How I Invest with David Weisburd
00:43:20 · David Weisburd
[00:43:20] S1: I want to also double click on something that you said that investors can't afford not to have venture in their portfolio. I'm summarizing. And there's two aspects to that. One is you could argue from the return standpoint over the last fifty years, venture capital has been the best returning asset class in the world outside of sports teams. I've been corrected on that. But if you if you can't afford, the Pittsburgh Steelers or the New York Giants, venture capital is the best asset class to own. On top of that, just 87% of $100,000,000 plus revenue companies are private. So there's actually another aspect to that, which is diversification. You can't just access technology in the public markets. You're just gonna have, like, a small portion of the portfolio. So in order to have a diversified portfolio, in order to to kind of have the the optimal risk and reward, there's an argument that you should be in venture just for the sake of diversification, let alone returns. Earlier on, you were talking about these three things that you look for in teams. I realized I cut you off. You said chips on shoulder and you also said team and team cohesion. What else is important when you look to invest in a venture capital manager?
00:44:27 · Brijesh Jeevarathnam
[00:44:27] S0: Three things, the team, which we talked about in terms of not just having the table stakes of being super smart, connected, etcetera, but also having team cohesion, complementary skills and a chip on the shoulder. Talked about... That's one. We talked about the track record, which is backward looking by definition. The third bucket we haven't talked about is investment strategy. By that, I mean, what is the investment thesis in this market? What's their worldview? What is their stage that they're pursuing to invest in? What's the portfolio construction? Do they have the right network connectivity and the now overused, I guess, phrase right to win in that particular part of the market, whether it's early stage seed investing or growth, writing bigger checks, etcetera. As you can imagine, as I talk about these three buckets, it's art and science, right? You can't just put this in an AI model and say, Hey, should I invest in this fund or not? There is a very heavy people element to this assessment. And we were talking before about the feedback loops. The LP job, the feedback loops are very long and we have to work extra hard to make sure we capture the feedback and then make changes to course correct. So we, for example, at Adams Street, we have three to four times a year, I'll call it kind of offsites or mini offsites, two main ones, one in January, one in June, with the entire team, the senior team, to talk about a number of things, including what we call our anti portfolio. What we miss three years ago, five years ago, and what was our underwriting then And how has it changed? This business is very humbling. David, when you invest, you make both kinds of errors, the mission and commission. And we don't need to be in every great fund, but every fund that we invest in, we want it to be great. And we make both those errors. So, anti portfolios things we missed. And then we make the other kind of error as well, which is, I think kind of four or five years in, we got this one wrong. It's in our portfolio. What did we miss? Or what has changed since then? And that kind of forcing function of looking at it two to three times a year helps hone our thinking. Human And nature is we forget this recency bias.
00:47:05 · Brijesh Jeevarathnam, David Weisburd
But by going back three, five, seven years and looking at what we did right and wrong, we've, I think, found a way to institutionalize and build upon this very long feedback cycle into something we keep improving every year. [00:47:22] S1: It's not an easy job to be an LP because you're constantly thinking about all these different frames, all these memes that you have to remember, all these mistakes that smart people make. And then you need to, the next day, go and break these same rules that have been hammered into you for decades, especially today. So many things are changing and you have to figure out what's changing, what's the same, what's... Doesn't even fit into this changing versus non changing framework. You have to truly think from first principles.
00:47:47 · Brijesh Jeevarathnam
[00:47:47] S0: It is challenging. Let me share some stats around that as well, because our job as allocators, both in venture and buyout, has become more difficult simply because there's way more firms and therefore getting to the signal and parsing out all the noise, it's just much harder. I think I said this before, there are something like 3,000 to 4,000 venture firms in The US alone today. And that compares to something like say the 400 to 500 in say fifteen, twenty years ago. And when I first started in the business, it was kind of clear of those, let's say 500 venture firms in The US, which maybe 10 to 20 were the ones that really were exceptional. And it was less about picking, it was more about getting access when I first started. Today with 4,000 plus venture firms that are active, there are years that are cycles where more new firms are created, venture firms, than ever existed when I first started in this business. So, we're meeting these new firms and all these names and these great people, etcetera. And this is especially hard in an outlier business. As I'm sure you've seen the stats, we're fortunate to have forty plus years of data. We've been investing in venture capital funds and companies since the 1970s. And we use the data as an offensive weapon. It's obviously proprietary. And our data tells us that 7% of venture backed companies generate all the gains and the rest, 93, is just a wash. It's just noise. Of course, ex ante before the investment, no one really knows which 7% I'm going to make it. If I go up one layer from that to the managers, right? Just like there can be an outlier effect at the company level, the funds that back these companies, some can get lucky or some can really find, to use the prior phrase we talked about, diamonds in the rough, etcetera. And it's really hard to predict in a linear way attributes of A, B, C, and D make for a great venture manager. We got to think really kind of outside the box and change our playbook, etcetera. So we invest in about, let's say 20 to 30 venture managers in The US. And we're picking, as I said, from a cast of about 4,000. Previously it was clear what those top maybe 30, 40 or 2,030 would be. Now it's not so clear. And it makes our job extra hard. And I think this is where the job of an allocator has become much more challenging because we have to move fast.
00:50:45 · Brijesh Jeevarathnam, David Weisburd
These funds, they come back to market, they raise capital quickly. We have to really broaden our coverage. We're looking at many more funds and sectors and so on than before. In that sense, to a certain extent, it's become like direct investing because you're investing in a much more noisy space with arguably even less data than before. And that's the challenge. And that's where we kind of leverage our, not just the data, but the investment judgment of the people we have. The cliche about it's a people business is true in investing. And we're fortunate to have a team that is now approaching 40 people strong that spends all their time selecting managers. And some of the folks on the team have been doing this for ten, fifteen, twenty years like myself. And it takes that long to kind of get the pattern recognition engine working. It's still not a static engine. It's got to be dynamic as we keep adapting to the market. [00:51:56] S1: I'm doing the math, 4,000 managers, 40 people, still a lot of managers for every professional to handle. How are you leveraging AI to help?
00:52:05 · Brijesh Jeevarathnam
[00:52:05] S0: I'll start with the premise, the true premise that venture investing is a people business. And some of that lends itself to using AI, a lot of it does not. So let me start with the top of the funnel. Part of our job is to know attribution, not just at the venture firm level, but at the individual partner level. And that we can use AI for. Let's call it individual track record. But to know which person with a great track record is going to leave their franchise in potentially three months, six months, who are they thinking about pairing up with? And we have a few live examples of this as we speak. I don't think AI can do this anytime soon. And this comes from us being in the ecosystem, talking to people and being in the Bay Area, especially, you'll appreciate this. You're fully surrounded by other folks. Go to my... Whether it's my kid's high school or a soccer game or a social event, you see somebody, you pick up some new information. That is very much a human process. And we got to go to where the puck is going to be because often when these great people start a firm, the best ones is an invite only list. And you got to get there before they really are kind of in official fundraising process. That's an example of where AI is not helpful. It's very much a human process. But where AI can be helpful is, I mentioned the data we have in house and we're always ingesting more data. This data is sitting in very, I would say disparate parts, unstructured data sometimes. We're using AI to help us connect those dots. Something could be in a meeting note somebody took versus a quarterly report versus an advisory board deck. Everything is confidential, but we got to learn about this particular GP or the particular company. We have to have at our fingertips the collective information base that Adams Street has within its four walls. And AI is really helping us do that. In addition to the more, I'll call it process elements of due diligence, writing the memo, for example, or doing some of the quantitative data cuts, for example. And we're still scratching the surface here. We can do a lot more with AI. But I would posit that on both the funder picking side and the early stage founder picking side, it remains very much a human centric business for now.