Latest / The Longest View with Dez Fleming / How Understory is Revolutionizing Private Market Data for Investors

Transcript

How Understory is Revolutionizing Private Market Data for Investors · The Longest View with Dez Fleming

• [Desmond Fleming] (0:00 - 0:19) Maybe we can kick things off. Eric, David, really excited to have you both on episode two of The Longest View early on, early believers. I think actually to start, would love for you, the two of you, to briefly introduce yourselves and briefly introduce Understory for people. [Eric Shapiro] (0:19 - 0:35) Sure. So I'm Eric. Prior to starting Understory, I spent 10 years in the finance and investing ecosystem, first as an investment banker, and then seven years in a hedge fund role where I was covering distressed credit, public equities, and private equity. [David Fine] (0:36 - 2:16) I'm David. I, prior to starting Understory, I also spent 10 years in an industry, but it was working at a couple of different startups, mostly on complex data problems and in customer success and product. And what Understory is, should we just jump in? Jump into it. Yeah. Great. So at kind of the top level, the tagline of what Understory does is it aggregates financial data from private companies and private sources for financial analysts. That's a very, I find that to be a very vague tagline to a certain extent because- Which financial analysts? Aggregation can a lot of things. Financial analysts can mean a lot of things, but I think at its core level of where it is today, if you are an analyst at a VC or a private credit fund or a hedge fund and you receive a data room and it's just PDF financials and you, in the past, that would have ruined your weekend probably because you're thinking, I got to stitch a bunch of data together across a time series to generate the basic data to create a model. What we do today is you just drop that into Understory. We extract all the tabular data, do the complex process of stitching it all together and then generate an Excel sheet that's well-organized for you to start building your model off of in 5 to 10 minutes. So that's the core of what we do today. We have ambitions to make that a bit bigger, but that's kind of the core product today. [Desmond Fleming] (2:16 - 2:35) We'll come back to that vision, come back to where you guys are today. I actually want to come back to the very beginning and not of just of Understory, of how you two met. So we're talking about it offline, I guess, of how you guys met in philosophy class. Can you tell that story? Sure. I think Eric tells it better, so I'll let him. [David Fine] (2:35 - 2:38) He probably also remembers it better, apparently. [Eric Shapiro] (2:39 - 3:25) It is kind of that moment where I can imagine how, you know, the cinematography would work if it hadn't been filmed in a movie, because it was a very special moment. So David and I both went to Columbia, and kind of the hallmark academic feature of Columbia's education philosophy is their core curriculum. So about a third of your classes there are predetermined. One of the classes that everyone has to take is called Contemporary Civilization. So David and I both were in the Contemporary Civilization section with a professor named Peter Pasolini, who we're still in touch with. He was kind of like a Robin Williams from Dead Poets Society meets the Dalai Lama kind of guy. And that was just like a great backdrop to form a friendship that was both intellectual in nature, but also just a lot of fun. [Desmond Fleming] (3:25 - 3:33) Was that professor there like 20, 30 years by the time you guys got there, 10 years in? No, he's pretty old. Yeah, he's older. [Eric Shapiro] (3:34 - 4:00) Yeah, he actually only taught this class, I think, because he felt he had had a prior academic career. And this class was like a moment where a lot of people discover ideas that they hold on to for the rest of their lives. And so he had that kind of effect in, you know, encouraging us to think deeply about the world and ourselves and our values through the context of basically like the Western canon of political and moral philosophy. Okay, cool. [Desmond Fleming] (4:01 - 4:10) And so how does that lead to, you guys meet in the class and then what happens? Do you guys immediately start talking about business or is it talking about, you know, democracy? Like how did it work? [Eric Shapiro] (4:10 - 4:28) We actually, so we did do a bunch of projects together, despite having different majors in college. So we ran or managed a literary publication together in college. We also organized a protest together. Oh, okay. And in both those experiences, yeah, we won't talk. [Desmond Fleming] (4:29 - 4:30) You don't, okay, sorry. [Eric Shapiro] (4:30 - 4:35) We don't need to get into it. We don't need to get into the protest. Columbia protest. We did not get arrested. Yeah, sorry. [Desmond Fleming] (4:36 - 4:38) This was in 2013. [Eric Shapiro] (4:39 - 5:14) But I think that gave us a high degree of confidence that we would enjoy working together. Yeah. But at the same time, our careers went very different paths. I worked for a bulge rack investment bank and David started a company with his brother in the sheets. And so I think we thought that would be the end of the line of our, you know, working relationship. Although we continue to live together after college. Yeah. But yeah, I mean, you can tell it, but I think we both reached a point in our career where we were starting to talk about, you know, things that we had seen in the world and how we thought we could make them better. Yeah. And I'll turn it over to you. Yeah. [David Fine] (5:14 - 7:23) I started this journey in a much more linear sense, in the sense that I had always thought of starting a company. I always felt like I was entrepreneurial. And where did that come from? Was that influenced by parents, by media? I think the major influence, which is kind of funny because I think it was actually a big professional turnoff for my mom. But her father, my grandfather, he immigrated from Germany during the interwar period between World War I and World War II. German Jewish guy fought in World War II and then made his way back to Minneapolis after the war with his brother. And they started, I guess, what in the Jewish world we call a schmata company, which is basically just like clothing manufacturing. And they turned that into a Western wear brand and company that still exists today. And so that story, yeah, it's pretty cool. They make like snap shirts and jeans with rhinestones on them and things like that. And that story always served as an inspiration, I think, to myself and my brothers. And so I think that that was part of it. And also in high school and beforehand, I think I was a bit of a nerd. I was very into science fiction, into technology. I read TechCrunch when it was really popular, just religiously, Wired Magazine religiously. And so I was very enamored with the concept of a tech startup. And so I think that was the origin of me being interested in it. At Columbia, though, I studied history. So during the summers, I would work at different startups as an intern or try to work my own ideas. And so when I graduated, it was very, it was the logical next step. I think most history majors at Columbia either become lawyers or consultants, basically. And for me, it was always going to be, I wanted to go into startups. [Desmond Fleming] (7:23 - 7:30) And so tech was on your radar. Yeah. How did you, the two of you, come to want to start Understory? [David Fine] (7:30 - 8:13) Eric and I, as good friends, by this time, were both married. We would grab beers together and just commiserate about work. And I was working at a company at the time that was dealing with a lot of different complex data problems and analyzing data and really, I think, kind of state-of-the-art manners. And at the same time, I was getting beer with Eric, where he was talking about the issues and inefficiencies he was facing at his job at Elliott, which is a top-tier hedge fund. And it occurred to me that these inefficiencies should not exist. They should be addressed in some manner. And so that's how we started tinkering on the problem together. [Desmond Fleming] (8:13 - 8:24) And if I recall, you were at Dynamic Yield, correct? And Dynamic Yield primarily served Fortune 500s or was a single client or a subsidiary of a big enterprise? [David Fine] (8:25 - 9:58) Yeah, it's actually an interesting story. For a few years, I was a McDonald's employee. So Dynamic Yield primarily served small to large e-commerce companies as a personalization and A-B testing platform for commerce online. And I was there for about six years. And halfway through that tenure, we were acquired by McDonald's to implement real-time A-B testing and product recommendations at drive-through menu boards and kiosks across McDonald's globally. And I was part of that project after the acquisition. And so we were just dealing with a ton of data on a regular basis. How do you take the billions and billions of data points you have around the way people buy things at McDonald's and try to convince them to buy more things or different things? And that was what I was immersed in on a very regular basis. Had an entire world-class data science and data engineering team behind our back to answer these questions, to build analyses, and then also to build the algorithms that actually try to accomplish this. And so I got a taste of what the state of the art looked like when you were able to bring to bear all of the tools that you could around a data problem. And I felt that what Eric was facing was kind of a lack of that tooling that was really custom built and purpose built for the problems that he was encountering. And so I think that's how we started talking about it. [Desmond Fleming] (9:58 - 10:59) Yeah. So let's transition a little bit to talking about early days with Understory, as well as the tooling for investors. When I think about kind of the venture tech stack, what do people use? They use email. They live in it. A lot of people use Google Docs for their notes or some form of note-taking, probably also on their phone. And then people will religiously and intermittently use things like PitchBook for data, CapIQ for data, AlphaSites also for data, but related to specific diligence processes. It still feels like we're in a very manual world. There's not a ton of automation. You guys are trying to solve component parts of that and have started with capturing unstructured data and then putting into structured outputs. But we'd love to get the two of your perspectives on where the software tooling is going for investors, as well as, hey, what's it been like building Understory in the early days? [Eric Shapiro] (11:00 - 12:58) You describe it well. There are a lot of vendors that historically have provided workflow automation tools wrapped around proprietary data sets. And that's kind of how I think about what Bloomberg or CapIQ or those products are. Increasingly, there's a cohort of companies today that, especially with AI, are taking that one step further and delivering workflow solutions for individual investment professionals. So, you know, summarizing data for you from a variety of sources or reformatting data, repackaging it somehow, assisting in the analytical process. I think where we see ourselves, while today we do sell a workflow automation tool that solves a particular problem for analysts, we see ourselves in a slightly different area, which is ultimately providing proprietary data to investment firms based on information that they already have access to. So a firm like yours, for example, you probably see hundreds of companies. You invest in a fraction of them. You have a really good sense of what good and great look like. You probably, to some extent, can go back and pull up old investment memos and see trends for a particular business in a very niche subsegment. You're probably not capturing all of the data, though, for the deals you pass on, for example. And at the early stage, though, those data points are not particularly quantitative. But if you're investing in mature businesses, you're getting a rather robust feed of information on how those companies perform, when they fail, you know, on a granular level, what things like working capital or, you know, margins look like at different points in the P&L. We want to unlock that for investment firms who are sitting on this information in PDF files, email attachments, board decks for past portfolio companies, and really enable them to get a quantitative view of that back book of intellectual property that they're sitting on. [Desmond Fleming] (12:58 - 13:37) I mean, you guys are attacking this, but what do you think is the gap? You know, if you walked into any set of GPs or people who sit on management committees of funds, they would say, yes, Eric, of course, we have all this data. We don't know what we're doing with it. We would love to know, you know, over the past three years, what have been the quantitative metrics for the top 10% of the performance of the fund? Like, people would love to have that answer. So there's a desire, but like, what have you guys observed about the gap from getting from the reality today to that future state that you guys are describing? [Eric Shapiro] (13:37 - 14:52) There's kind of two places. The first is just getting clean data in a single location. So being able to take all those documents and synthesize them, but then put it in a database on which you can run analytics and things like that. I think that's a huge problem, and that's why we've started at the point of trying to structure data from those source filings. I think the next point is, you know, investment folks love Excel, and there are good reasons for that, particularly when you're looking to do portfolio-wide analytics. Building a dashboard for a one-off piece of analysis, you know, in a web platform is really not the way to do it. Ultimately, what you want is just to tell some analysts, like, you know, color code this, move that column over, et cetera. And Excel is really good at that. So I think once, even once you have that structured data, it's putting it into a format that can be manipulated and digested by people who are used to looking at spreadsheets and PowerPoint slides. And that's kind of the direction of travel for us. But I think those are the two places where just because you have a team of data scientists or just because you have, you know, an internal dashboard that you have committed to using for whatever reason, getting the data into that format is where I think there are a lot of challenges remaining to be solved. [Desmond Fleming] (14:53 - 14:56) What's been something unexpected about building the company? [David Fine] (14:57 - 15:21) Almost everything. Just because you're selling into a very high margin, kind of high capital business, like funds, doesn't mean that it's going to be an easy sell. In many ways, these funds operate almost like small to medium-sized businesses more than they do, like, an enterprise that usually buys technology. [Desmond Fleming] (15:21 - 15:21) It's seething. [David Fine] (15:21 - 15:24) They're like, yeah, we've managed 40 billion of capital, but it's a hundred people. [Desmond Fleming] (15:25 - 15:25) Exactly. [David Fine] (15:25 - 16:02) Exactly. And there's like two people who are dealing with technology and they're underwater. And so there's not the same sort of rails around a sales process that you might have when you're doing a traditional B2B sale to, you know, an e-commerce company or something of that sort. So that was pretty surprising to me. I had always expected something different from selling into funds than what we encountered, but it's been good because it's made us kind of hone the product in a way that has made it more useful to our end users. [Desmond Fleming] (16:02 - 16:25) So there's one element of accelerating go-to-market, which is just improve the product and make it undeniable. You know, you put it in front of someone, put the demo and they're like, oh my God, I need this. Right. Excluding product, have there been tactics from a go-to-market standpoint that you know now that when we rewind the clock, I think 24 months plus ago when you guys were starting, that you wish you knew? [Eric Shapiro] (16:26 - 17:39) We started the company with this philosophy of let's build a product that works in single player mode because that way we don't need to necessarily run an enterprise motion in order to get adoption. I think we've been, certain assumptions underpinning that choice have been, you know, humbling to learn more about. Ultimately, you know, I worked in these environments for a long time. You think things like efficiency, speed to deploying capital, greater visibility into your data are what folks care about. Ultimately, I think what we've learned is that you really do need to solve an institutional need rather than an individual need, even if the buyer is only evaluating one seat. And so, you know, I don't think that changes anything about our roadmap necessarily, but it does impact things like positioning. It impacts who you talk to, you know, because each one of these firms is their own beast and the decision maker can be different in each one. So I think being more humble about like, what are your organization's priorities rather than, hey, this is going to make you a better analyst. Yeah. Don't you think you should buy it? Yeah. [Desmond Fleming] (17:39 - 17:47) And that could be true, right? It could make someone a better analyst and also not fill the institutional need at the same time. [Eric Shapiro] (17:47 - 17:49) Right. That's probably been the biggest learning. [Desmond Fleming] (17:49 - 18:36) Yeah. How do you guys go about hearing or listening for institutional needs? Again, going back to, it could be, you know, managing directors, co-founders. A lot of these funds are still ran by, you know, founder CEO types. They're hard to get in front of. They're busy. Their priorities may shift from, you know, saving a portfolio company to preparing for fundraising. Not all of them may be thinking in a world where, hey, we need to make the 20 analysts that are super highly paid, you know, 20% more productive. So what do you guys, without necessarily naming names, what have you heard are like signals of, hey, this is the institutional need that we can solve for, or we can work backward from, here's a need I'm hearing and we can solve this problem for them. [Eric Shapiro] (18:36 - 20:10) I think on one level, you have people who get hired and maybe there's a new sleeve of capital being raised around a new team or a new strategy. That's obviously a signal where going to them and saying, hey, you know, having the infrastructure to go deploy this capital or be able to talk about it with your investors that you have this new piece of infrastructure. That's definitely one signal. But in a lot of cases, it comes down to just like, is there a person who's empowered to be thinking about these issues? David kind of alluded to this, but even at a large investment firm, the human capital side of the business is very different from the way that they make money, which I think is kind of unique to how investment firms operate. You know, if you're really good at investing, that doesn't mean that you're the right person to be thinking about organizational needs or institutional scale. We have found that while we can serve, you know, the needs of an individual team, some of our customers do have people who are really empowered to be thinking at the platform level. And that's usually where we can get the most visibility because they're unconstrained by, hey, sorry, I have an earnings call today. You know, we can't get together or whatever. But that is their priority. I would say, though, that that is, you know, while a trend going in our favor, you know, I think in a lot of ways, data strategy, IT strategy, you know, those things are historically not inherently part of an investment firm's operating model in the way that investor relations or, you know, Bloomberg Access are. Similar question. [Desmond Fleming] (20:10 - 20:21) Without naming names, how did you get your first customer? Through our network, you know, I called my friends, which is how a lot of people get their first customer. [Eric Shapiro] (20:21 - 20:42) But what I really wanted to say is another thing that's been very interesting is I'd say a good chunk of our customers come from people we knew and were able to kind of not call in a favor because no one's buying anything from you if they don't need it. But people where you can at least get the mind share in order to get a foot in the door and make your pitch. [Desmond Fleming] (20:42 - 20:52) You have, you've trust built up with them in some other domain or realm. So spending 30 or 45 minutes to hear you out, they will happily do that. [David Fine] (20:52 - 20:52) Yeah. [Desmond Fleming] (20:52 - 20:56) But they're not just going to say, yeah, I'm going to cut you a hundred thousand dollars because you're my lit. Totally. [David Fine] (20:56 - 21:28) I've also found, by the way, just sorry for cutting you off, but it is to sing your praises, is that Eric is not a typical sales person in this in this space. And because of his investing acumen and his experience of where he has invested at previously, people kind of jump at the opportunity to talk to him, I think. And they're just interested. They're like, wait a second, you you left a large hedge fund job to go start a company. What's going on here? [Eric Shapiro] (21:28 - 21:29) Can I do that? [David Fine] (21:29 - 21:40) Or should I do that? Or maybe I shouldn't. But at least I want to talk to you and understand what it is you're doing. So that's been that's been a unique in as well. Yeah. The question probably for a lot of people is why? [Desmond Fleming] (21:40 - 21:55) Because lots of people want to get to to where you were. And so to be at that peak, let's call it, and then to start over from the ground floor of building your own company, like why make that trade? You don't have to answer that. [Eric Shapiro] (21:55 - 22:16) No, no. I mean, I'm a little bit of a glutton for punishment. But in all seriousness, I'm extremely passionate about investing philosophy. Yeah. You know, like the meta thinking around investing, you know, how to think about risk. I mean, that's what I think about in the shower, you know, or at night when I'm trying to fall asleep. [Desmond Fleming] (22:16 - 22:28) How do you think about the risk associated with understory? Because I think that's a entrepreneurs have a more unique appetite for risk than than most people. [David Fine] (22:29 - 22:49) It's a good question. I mean, both of us have families. I think both of us are very fortunate to have wives who are in high powered, well-paying jobs. So that has mitigated the personal risk a little bit, which I don't know. Now that I'm answering this, I'm not I'm not sure you want us to talk about the personal risk or the. [Desmond Fleming] (22:49 - 22:59) No, I think it's organizational. Right. That's a component of of being able to say, hey, yeah, I can go from making X amount per year to zero per year for some indefinite period of time. [David Fine] (22:59 - 24:05) I think actually Eric and I probably approach this very differently because the way I thought about it was I I was I was at Dynamic Yield, which was acquired by McDonald's, and then it was acquired by MasterCard. And I actually really loved my job and I love the people there. And I could I could just see myself if I stayed at MasterCard, just waking up in 10 years being like I should have left a while ago to go start a company or go change my career and somehow. And so and so it felt like the right moment to do so. And as as a result, I was able to underwrite the risk because it just it just felt like the marginal utility of my current trajectory was not as high as an alternative like starting a company. And so I'm just constantly re-underwriting that. And it still feels that way. Like I'm still learning so much. And I feel like building equity in something that's that's really valuable, either hopefully commercially valuable, but at least from an experience level valuable as well. [Eric Shapiro] (24:05 - 24:15) Yeah. For once in my life, I tried not to think about it too much. Yeah. I mean, the risks are so obvious, you know, like especially for me, I have no technology background. [Desmond Fleming] (24:15 - 24:15) Yeah. [Eric Shapiro] (24:15 - 24:42) So I'm just like, this is clearly dumb. Yeah. But, you know, I would spend we had like a notion document and I would spend my nights and weekends, you know, looking at tools. And this started from a whole different place. Right. But we're just talking about it. And it was so energizing. And then just the opportunity to work with David. Yeah, it was really special. So I don't know, it felt like the right time in life. And you just can't think about it. [Desmond Fleming] (24:42 - 25:49) Yeah, I think I think I think one component of being able to work with a trusted friend pays for itself in spades because it's a unique personal experience. Like, I'm sure this you guys already passed this, but let's say worked on it for three years and went nowhere. Yeah. Well, I got to work with my homie all day. So that was pretty cool. And then the other thing that I like to put out into the tech ecosystem environment, especially for founders and prospective founders is, hey, actually, the risk of starting a company. There's a high failure rate within your own company. But the risk to your career actually is not as high as people think, because within at least in the tech environment, people put a premium on people with past founder experience. And obviously, there's always this kind of Darwinian process of companies being created, companies dying, new companies being formed. So there's always a spot for talent, at least in the startup ecosystem. I don't know what it would be like to get back into dynamic yield or MasterCard or to go back to Elliott, but the startup ecosystem lives on that kind of recycling of talent. [David Fine] (25:49 - 25:59) No, that was definitely part of the calculus for me. And you articulated it much better than I could. And at dynamic yield, I don't know if they still do this, but they had most likely to do X. [Desmond Fleming] (26:00 - 26:02) Oh, yeah. Superlatives. Senior superlatives. Yeah. [David Fine] (26:02 - 26:04) So they had superlatives. [Desmond Fleming] (26:04 - 26:05) That's kind of fun for a company. [David Fine] (26:05 - 26:46) Yeah. It was a great, fun environment. They had superlatives every year, at least in the New York office. And the first year I was there, I won most likely to start a company. And I was quite honored. But then I kept on winning it. And by like the fifth or sixth time, I was like, I need to like start a company or take myself out of contention for this award. There's definitely, I think, a much more welcoming environment to leave your job and start a company than there is probably in Eric's industry, where I think some people probably looked at him like he had, he was having like some sort of mental breakdown. [Desmond Fleming] (26:46 - 27:12) Yeah, they just do the EV math and like this expected return is lower than this one. Sure, it could work, but probably. Let's talk a little bit more about Understory, the product. So David, for you, would love to hear, you know, how do you think without giving away the special sauce, how do you think about your tech stack? What have you experimented with over the past two years? Remind me, you guys got started at the end of 22? [David Fine] (27:12 - 29:48) Yeah, our third co-founder, Alex, who lives in Canada and previously worked at a company in the space called Canalys, he joined us January 23. So that's really when we started coding in earnest at all. And the way that we've built the tech stack basically is a deterministic data pipeline in the sense that you have data from sources being currently PDF files with data and those are kind of source A and they need to get to point B, which is a nice looking organized Excel that has a unified time series. And so we have an entire algorithm or distributed app to do that and that's the way we built it initially. And then what we realized is to really do that very effectively, we needed to start building in semantic understanding of source A to get to point B. And for that, we started integrating a significant amount of LLM inference points. So when we started building under story, you could generate a model and it would work relatively well, but there would be edge cases where it didn't work super well. And the only way for us to address those edge cases would be algorithmically. Today, what happens is it still goes through that deterministic pipeline, but there are a bunch of different waypoints and parsers and little tiny inference points where an LLM comes in and says, actually this is what this table is referring to or this is how this table refers to time. And so by the end of that process, a single model could call hundreds if not thousands of inference endpoints to generate what is a much better looking and more consolidated and much easier to use Excel sheet today. And so that's effectively how we built it to be robust enough to handle a lot of different ad hoc types of documents, right? Like every single PDF file is its own special snowflake, which is why this problem is so complex and difficult to solve. And the emergence of really robust LLM technology has allowed us to build a relatively resilient system to handle that problem. [Desmond Fleming] (29:48 - 30:21) And it feels like your system is constantly expanding. So if I, you know, drew out a hypothetical PDF today and had some, you know, I don't even know what to describe, but like an example of an edge case, you know, I could digitize that and then put it into your platform and that could be viewed as the nth permutation of a new document that, oh, we haven't seen this before, but now we have. So when it comes again, you know, this will be a type of document that we feel very comfortable with parsing. [David Fine] (30:21 - 30:41) Yeah, exactly. And everything is very modular in the way that we built. And so we can usually tell where exactly that novel document is failing and fix it. And then that fix is scalable across all other types of documents that even kind of marginally look like that. It doesn't have to look exactly the same. [Desmond Fleming] (30:41 - 31:36) So how do you guys deal with the customer service element of this market? So financial services, investors, persnickety bunch, they want accuracy and they care about accuracy and precision. So do you, are you guys kind of like in this kind of iteration loop where you feel very confident about the documents that you have seen, but if you come up with a document that you haven't seen, then you kind of have to fast follow that being flagged by your client? Or how do you think about, you know, managing the exceptions or use cases that come from them largely from, hey, you guys are ultimately still in the service business, so you want people to be happy. If you said, hey, 95% or 99% of documents we see work out of the box, but there will be some exceptions and that exception can come up. So how do you solve that problem? Part of it is just what the product is. [Eric Shapiro] (31:36 - 32:57) We're effectively delivering relatively usable data much more quickly than any other service could give it to you. And whether that service is calling the CFO or head of FP&A at a company you're invested in and asking for XYZ output, whether it's an offshore team or a junior analyst, ultimately you're waiting some amount of time, even if you're spending that time doing something else for that data to come to you. We deliver it very quickly. So we don't always get it right. We, you know, it's very frustrating when we don't because that need is so immediate and it is kind of time bound. But I'd say a lot of our customers don't necessarily have the expectation that it's perfect. They just have the expectation that it's quick and available. If though there are major errors, like we will more than happily get on the phone with them and try and work it out. Not so much with software because I think in order to build a system that's resilient, we can't go in and make tweaks super quickly to the code. But it's a spreadsheet. We can fix that. And, you know, I've worked my fair share of long nights, like every once in a while, it's really something we're happy to do because that means that it's a need that we can get closer to our customer. So I think, you know, if we do that every once in a while, it's we're more than happy to do it. And it does build a good relationship. [Desmond Fleming] (32:57 - 33:42) Yeah. Yeah. Yeah. You kind of forge yourself in the fire alongside the customers. You guys alluded to where where you want to go and your vision for for understory. I'd love to hear what that kind of update is because one component of it is, hey, you guys are not wholly levered to this application AI wave, but you're competing in it. And there's a lot of competition in the space so people, you know, less so a lot of people saying, hey, we'll let you chat with your document, we'll pull out the relevant information. And that comes in a ton of different variations. We don't need to name the names. But how are you guys thinking about a how to navigate that competitive environment? But B, where do you want understory to kind of go as a product and as a business? [Eric Shapiro] (33:42 - 35:39) There are a lot of products that are more overtly LM forward, if you will. And I think they're amazing. I would have loved to use all of them. Basically, I'd say one niche that we're in is is just focused on quantitative data, at least for now. And we think that's differentiated because you do kind of need this pipeline and breaking down quantitative reporting into all its constituent parts in order to get an output, rather than just asking LM to, you know, kind of like produce a time series to do a great job of that. In terms of where we're going, I think we're really guided by this idea. And it's a problem that I encountered in investing all the time. I used to invest in the aerospace supply chain. And so people would say, like, hey, here's an engine widget maker. We've looked at a bunch of these before. Maybe we looked at them in a private equity context. Maybe it was a credit deal that we were queued in on, but, you know, it's not broadly aware, or the market's not broadly aware of. Is there like a margin opportunity for this business? Is there a sales growth opportunity for this business? Ultimately, those inputs, particularly in private markets, they're not broadly available. So you're kind of just guessing. But if you can find like an intellectual ballast in your underwriting that says, hey, you know, I think there's an opportunity to grow profits here or grow end markets here because we've looked at a bunch of these other businesses and we know, you know, what the what the metrics should look like. That's something that is extremely important and valuable to an investment firm. And I think in a lot of cases, like the best of the best, they kind of synthetically create that rather by just investing in very niche subsectors repeatedly. But we want to unlock that for firms that don't have $100 billion of scale or don't have specialist investors where we can surface that information for you. And when your IC says, how does this compare to the five other deals we've looked at recently? Instead of that taking three days, you can pull that up in the meeting. That's that's our vision. [Desmond Fleming] (35:39 - 36:34) You know, every investor is like, yeah, pattern matching is a part of the job, right? It takes pattern matching from being an art into a science. I'd love to hear a little bit more along the direction of where you guys want to go of how do you guys think the investing world is going to change, especially in this kind of AI era, you guys are, let's call it part of the vanguard who's pushing people to adopt these products. But you're also realizing, hey, there are kind of speed bumps. And those speed bumps are largely also people, for lack of a better word, you guys could say, hey, we can make this happen today. But people are like, it's not a priority. So like, how do you guys think the investing world will change? Do you think from what you've observed, people are on the leading edge? A lot of people are talking about it. But what's your perspective? [David Fine] (36:35 - 38:29) I think even since we started, there's been a bit of a sea change in the mentality of many of the end users and potential customers that we're talking to. And all credit due to effectively chat GPT. I think at some point, it seems like every single head of fund CEO got together and said, What are we doing about this? And then they went to all their MDs and portfolio managers and CEOs and CTOs and said, Hey, what are we doing about this AI thing? And so I think we've actually been a bit of a beneficiary of that to a certain extent where people are much more open to talking about workflow efficiency and analyst efficiency than maybe they were five or 10 years ago. But at the same time, that makes it a much crowded space, much more crowded space. In terms of just where I see things going, what Eric just described as our vision, the way I see it is a lot of innovation over time, outside of even just like standard technology. If you if you look at like the Renaissance and the way that really intricate art happened, a lot of that was just enabled by innovations around tooling and what people could accomplish. So really advanced silversmithing was enabled by kind of like metal rollers being much, much better, and then being able to roll silver very, very, very thinly that then enabled silversmiths to have much more intricate designs, much more intricate work. And that's kind of the way I see AI working with an industry like finance. You have a bunch of silversmiths currently who are trying to build something intricate, but oftentimes they're using very clunky sort of silver and companies like ours can be the metal roller to make it much thinner. [Desmond Fleming] (38:29 - 38:46) So does that mean the throughput of investing increases? So does that effectuate itself in more AOM being unlocked for the industry, or is it more on the back end of saying, hey, the return streams within investing will be higher? That's a great question. [David Fine] (38:46 - 39:22) I haven't thought about the analogy all the way through there, but I think in many ways it's probably going to end up being the throughput, right? Like that, as far as I can tell from the outside looking in, seems to be the great limiting factor for a lot of these firms, right? A lot of them have raised a ton of money. There's a limited talent pool, and they want to deploy capital efficiently, but also effectively and in a way that generates alpha. And so that is a difficult problem to solve, as you're well aware of. The painful problem. [Desmond Fleming] (39:23 - 39:23) Yeah, exactly. [David Fine] (39:23 - 39:53) Anything that enables comprehensive analysis and novel analysis and unlocking proprietary insights in a really easy kind of click a button sort of way, I think will just enable the good investment firms to make better investments faster. And that, to me, seems to be a value proposition that a lot of these firms will bite at. But I don't know, Eric probably has a better perspective on this. [Eric Shapiro] (39:54 - 41:23) Yeah, I mean, on some level, if you can reduce the amount of friction that it takes for particularly like a private company to access capital markets, that's good for everyone, right? And I think what it ultimately causes an investment firm's edge, and it really is this way now, like it's not that you have the best ex-banker associates who build an LBL model for you, right? It's decision making within the firm. And I don't think that will ever go away. I think, you know, having the intuition of an investor is a craft, and I don't think you can replicate that with artificial intelligence, or at least I don't know. I agree. You know, it's sourcing and deal origination, and it comes down to expertise. You know, AI will help you make decisions faster on the basis of criteria that you pre-specify, and it might cover some blind spots. But ultimately, if you know how to invest in a, you know, an auto parts company, and you know what to look for, if you can find the opportunity, diligence it quicker, and then basically encode your judgment around what makes a good versus bad investment in a process rather than people who, by the way, only have the career longevity to operate at that high pace for a few years before they need a promotion opportunity anyway, it'll actually allow the DNA of an investment firm to be sort of more true to itself rather than dependent on specific people. And then it's just a question of bringing Yeah, that's really interesting. [Desmond Fleming] (41:23 - 42:12) It's about, I'll use a word, encoding the processes and decision making of a firm, right? So as long as the, if you know what your quantitative and qualitative inputs for success are, as you identify those, and as you gather that data from the market, for lack of a better word, then you can kind of help people make that a more repeatable process over time, right? Because, you know, every firm goes through processes of recruiting people, and they think people will stick, and you know, maybe someone is meant to be an investor at Tiger instead of Ballpost, or they're meant to be an investor at Ballpost instead of Co2, right? And I imagine, I'm just picking on those because they're pretty big brand names, but you know, there's a certain style of investing in each one of those firms. [Eric Shapiro] (42:12 - 42:19) Of course, I think, and I don't know if you guys have this expression, but when I was at Elliott, there was always this expression, is this an Elliott trade? [Desmond Fleming] (42:19 - 42:26) Yes, we talk a lot about, is this a first mark deal? Is this a business model deal? That always comes up. [Eric Shapiro] (42:26 - 42:54) Right, and what does that mean? I mean, question for you, but like, I think it comes down to why are we getting this opportunity? How did we source it? Does it have the structural features where we are good at managing this investment after we make it? And do we have the insights that give us a differentiated view? I don't think those things will go away, but if, again, if you can plug it into a systematized process aided by technology, that's really powerful. [Desmond Fleming] (42:54 - 44:04) Not to oversimplify it, but it's a little bit of helping you put up that objective filter, right? Of like, hey, yeah, you may be really excited about this opportunity, but it's not a business model deal. It's not a Elliott trade. And it helps you, again, summarize that. So I know one of the features you guys have is like the click through, how do you guys describe it? Auditability. The auditability of the platform. So you could even, or at least I'm getting a little future tripping here, but I could see at some point in the platform where it's like, hey, here's the auditability of the summary of why we don't think this is a good deal or not for your own company. Guys, we covered a lot. This was awesome. I think maybe the last question I have, and this is something I care about a lot because I think selfishly, but also I genuinely believe this, I think more people should start companies. I think more people should be exposed to the process of entrepreneurship. And so you guys have obviously taken that leap. And if you were speaking back to the version of yourselves two years ago, three years ago, what is the advice you would give to that version of yourself? [Eric Shapiro] (44:05 - 44:07) I know what I'm going to say. I'm just waiting for you. [David Fine] (44:08 - 44:08) You go first. [Eric Shapiro] (44:10 - 44:14) I think you need to find your David. Yeah. That's very sweet. [David Fine] (44:14 - 44:15) That is very nice. [Eric Shapiro] (44:18 - 44:57) It's an emotionally volatile experience. You have to go into that eyes open, by the way, like there's that, there's that kind of advice. Like it doesn't always work out. Don't look at, you know, your friends raising tons of money and think that that's inevitably going to be you, et cetera, et cetera. But part of the fun of this is that you get to choose who you work with, which is a unique feature of this job versus any other. Yep. And that choice should be one that you make intentionally. And that gives you a lot, just an enjoyable experience. And so I get to work with my best friend. I've also found out all these ways that he's talented that I didn't even realize. And I mean that. What's an example? Great at making late night snacks. [Desmond Fleming] (44:58 - 45:00) He knew that. [Eric Shapiro] (45:00 - 45:14) We were roommates. I don't know. We needed a website. He just made one. Even now he's developed all this expertise on LLM technology, which no one knew about except for a select few, a few years ago. [Desmond Fleming] (45:14 - 45:19) And it's an emerging skill, highly valuable again, going back to the recycling in the industry. [Eric Shapiro] (45:19 - 45:42) And also, you know, we should call out our third co-founder, Alex. Like he is also in that vein, just a jack of all trades. He works super hard, very, very smart and a great person. And so there's a lot of things about being a founder that suck. But for me, the people I get to work with don't ever suck. And that's very powerful. Yeah, that's awesome. [David Fine] (45:43 - 46:10) Yeah, that's, thank you. And the interesting thing is for me, as I said at the beginning, I felt like I always wanted to start a company. But I think because I read about it so much and thought about it a lot, like I always knew it was really essential to find good co-founders. And so that to me was table stakes. And when Eric kind of, I never expected Eric to want to start a company. [Eric Shapiro] (46:11 - 46:11) Right. [David Fine] (46:12 - 48:07) And so when he, we were just talking about this and when he was like, maybe I should, maybe we should just go start a company. I kind of leapt at that opportunity. So to me, that was very much table stakes. And likewise, like being able to work with your best friend and being able to just sit and think about complex problems with them every day is amazing. In terms of the thing that I would say is we got, I would say we got extremely lucky with Alex, our third co-founder. I'm semi-technical. I know some data science stuff. I know some engineering stuff, like maybe like intern level engineering, junior engineer sort of thing. And Eric obviously is a subject matter expert around what we're trying to build. So we obviously needed a technical co-founder, someone who could really quarterback the development process and the engineering process of what we were trying to build, which is rather complex. And we have a I think a lot of them work at larger companies and we're not interested in the startup life. And I actually am, at the time I was a little despondent about this, but I'm actually very grateful that that happened because what I realized with Alex, who likewise, similar to Eric, I think has more of a passion for investment decision-making and what it is that we're actually doing than any of the other people, our friends who we talked to about this idea. And so I would have told myself to try to kind of skip to somebody exactly like that. Instead of waiting through kind of all these really brilliant engineers that we knew, but it was very much trying to fit like maybe a circle into a square peg. [Desmond Fleming] (48:07 - 48:21) Great on paper, but lacking the passion. Also, I will put this out into the world. I think it's startups should no longer be framed as people not being willing to take the risk. They're averse to opportunity. That's the reframing that I'm putting out. [David Fine] (48:21 - 48:25) Yeah, we tried that framing a lot. It did not go over well. [Desmond Fleming] (48:26 - 48:36) Can't you see this opportunity in front of you? David, Eric, appreciate you guys so much for stopping by. This was awesome. Hopefully you guys enjoyed it as well. [David Fine] (48:36 - 48:37) Thanks for having us. Yeah, it was great. Thank you. [Desmond Fleming] (48:38 - 48:38) Awesome.