeting. They all read those memos. And then because we're not selling your potential investments at that meeting, you're just discussing them. I feel much more open to tell Jeremy like, hey, I read your memo and I think you totally missed something. This is how I would have thought about it. And that's not really threatening or offensive

Starting point is 00:28:57 to him because he knows he can still do whatever he wants. And so we can have this open intellectual conversation and everyone is there to hear it participate in it, tell me I'm wrong. There's just a ton of that open dialogue in that couple hours we spend every Monday together. And that's massive learning for all of us, even those of us who've been here for a long time. The irony is our three most successful vertical software investments, Shopify, Toast and ProCore, were all relatively unpopular investments when first discussed in the partnership meeting. And our autonomous decision-making model made those possible. And when you're a young investor, you see that dynamic. You put yourself in a different headspace. It's not about doing what's popular. It's not about chasing momentum.

Starting point is 00:29:42 It's not about pursuing the consensus opportunities. It's about developing a thesis and pursuing something that may be a little off the beaten path. I'm sure that during this apprentice period where someone's effectively with a partner all the time, that there's this development of taste that happens, some of which I'm sure is borrowed from time spent with the partner, but some of which in a perfect world becomes their own. I'd love for each of you to describe your taste in companies, what that term, brings to mind and maybe even a little bit about how that taste was developed. It's funny because I think we have three different tastes.

Starting point is 00:30:19 If I had to project on the two of you, I'll start. Over time, Patrick, mainly as I've made some good but some bad investments, all of my learnings have come from my mistakes. And as I think about them, I've slowly honed in on when I've gone wrong, it's that I haven't started with just very clear radical product advantages. So the investments that worked have had such clear, better products than the existing alternatives at the time of our investment. Almost all of the ones that have failed, that was more of a question mark.

Starting point is 00:30:54 Either the product wasn't fully formulated yet or it was out there, but it wasn't just a slam dunk in the minds of the people buying it or the consumers using it. I've just honed in tighter and tighter and tighter on how can I get to the truth of, is this product radically advantaged on day one? and try to eliminate a lot of other noise so you can talk yourself out of any early stage investment. It's like really easy to get scared about something. You can find something to be scared about. But if the business has this strong fundamental product value on day one,

Starting point is 00:31:25 try to like screaming for the other noise and not scare myself because the market size or there's some regulatory concern that could come up that some of those noisy things that I think have historically talked us out of some potentially great investments. So that's my taste. I think Jeremy and Brian are very different. you guys fill it in for yourselves. I hear Jeremy's answer. I've never asked him this question. Okay, I'll go next. The thing that I typically fall for hardest is wild capital efficiency.

Starting point is 00:31:50 I love investing in companies that don't really need my money. I think that's the dirty secret of all investing. And that was really true for a lot of the companies in which I invested, although ironically, I think the most successful investment I've been a part of from a penny's returned is Pinterest. And Pinterest, we invested in the Series A. which was maybe like an $8 or $9 million investment, ended up raising a billion dollars of equity capital in public, which is the opposite of highly capital efficient. And so even sometimes you get led away from your original tastes,

Starting point is 00:32:20 and I think if you're not willing to have a bit of flexible mind, you're going to miss a lot in this business. Even with flexible minds, I think we still miss a lot. As evidenced by the anti-portfolio we've been on our website, I fall hard and fast typically for capital efficiency. It just solves for so many potential errors down the road. From a business perspective, I think I over-index on path to market leadership. I cover vertical software, and in these vertical markets, there are these virtuous cycles.

Starting point is 00:32:47 You're the market leader, you're the number one, number two player. You get all the treasure. And so there's no prototypical path to get there. I want to have a conversation with the team and understand how they see their path to market leadership. As it relates to the founders, I think I index away from like prototypical 20, something-year-old tech genius who dropped out of Harvard or MIT. Most of the founders I've worked with look a lot like Tui, the founder and CEO of ProCore, older, came from industry, built his business

Starting point is 00:33:19 in Santa Barbara, not Silicon Valley, just deep empathy for the domain and deep caring for the customers. I think particularly in vertical software, that translates to long-term success. Kent, when you're evaluating a product, given your preference for like obvious product leadership or 10x better product or whatever, what have you gotten really good at to be able to do that effectively, especially if it's not a product that you personally would be a user of? I really don't overthink it. I think that's what I've gotten better at, which is recently invested in a company that's serving the construction industry. And I am not a contractor. I can barely replace a light bulb. So I immediately got on the phone. and talk to 10 contractors and said, hey, here's a product. I'm going to describe it to you. It was a very simple product to describe, which was it runs your payroll, your current payroll service, but it also provides for you a delightful piece of software on top of your payroll that runs all the compliance and the union reporting and all these things. And that is free.

Starting point is 00:34:20 And so I pitched this to 10 contractors and six of them were like, where is it? Can I sign up for it? I need it. And so that signal so clear, I understood the economics of the product with the lifetime value or the annual value could be, and I understood that I had basically just generated a pipeline with an afternoon that was going to be an excess month of the value. So that's the clear signal. I don't overthink it, and I recognize that if I engage in 20 deep discussions, I'm only going to get that level of signal one or two times, and that's when I should invest. And if the signal is fuzzy or I can't understand it because I can't empathize and I can't find customers to tell me they love it, then it's just not for me. Jeremy, when you're assessing

Starting point is 00:34:59 capital efficiency, especially as you get earlier in the spectrum and there's less historical financials, there's less to go on overall. Maybe there's not even like a mature product in pricing yet. How do you think through that problem? At the earliest stages, capital efficiency shows up in just a scrappiness, a willingness to do things that are the opposite of gold-plated. In fact, a slightly related anecdote. So within our partnership, as we've told you, we've got 20-ish partners and we all have different views, there's been a great debate over the years as to whether our offices should look like that

Starting point is 00:35:35 of your typical dentist or if they should look like that of your typical fancy investment bank. I've always been a fan of the dentist's office, low-key, unordained, because when you work with scrappy entrepreneurs and they see you working in a scrappy environment, they're like, oh, okay, I get it.

Starting point is 00:35:51 Of course, the counter is, you want really fancy offices to show people how successful you've been so they want to work with you. But for me, I'm all about the scrappiness end up the early stages. You can see that in just the environments, the people that the team hires. They're typically less experienced and risky or in cheaper hires.

Starting point is 00:36:07 But if you get really talented people who are less experienced and risky early in their career, they end up growing into the next Toby Lutki, the CEO of Shopify. And I think too many entrepreneurs misunderstand the objective, which is like how do you build something really great without burning lots and lots of money to do so? Because burning lots of money means you've got to spend more time raising money. and you end up owning less of it in the end because you had to sell more of it off to get access to the capital.

Starting point is 00:36:32 And I'll tell you one other related anecdote, which is when Shopify was preparing to go public, they were doing like a non-deal roadshow, which is basically when you go talk to a bunch of investors privately, not for the purpose of raising money in an IPO, but to try to get your story out well ahead of an IPO. The then-CFO, Russ Jones, was showing me the deck he helped prepare

Starting point is 00:36:50 and he was focusing on the fact that Shopify had raised $100 million. And I'm like, Russ, that's not something to advertise. In fact, you're missing the point, which is that Shopify actually only burned $2 million in its entire lifetime, because of the $100 million it had raised, I think something like 25 was for secondary selling of existing shareholders, and the other 75, of that other 75, 73 was still on the balance sheet. That's what you should tell investors. You built this incredible business while burning $2 million. That's an accomplishment. Unfortunately, the media headlines are often about how much the company raises. But in fact, the true

Starting point is 00:37:27 headline that you want to understand is like, well, how big are you relative to how much you've raised or how profitable are you relative to how much you've raised? But of course, those stories rarely get told. Brian, on the go-to-market side of VMS businesses, talked about path to market dominance or market leadership. What have you seen the best in the world at accomplishing that do consistently well? Every story has its own idiosyncrasies. I think in general, it's much easier to build a great business quickly if you're going after at Greenfield opportunity. Shopify catering to online retailers who were coming online in a way that we've never seen before. ProCore catering to construction companies that suddenly had internet on the

Starting point is 00:38:07 job site and mobile devices for workers to interact. Being able to deliver software to an industry that's never had it before is the ideal setup for market leadership. Now, there are exceptions, and I give Kent a lot credit for the Toast investment because Toast was attacking a market that was fully vended. At the time, you had NCR, you had micros, you had countless restaurant point-of-sale systems, so they did not have a greenfield opportunity. And in those situations, what we've observed is the path to market leadership requires a dramatically better product, 10x better, or dramatically cheaper, 10x cheaper, in order to cut through the noise and drive that replacement cycle. And it's not easy to do. Ken can tell you the toast story, but my understanding is they did both. They showed

Starting point is 00:38:56 up with an incredible tablet-based product in an industry dominated by incumbent crap, and they were able to sell it much cheaper because they were the first to bundle payments with the software so they can underprice on software and make money off the payments. So they're both better and cheaper. That's a way to cut through the incumbents, but otherwise, I would recommend trying to find Greenfield opportunities. Absolutely not enough to just show up and say, hey, This product's in the cloud. Nobody cares. You're running a restaurant.

Starting point is 00:39:25 You don't care what the cloud is. It's not doing anything for you. But if you show up with a tablet that is 10 times cheaper because it's built on the Android platform than the $20,000 legacy thing they had to buy, and it's more functional. And you're told that you can change your menu when you're in bed that morning

Starting point is 00:39:42 and you don't have to drive in. And it bundles in the payments, which you didn't necessarily even know the name of your payments provider historically. And by bundling in payments, all of a sudden, you can look up that receipt from three weeks ago instantly in the POS. All that comes together and it's just a complete no-brainer.

Starting point is 00:39:57 And that's when you get path to market dominance and market they may have some incumbents. Again, it comes back to product led for me, but it's funny to think that, Brian, your path to market dominance taste, I think could have gotten you there with toast. I mean, I think all these things converge. Honestly, it's three sides of the same coin because when you have that incredible product, it leads to efficiency. Like, I think all of Jeremy's efficiency stories are also stories of incredible product that made the sales, payback, unbelievable, even Pinterest. Yes, it was inefficient, but they drove hundreds of millions of users without spending a dime on the user demand. It was like incredibly efficient growth of the consumer engagement. Yeah, put another way, if you've got an incredible product going after

Starting point is 00:40:36 a greenfield market, you're going to check Kent's box because you're going to talk to customers and it's going to fly off the shelves. You're going to check Jeremy's box because you're going to be picking up the phone and taking orders, and you're going to check my box because you'll be able to articulate a path to market leadership. in this period now we're spring of 2023 and the period of furious outrageous quick deal cycles and huge funding amounts is a bit behind us. We have a bit more perspective on the sort of madness that was the 2021 market with benefit of having gone through that but also invested for a long time. Do you think that our industry has the right tools, specifically like a preferred equity style

Starting point is 00:41:17 of investment, a preference stack. Do you think that that continues to be the right primary instrument for investing in these kinds of companies, given that I have to imagine some part of the run-up in valuations was due to this preference being in place? I'm just curious if you think that we're properly equipped for the next 20 years of funding these kinds of companies with the traditional toolkit that we've had. I'll offer an opinion, Jeremy, correct me. But I don't think the high valuations were necessarily driven by preferred stock in many cases. I think for some of the pre-IPO rounds where you had a structure that would be repriced in an IPO effectively, maybe some of the later stage investors were doing that math.

Starting point is 00:41:57 But for many of us, I think we've come to realize over the years that value of our preferred stock to protect us in a downside scenario is actually not all that material to our fund returns. And so actually, we're thinking about the returns. Most of the return profile is driven by the share price. not necessarily that it's preferred stock. Problem if we were to invest in another vehicles, there's a collective action problem. If all of our peers are just investing in preferred stock

Starting point is 00:42:23 and we buy common stock and then we're buried under a mountain of preferred, that doesn't feel good for us to somehow pioneer to what end. And so I think there's a little bit of a just perpetuating cycle. That's the standard. But there are other factors at play, certainly, that led to the valuations aside that structure. I agree with your conclusion. I think there's two other comments to make.

Starting point is 00:42:41 One is, I think investors, on average, paid truly stupid prices most of the time in 2020 and 2021, and that they used any excuse that could come up with to justify those prices, and maybe the existence of preferred stock was one of them, but it certainly didn't help. I think the other comment I'd make is that preferred stock really makes sense in the early days of a company, because you have this company, which might be an idea on a napkin, and you're investing, say, $3 million for a 20% stake, and therefore declaring, that the company is worth $15 million, it can't be worth $15 million. It's an idea on a napkin and a couple people.

Starting point is 00:43:17 So the construct of preferred stock, I think, was invented by somebody cleverly to say, look, entrepreneur, if you go and sell the company tomorrow for $2 million, you got to give us our money back. You can't just give us 20% of $2 million because all you did then was eviscerate a million of value by turning three into two and now you made money. It makes no sense. So the preferred stock construct makes sense to ensure interest alignment in the early stage of a company. And then at some point, it becomes irrelevance, but precedent is set from

Starting point is 00:43:47 that early investment in how almost every future round of investment will look like. So people tend to just take the last structure that was used to make the last equity investment in a company, copy it, change a few of the numbers around, and then hit print. And so I think the reason why we have it at all bases is because it makes sense early. It's almost necessary for alignment, and then it just sticks. One of the best predictors of future return in public equities is just some sort of normalized prevailing valuation. There's lots of measures you can do it with price to sales or book or normalized earnings or something. Pretty predictive of what forward returns are going to be. And I'm curious what you think the forward perspective

Starting point is 00:44:25 returns are for this style of early stage investing today versus the 15 to 20 years that you've all been doing this in the past. I read that mind-body memo in preparation for today. And one of the things that pops pretty quickly off the page is that two and a half or three times ARR multiple that was paid for that equity back at the time. And it was 12 million bucks of revenue or something like that. It was a real established business that now that would have commanded a multi-hundred million dollar valuation a couple of years ago. So valuations matter for returns, even though sometimes the worst mistake you can make in this style of investing is not being willing to pay a rich price, but there's too rich of a price that can ruin any investment. And I'm curious how you all think about

Starting point is 00:45:05 the prospective return opportunity in this style of investing over the next 10 or 20 years. It's lower. That I can say with confidence. Because as you said, when the entry prices are higher, the returns have to be lower. There's a bear and a bull argument on today's prices. I think we're generally in the bull camp. So I'll tell you the bear argument first. The bear argument is, gee, when you're paying higher prices, which is, by the way, driven by the fact that there's much more capital out there, more venture funds, more growth equity funds, more firms who are at scale or scaling up, and you have to compete to invest in the best companies, and therefore you're paying higher prices, therefore the returns have to be down.

Starting point is 00:45:44 And that's the bare case. And I think there's real credence to that. In fact, I think that's true. The question is whether it offsets entirely or only just partially the bull case. The bullcases may be best exemplified by the following anecdote, which is when I joined Bessemer in 2001 and started thinking all the time about tech investing, and I looked at what are the successful outcomes of great software and internet companies. I think there were maybe a handful of software companies in the world that had a billion dollars in revenue, in revenue, Microsoft, Oracle, HP, SAP, but it wasn't a long list. Maybe parametric technology had a billion back then. And maybe I'm undercounting by a bit, but not by many. And internet companies,

Starting point is 00:46:27 the most valuable ones were eBay and Yahoo. And I don't think either one of them had a $50 billion market capitalization. And if you fast forward to today, I can't even count on all my fingers and toes the number of companies that had a billion dollars in revenue, maybe even just from the Bessemer portfolio alone. And so the outcomes are so much bigger. And when it comes to internet, there are many internet companies with $50 billion market caps, including a few with trillion dollar market caps or maybe $500 million plus market caps. The bullcase is the market sizes and the potential outcomes for these companies is so much bigger that while sure, you'd always rather as an investor pay less than more, the outcomes allow you to afford stretching more on price. And of course,

Starting point is 00:47:08 if you take anything to the extreme, you get into trouble pretty quickly. And one exercise I go through all the time on an investment is if I take this deal and I pay 10% more, is it still okay? And the answer is always, yes, it's still okay. The problem is if you then take that number and you say, well, what if I pay 10% more than that? Is that still okay? If you do it enough times, you've gotten so far away from your initial investment that it becomes nonsensical, you could literally pay 100 times more. But each individual question, the answer always has to be asked because we just don't have enough confidence or precision to know what a 10% difference is. The conundrum, if you will, of investing particularly at the early stages.

Starting point is 00:47:44 In practice, the way that this manifests across the software portfolio is we have different flavors of ice cream. We've got growth at a reasonable price where your entry multiple matters, your exit multiple matters, the model matters. And then we've got investments where we're paying a hundred times revenue and swing for the fences. On the aggregate, it tends to work out. You've got this concept of best of the centaur, like a better unicorn because it's based on an actual fundamental metric, which is companies that have achieved that $100 million milestone. Why is that milestone significant and interesting T-wall? What does having, coined that term and written about it and thought about it due to the way the partnership

Starting point is 00:48:22 thinks and operates. I'm curious why it's useful. So a year ago, as we were reeling from what we'd labeled the Sassica, all the multiples are way down, we sort of took a step back. We'd remain, I'd say, reasonable players through the two frothiest years, not to say that we didn't do some unreasonable things as well. But we took a step back and it was kind of like, what the heck just happened to this entire industry? Where did we go wrong? And it was pretty obvious. We had all fallen into the valuation trap of unicorn counting and having press headlines tell us how great our industry was doing.

Starting point is 00:48:58 But what hadn't changed was that these businesses, especially these cloud businesses with recurring revenue, they remained incredible businesses. I think the greatest business model I've ever seen. Said the basic thing, which is like, what's an indicator of the overall health of this market, regardless of valuations? Let's count how many businesses are doing 100 million of revenue.

Starting point is 00:49:16 The reason we picked that number, a nice round number. Also, we couldn't tell you about a hundred million recurring revenue software company that had gone out of business in our recent career lifetimes. So it felt like a measure of immortality. And so taking that count of centaurs, first of all, we're like kind of heartened that the number last year was on track to be the highest number we'd ever seen. So despite massive collapse and valuations and unicorns, wow, this indicator of the health of this economy still looks pretty good.

Starting point is 00:49:45 and we just knew going forward. It's something we could count to ground ourselves in the reality of our new, future great software businesses still being born at a pace. Is that pace speeding up? Is it falling off? Truthfully, we counted it again this year after inventing it last year. And it was slightly down this year versus the year before, but it's still the second highest count number in the recorded history. So we still think the pace of fundamental creation of future great software companies is still about as high as it's ever been. It checks with our experience that We're seeing incredible companies founded every day, even in like a gloomy macro environment. The grounding is important because I think the biggest sin that was committed by our industry or asset class during the boom

Starting point is 00:50:28 was convincing ourselves that the value of the business was equivalent to the valuation of the last round. And the performance of a venture fund was based on marks set at the last round price. We all know the last round price is not the value of the business. it's one investor, one bidder, buying an option. You can't sell the unicorn for a billion dollars in many cases. And so the Centaur concept was a way to reground ourselves in what matters, which is the financial performance of the business. When I think about investment firms that do a good job on, I guess I'll call it the content side,

Starting point is 00:51:06 I think of Bessemer in first round as being really interesting firms. And having created a lot of concepts like the Centaur, like the Cloud Index, that Bessemer has created that's cited all over the place in the industry. You've done a ton of stuff like this. Like if you go on the website, there's just an embarrassment of riches of old memos to read and frameworks to read through and roadmaps and ways and thinking about different kinds of businesses. There's obviously a ton of effort that's gone into all this stuff. How does that help you? Like, why are you doing all that? Does it manifest as better deal flow? Is it a brand building exercise for Bessemer? It seems like everyone sort of feels like they need to do this now.

Starting point is 00:51:42 Bessemer's been doing it a long time. So can you give us a report on whether or not it's worth it? First of all, it just stems from the work we do to hopefully sharpen our investment judgment, and that's probably the biggest benefit. And often we will have an idea that we will circulate internally, and we won't tell the world for a while because, oh, gosh, we're paying attention to payments being bundled in vertical software. Like, why would we announce that the first couple of years?

Starting point is 00:52:07 Then the moment tips often where it becomes a little more obvious. And we make the decision to launch that content externally where we can earn some thought leadership. It may drive some incremental deal flow when we do that. I mean, we're not celebrities. We're all pretty nerdy, not particularly good at self-promotion. I think the most valuable thing it does once we've launched that content is that when we meet entrepreneurs who are doing something that we're going to like, they often see this

Starting point is 00:52:36 welcome mat as they're coming in the door that is a white paper that speaks to their business in a way that they're like really nodding their head. The number of times I've had a first conversation where there's just a mutual sense of, okay, you understand what I'm talking about and you get this. That's hugely valuable. It leads to some self-selection of the business models that are like,

Starting point is 00:52:54 Jeremy's going to understand my user-generated content business model and I think may have them select us versus another investor who may not have already had that thought or at least put it out there. The only thing on our website that we created just for the purpose of marketing is our anti-portfolio. And we try to be funny. The rest of it is really repurposed.

Starting point is 00:53:15 And sometimes we'll modify it or polish it or as Kent then delay it before we put it out to the public. But we're creating it internally anyway. And so it's a free opportunity to show people more about who we are. And it actually does lead to people reaching out to us because we publish something relative to them. Like, oh, I didn't even know you invested in that area. I want to talk to someone at Bessemer.

Starting point is 00:53:35 Speaking of which, the reason that four of us originally got connected with, was because of a vertical market software, that topic, my firm was doing a big deep dive into it. And Bestberg came up over and over and over again. Like, if you're interested in this topic, you need to talk to these guys immediately, having made a lot of these investments like you mentioned in ProCore and Toast and many others. How mature do you think that whole style of business building is? Because if I do some quick Googling between your site and Tidemark site and a few others, you can get these incredibly rich, incredibly detailed.

Starting point is 00:54:08 playbooks for how to build a VMS business. It feels to me from the outside quite mature in our understanding of how to build one of them. But I'm also curious how mature you think it is in terms of how many markets there are that have yet to be addressed by a VMS solution for a given industry. Obviously, you've made hey here. It would be hard to say that there's no more opportunity, but where do you think it is in its life cycle? I have 12 vertical market software investments. I actually counted before we hopped on the call. I was surprised at that number. Five of them I made in the last 12 months. So, but I would say that there have been some cycles of this model and what it looks like from mind-body, which started as pure SaaS initially, bundled in

Starting point is 00:54:52 payments. And then may have, by doing that, inspired and kicked off a next wave of vertical software where embedded financials were a big part of the business model. And we saw some derivations of that with business-to-business transactions being part of certain platforms. And that's another sort of embedded financial wave. Payroll, as we've had API-driven payroll opportunities, another wave that is powering some more. And then along the way, we've seen the percent of total transaction revenue going through these markets that the core platform that provides not only the software,

Starting point is 00:55:24 but maybe some of the embedded financial products can command over time. And as that ticks up from half a percent to one percent to one and a half percent, to 2%, it just continues to open up other markets that 10 years ago we might have thought were too small and now look like absolutely opportunities for billion dollar plus outcomes. I would agree with you that clock is getting late-ish. I'm sort of betting it's seven at night and we're not sort of at 11 yet. And then I look at this wave of language model-driven AI. We had a conference last week where we had a dozen plus vertical software leaders

Starting point is 00:56:02 in the same room, and they were all excitedly and furiously scribbling down notes of what their peers were doing to innovate on their products. And certainly for the embedded vertical software leaders, I think there's a huge moment of adoption of these language models, which I think power obvious things across that suite of technology. But it's also going to open the window for many verticals that previously didn't have a reason to adopt software strong enough to have a real leader. And then for some of the verticals with a sleepy incumbent, watch out. We could see another wave. One other thing to note, which is that if today cloud software gets the most attention within Bessamer and obviously other areas, biotech and healthcare and consumer and deep tech and

Starting point is 00:56:44 SpaceX, they all get attention also, but the biggest and most important category over the past, let's say, 10 years has certainly been cloud software. And as you pointed out, we're constantly worrying about when is the end. And not that these companies are going to go away or there'll be less innovation, it'll just be harder for to remain the most important category. When I joined in that was shortly after, by far, the most important category of investing for Bessemer, had been telecommunications equipment. I'd say more than half the firm's returns from 1996 to 1999 came from telecommunications equipment. I can't remember the last time we made a telecommunications equipment investment. And so to your point, Patrick, the most important thing to do is get off

Starting point is 00:57:22 the train before it runs into a brick wall at the end, and we think about that a lot. I agree with Kent. I don't think we're at the last stop yet, but we can sense it coming much more now. than we could even just a few years ago. Is your sense that the markers of that might still be software, just something totally different? LLMs usher in some whole new type of software. It's hard to even imagine what that might mean. But is that the sort of thing that you're watching for?

Starting point is 00:57:47 A change in platform or enabling technology that might just render the past generation of style of company less relevant or just create a new green field that's way more exciting than the telecom of 1996, 1999? It's hard to imagine the thing that's not software that is innovation. I think it's going to be another wave of software, but it may look radically different than it does today. And I think what we will look at is we've got close to 200 portfolio companies who are reporting back every day, what they're seeing in the field, what competition they're seeing out there. So that gives us a real edge when it comes to looking at a next wave of type of companies. I think more likely what we'll see is this, just as we saw with huge innovations in the last,

Starting point is 00:58:32 years in the cloud software, this software will just continue to evolve itself. People building these companies are paying very close attention to the language model technology and what that can do for them. And in many cases, it doesn't disrupt them. It just feeds right into their products. My bet is it's not going to completely flip their businesses upside down if they're awake to it. Yeah, Patrick, I come back to your question. Where are we in the evolution of vertical software? I kind of segment the opportunity into three buckets. One is the Greenfield play that we talked about. And I think you're right, I think there are fewer greenfield opportunities today than there were five, 10 years ago. What's different today is, as Ken points out, you can go after a smaller greenfield

Starting point is 00:59:12 market opportunity and build a much bigger business because you can capture more take rate than you've been able to in the past by layering on multiple layers of software and then providing financial services and value add services on top of that software. So I think the greenfield opportunity is still there, less obvious, but someone with domain expertise can find it. then there's a replacement cycle opportunity. And the way I think about that is, if you look across all the private equity firms, all the aggregators like Constellation, all the publicly traded legacy software businesses, there's tens of billions of dollars of ARR that's locked up in incumbent vertical software

Starting point is 00:59:51 that is ripe for the picking. Now, you've got to be 10x better, 10x cheaper. You have to take advantage of platform shifts, but it's there for the taking. And so that's the second bucket of opportunity. And then the third is what you alluded to, this new type of model that likely leverages AI. And the way I think about that is within these verticals, there are software businesses, and then there's billions of dollars being spent on services.

Starting point is 01:00:17 And a lot of that services dollar is going to be digitized and automated using LLM technology. As an example, in the legal vertical, there's a company called Even Up. and even up is taking the process of generating demand letters, which are these 100-page book reports that lawyers sent to insurance companies to help their clients get paid. That historically has been done by human beings with weeks of labor compiling these book reports. You can now use LLMs to generate it in a heartbeat. Those types of services offerings will be digitized and become software. It's crazy to think about all the shifts that we might see in the next five years because of this new

Starting point is 01:00:58 layer of technology. When companies come to you, let's say you're on the board of some software company, when they come to you, they're not naturally an AI company, they're just a software company. And they ask you, what are you seeing? Like, how should I think about this? How should I prepare my company to be ready for this new technology? What are you telling them? I have to plug, Patrick, last week we put out our annual state of the cloud report. And we spent 65% of our time answering the question you just asked. So we talked a lot about the current land. language model moment and practical tips for application software leaders. We did that not just specimen whiteboarding and coming up with our idea because the three of us couldn't operate a New

Starting point is 01:01:40 York City deli, but we talked to a couple dozen of our software leaders who are doing things about this already. So we got some practical tips. And the general answer is you don't need to wait until you have a language model PhD on staff to do something immediately. There are tons of companies who have pulled aside like a 20-year-old hacker and said, hey, why don't you spend this weekend playing around with this stuff and see what you can do? And they've come up with real applications. And we've got dozens of examples of product features that are already rolling out. The advice to companies that haven't yet started is to find that willing hacker advocate for this technology and empower them and say, look, you're the green team captain of the language

Starting point is 01:02:19 model tech. Go figure out what we can do here. A second version of that is to host an internal hackathon, $100 bonus for anyone who can come up with a cool project idea in this hackathon that leverages this language model technology, both for our internal efficiency, all the things we could do to be twice as productive with the people we have, as well as for product roadmap. I guarantee you any application software leader who does this will come up with 10 actionable, great ideas like this weekend. The only thing that's a little unsettling is this technology is changing every single week. The best in class idea you come up with this week, A month from now, you could come up with a set of better ideas because just the speed of

Starting point is 01:02:59 enablement of that platform is changing a lot. So it's probably a hackathon you need to start having every month for at least the next several months as this technology continues to stabilize. Anything about this scare you about LLMs? Yeah, absolutely. So leaving aside the existential questions. I haven't seen anyone point to a direct existential path to risk. But obviously, we're all paying attention to that conversation. But I think to bring it to where our pay grade level is applicable, yeah, this is going to shift the ground for all of these categories of software and for all these businesses. Any of them who are asleep at the wheel, I think, should be very scared. And us as directors of those

Starting point is 01:03:41 companies should be very nervous that somebody could come in with just a much better, more efficient mousetrap if we're not careful. So I think there's like a healthy, actionable paranoia for these companies just as software leaders that they need to be on top of this. Jeremy, does this feel to you like other platform shifts that you've seen since being a Bessemer starting back in 2001? That was right in the heart of the big one in the internet. How does this feel relative to cloud and internet and mobile and some of these other big shifts? It certainly feels like it could, but you mentioned cloud, internet, mobile, and with the benefit of a little hindsight, it's quite clear how big and important.

Starting point is 01:04:21 those were. But when they were first emerging, I'm not sure they felt all that different than VR and blockchain. And so these things all feel incredibly exciting at the beginning. And some of them make it and some of them don't. And oftentimes, even the ones that don't, sometimes there's a company or two that kind of make hay out of it. So we've got to look for. And so, yes, it certainly feels like the potential is there. But I think it's way too soon to declare victory. And I guess 10 years in, we'll know for sure. And five years in, we'll have some confidence. And now that we're maybe nine months in to the true arrival of these LLMs on the scene, I think it's very fuzzy. But we're certainly enthusiastic. We think there's much more likely than not to be a huge impact.

Starting point is 01:05:05 Whatever Jeremy's enthusiasm is I'm taking the strong over, just in classic Bessemer spirit, we would argue about this internally, openly. But the speed at which we've seen our portfolio companies build things with this is way different than what we saw in terms of mobile technology adoption. Mobile came out when the iPhone came out. We saw new companies show up within the year, certainly, but it took a while for our portfolio to see what they could do with it and have it change. I'm trying to think of the first portfolio company it impacted in a huge way versus a new company, whereas, again, I can point to a couple dozen of our companies who have features in development right now that they can describe to you that are obviously jogger.

Starting point is 01:05:46 dropping, wow, that would be incredible if you could hold out. A very simple one is the potential for any SaaS company that has an operating system to empower their customers to use natural language to query their database. Because most customers can't write SQL queries, but LLMs can. And so with some guardrails on this, you can turn any SaaS app into having its own natural language-driven reporting. That's wild. It's a nerdy, actionable example. There's a company in my portfolio that built that last week. It's a good example, and it's a measure how quickly existing companies adapt or adopt these new technologies.

Starting point is 01:06:25 And I was thinking back to Pinterest. The iPhone was 2007. We invested in Pinterest in 2011. Pinterest didn't have an iPhone app until 2012, a full five years after the iPhone. And so on that measure, Kent is absolutely right. But I wouldn't overweight that one. It brings to mind the question of defensibility. I'm always curious about how investors, like you all,

Starting point is 01:06:46 about this concept, especially because it's kind of hard to squint and see it early on in the business's life cycle. But how do you think about defensibility in light of a technology change like this? Because I personally, just one guy, have seen, I don't know, 20 companies that effectively say they're going to be this service you just described, Kent, like a BI analyst, almost like in a box, in a chat interface or something like that. And I'm sure an amazing one or 10 will get created, but it makes you wonder, like, how the hell are this business going to make money over time if there's so many options. So how do you think about what something like this does to existing companies' defensibility and new companies, how they might build that over time?

Starting point is 01:07:24 You could have asked the same question about cloud technology. I think this is largely going to be a completely ubiquitous commoditized offering that will be equal opportunity for all these software companies. I don't think with maybe a few exceptions, you'll hear people discuss if you have a proprietary set of data. Maybe you could build some custom language model that works only on your proprietary set of data where you don't let it out. I've heard very smart people say like, no, as these models get bigger and bigger, like they don't really need access to the proprietary data to be really, really effective. So I think it's like everything, it's just another golden hammer in the tool belt of incredible software. And you just need to use it to build the best possible product

Starting point is 01:08:07 and own your market that way with the best possible product. And the long term defensibility comes there and occasionally we'll have a company that will be lucky enough to have a true network effect. I don't think this AI is more likely than other technologies to lead to a networked effect. I do think when the companies that leverage this tech to make a jaw-dropping product roll their product out, it will be adopted even faster because it will be a holy crap moment. Like, oh my God, that product is incredible. And like the word of mouth that's a key feature in the spreading of a lot of great companies will be even more accelerated with this.

Starting point is 01:08:38 the ability to onboard customers leveraging this technology will be accelerated. The efficiency of these companies will be accelerated because they'll be a more productive per person. So as you can tell them, a little bit of the bull in this group on this tech. But I think the long-term defensibility is just classic software long-term defensibility, speed to market, own that market, build a great team that provides the best product and service in market. That's the real defensibility. This change brings to mind another question as we wind down here about building a great investment firm, which when it comes to the partnership, the partnership is,

Starting point is 01:09:08 the firm at any given point in time making these investment decisions. And I understand how you recruit and train through the apprentice model great investors over time. What's the other side of that? Like, how do you get people that just aren't working out as partners out of the organization, whether that's a young person that started and just isn't a fit or someone that's been there forever and is towards the end of their useful life as a partner at Bessemer for whatever reason. and needs to be a new young partner that is more specialized in AI or something like that. How do you think about, I guess I'll call it pruning and tending to the partnership, especially as it relates to having people leave when they're not right for it any longer or in the first place?

Starting point is 01:09:50 Well, like anything, practice makes perfect. In the 20-so years I've been here, I've probably seen 50 to 70 professionals leave Bessemer. and virtually all of them have gone on to do something really successful in many instances in the VC industry. So one of my favorite games to play is just talk about the people who are partners at other firms who spent an important year or two or three at Bessemer, Sarah at Benchmark, Christina and Chris at Injores-Norowitz, Philippe at Excel, Mitchell and Brian at Lead Edge, Larry at Volition, Lisa at Norwest, Anna at CRV, the list goes on and on and on. I think Bessemer to this day has exported more professionals to the rest of the venture capital industry than the rest of the industry combined. And so the good

Starting point is 01:10:44 news is you can find it's not the right fit for you at Bessemer and still be really successful in venture, let alone in other places. We have people who are writers and comedians and in politics and so forth. But the bad news is it's still hard and it's awkward. And so I think the important thing is what we realized, and it's sort of obvious when you say it, in order to attract the best people, they need to feel confident and comfortable going to a firm. And when the firm can show that you have a really good chance of succeeding at that firm over a long period of time, or succeeding outside that firm, it's a no-risk proposition. The scary thing is, when you join a firm, it's either you're successful here or we're going to publicly humiliate you if you're not successful

Starting point is 01:11:26 here and good luck with the rest of your career. That's a much tougher proposition. We want to get the best people, even those who might be a little bit nervous about what the future holds. So we're really careful to do that pruning or the trimming as you describe it in an extremely discreet way. So nobody really knows other than the person. And by the way, it's often not about keeping someone's weaknesses a secret. It's often about just it's the wrong fit. Like we have a particular model. The model works really well for us. There are many people who want to do it in a different way. The most sharp contrast I might offer is the hold hands, say kumbaya together and make an investment together. That's not what we do. We support each other, but as we mentioned

Starting point is 01:12:04 earlier in this conversation, we do it in a way where we're individually empowered to make decisions. That's not for everybody. And so in many cases, the folks who don't stay at best and for the long haul are really talented. They just are going to thrive better in a different system. Jeremy spoke to the up-and-comers. You asked about the veteran investors. One of the healthiest aspects of our partnership is there's no tenure. Every year we evaluate each other's performance. It doesn't really matter what you did 10 years ago. If you're not performing today, we have a very graceful and elegant model for transitioning folks out and making room for the next generation. I go back to the intellectual honesty where we're having open conversations with each other all

Starting point is 01:12:46 the time. So it's not like there's a surprise moment. It's like, hey, Jeremy, you lost a billion $10 last year and you haven't performed in 10 years. That's not true for Jeremy. It's a constant open conversation, so it's not political, it's not personal. It's as low friction as it can be. I've loved learning about Bessemer from U3. I think it's such a unique firm. I love collecting these different models. There's lots of ways to do this, but really like the interesting advantages that your model has conferred on the partnership and obviously the results have spoken for themselves. My traditional closing question for everybody is to ask what the kindest thing that anyone's ever done for them is. Kent, maybe we'll start with you.

Starting point is 01:13:22 Oh, boy. I mean, so many people. I go back to Felda Hardiman, professionally, my first mentor at Bessemer, who did sit next to me in a room for hundreds of hours to no benefit for himself of just telling me, like, here's why I did that thing that way. Half the time I just sat there, we're like, why are you wasting your time telling me these things? He was just a natural teacher and professor and he certainly did not have to do it. He could have been having a lot more hobbies in free time if he hadn't bothered educating me. So that was professionally the kindest thing that anyone did for me. Brian, you want to go next? Yeah, I was born in the Soviet Union. And when I was

Starting point is 01:14:02 four years old, I got on a plane with a box of Legos, came to America. And my parents uprooted their lives and came to a country where they didn't speak a language and had no money and built a better life for me. So I'm grateful for that. Amazing. Jeremy? that's a hard one to compete with. Yeah. Sorry. You can't be successful in this industry or a little in any industry with an enormous amount of luck.

Starting point is 01:14:28 And luck typically comes at the extension of help from somebody else. And I don't think any of us would be here without it. I'll pick on one specific professional situation. And it was actually Sarah Tapp, who's now a benchmark, who was working at Bessemer at the time. She was the first person who I ever managed in my life, in my career. we were a couple of years into working together. And she wrote a public blog post about me, which she didn't run by me first,

Starting point is 01:14:53 where she mostly just said, thank you. And it made me feel so good that I basically resolved at that point to say, like, I want to do this 10 more times in my lifetime, find someone who's younger than I am, less experienced than I am, and help them. Had she not done that?

Starting point is 01:15:06 I don't know if I would have taken to it in quite the same way, but I'm really grateful for that. Wonderful story. Love that point of connection. And I'm very thankful to E to you for your time and for being willing to share a lot of how the business runs and what you've learned in the past. Thank you. It was a pleasure. Thanks, Patrick.

Starting point is 01:15:23 If you enjoyed this episode, check out join colossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.

Invest Like the Best with Patrick O'Shaughnessy - Bessemer Venture Partners - Building a VC Firm that Lasts Centuries - [Invest Like the Best, EP.327]

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.