Matt Ober

Clara (0:00): Matt started at Bloomberg, moved to WorldQuant, one of the most secretive and sophisticated quantitative funds on earth. Then he became Chief Data Scientist at Third Point, a $20 billion hedge fund. Now he's Managing Partner at Social Leverage. He's one of the few people who can tell you what data is really worth because he spent his entire career finding out. Welcome to Conversations in Data by Tech Accelerator. I'm Clara, head of people and I'm so excited for our chat today.

Matt (0:27): Excited to be here.

Clara (0:28): Absolutely. Let's get started with a couple of questions about the data career nobody planned. Bloomberg, Quant, Social Leverage. Every stop in your career has been at the absolute frontier of using data in financial markets. Was this a plan you had or did you just keep following the most interesting problem in the room?

Matt (0:48): Uh I think you know Bloomberg definitely wasn't planned until I landed there in San Francisco and then eventually moved to New York. I think I quickly knew working at Bloomberg that I wanted to be on the other side of the terminal. You know, I started answering phones at Bloomberg and then it was the instant Bloomberg chats where you get yelled at from people that are using it to make money. And so I spent a lot of time and effort getting to the other side which landed me at WorldQuant. I don't think a quantitative hedge fund is what I thought I would be at. But it ended up working out great and you're surrounded by people that are, you know, all smarter than you. So, it's always nice to be in the room when you're the least smartest person in there, but also provide a lot of value because you have a different skill set. So, I was, you know, working with all, you know, arguably PhD rocket scientists. And then I think from there, you know, I on a whim met a recruiter that was working for Dan Loeb. And over a year of meeting him and at the time his head of TMT, they asked me to join as their Chief Data Scientist, which I think, you know, at that point, Chief Data Scientist was really just like the title. I think they really wanted me to think about data, technology, risk, everything they did to make the investment team, you know, bring it to the 21st century. You know, make better decisions using data and technology. And then, you know, within WorldQuant and Third Point, I did venture investing. So, WorldQuant, we started the ventures team, me and a guy named Steve Lau. Third Point had, you know, a couple billion in venture investments and I got to lead a deal or two there while I was working there. So, it was always something I was interested in. I was an LP with Social Leverage for a long time. So, it was just a matter of time. Venture is a different game than the hedge fund world. So I, you know, I think it was the right time post-COVID to make the move.

Clara (2:36): Quant is famously secretive about how it works and without getting into anything proprietary at all, what is it like to build data strategy inside an organization where the edge is the data and protecting it is existential?

Matt (2:52): I mean I think my experience probably is different than people that have been there in the last 10 years because like when we were there we were kind of like building the strategy from day one, right? I think data was part of the strategy at WorldQuant. I think it quickly then was a realization we had, and Igor had mainly, that like if we could consume more data than anybody in the world we could manage more money. So then it was really about you know how do we do 100 times more than we did last year, right? Every year it was like how do we 100x this? Igor was pretty famous for always saying that, and so you know when you have people around you that can build anything you can think of, my job was just thinking about, you know, as like a factory, how do we continue to fill the funnel and what are the roadblocks to speed that up? You know, we kind of, I think, changed the hedge fund approach of, you know, not taking phone calls and not treating vendors with respect to being like, we'll take 100 vendor meetings a week and bring them all into the office. And so, I think like we kind of flipped everything and you know, we also kind of were cold calling and cold outreaching to companies that never even thought about selling their data. And that's how I met my partner Howard Lindzon today. He started StockTwits and I cold called him and, you know, we wanted to buy the entire API from StockTwits.

Clara (4:10): Very interesting. And well, I guess cold calling works, right? I go on LinkedIn every day and I see posts of SDRs or Hunter accounts saying, "Stop cold calling." No, cold calling is the best, but yeah, at the end of the day, I think it works. And how were you getting that contact information?

Matt (4:29): I think we were just really good at Googling back then. We also showed up at every conference in the world. I mean, as much as there's these data conferences now, we were at weather conferences. We were in every country, Russia, China, India, everywhere in Western Europe. Our travel budget was pretty large. And we kind of saw the world as, you know, it wasn't really a question. If I wanted to go to a meeting in Beijing, there wasn't really a "this is a big deal." Like we were opening offices in every country. So it was kind of like get your visa and go.

Clara (5:03): That's definitely the goal, right, at Tech Accelerator. So you know Tech Accelerator recently got acquired by Data Sharp which is a company from Belgium. Tech Accelerator is from the US, Miami but now we have under Data Sharp, which is a company that I also work at, we have Tech Accelerator in Miami, we have Woomap in Paris, Montpellier, London and Brussels. So that's definitely the goal to travel and continue opening offices. We have been debating opening in other countries as well, hiring people all over the globe. So that's the goal and this growth is very very exciting. You went from being one of the best data scientists at a hedge fund to being a VC who bets on data companies. I love taking bets. Those are very different things though. And what did you get wrong about venture capital before you actually did it?

Matt (5:56): I think just how big you need to think. I, you know, I was doing venture investing at WorldQuant, but we also had a strategic angle there, right? Where like we didn't need companies to be maybe venture scale because any returns were good if it also helped the hedge fund at the same time. I think when you manage money externally and you have LPs, you know, like I tell founders when we meet them, like I need to assume I can get a 10x on my money. Not all the companies are going to get there and most aren't going to succeed, but like I have to underwrite in my mind that you can return the entire fund with this one investment. So if I manage a $100 million fund and I invest in you at 5 million and I only own 10%, like I need you to be a... do the math of like how big I need this to be.

Clara (6:36): Yeah, definitely. Always thinking big. And you co-authored a book on options strategies for volatile markets, data scientist, hedge fund executive, venture capital analyst, author. What's your favorite part of all this that you've done?

Matt (6:51): My favorite part, I mean, I think the experience of living in New York and working at WorldQuant was... I couldn't write a better story. You know, WorldQuant was small when I got there. We had maybe two, three offices. We expanded to like 25 in 6 years, right? We went from like 80 people to 600. So I think that experience, you know, I went everywhere from Russia to Thailand to China, like every country in between. And you know, Third Point in my mind is more of an actual real hedge fund. WorldQuant in my mind is more of like, you know, it's a quant shop but it's like working at a technology firm. We didn't look at the markets all day. When you're at Third Point like you live and die every day like seeing, you know, ins and outs of every company. You know the company well, we were activists, we were taking over companies. It's also a different scale there and a different mindset, like very much like you sometimes see on TV. I would say those two things and just like the experiences and the stories and the travel it was not something I could have ever imagined and it was a great stepping stone to where I am now.

Clara (7:58): We're going to talk a little bit more about what good data actually looks like and we'll jump a little bit to Social Leverage of course that's the main goal. But at Third Point, you built the data and analytics platform for a $20 billion fund. What's the difference between a data strategy that really changes how a firm makes decisions and one that looks impressive in a board deck but doesn't really move the needle?

Matt (8:23): I mean, I think it's all about the culture within the firms, right? You can build all these things and it looks cool, but reality is if there's not buy-in from the analysts and PMs that are supposed to be using it, then like you're kind of just wasting your time building cool things and cool gadgets. And I think like that was one of the things we did really well at Third Point is like we built a research management system where all the notes and the insights and everything that was important to the investments was transparent, and then allowed the data science team to essentially deliver value rather than hoping and dreaming that what we're doing is helpful.

Clara (8:59): I really love that you said that, that you know a lot goes into culture and what actually really matters, right? I'm trying to build culture every day so I resonate with that a lot.

Matt (9:10): I mean listen, it sounds like a joke and people can roll their eyes about it, but like you're going to invest all this money and time and data and data science teams and all these things, especially at a traditional investment firm, like if you can't get buy-in and it's not going to be used, like why is everybody working hard to build these things? I think you got to have buy-in from the top and you've got to have, you know, you fire people if they don't want to buy in. It's no different than what's happening with AI right now. Like it's a big push from a lot of firms that like we're going to leverage these tools. Not that we have to get rid of people, but if you're not going to understand that you can get, you know, 3 to 10x more out of yourself from an efficiency perspective, you're not really leaning into what these tools are capable of.

Clara (9:52): Yeah. That's why I think in the moment of hiring, you need to be aligned on those aspects, right? If you're not aligned, then it's just a waste of time and money because this person's going to leave in six months, either, you know, quitting or getting fired. So, real prop. But um, alternative data has been the gold rush of the last 5 to 10 years. I would say everyone wants it, right? But a lot of it is noise. And in your experience, what percentage of alternative data actually earns its price tag?

Matt (10:24): I think if you buy it at the right price, there's endless value in all these data sets that are out there, especially if you're using them in different ways. Is there a lot of slop and there's a lot of headlines about these multi-million dollar data sets? Of course. And I think like that's noise, but I do think that, you know, if you're buying things at a fair price for the value that you get out of them, I mean, it's why people are spending hundreds of millions of dollars at different funds, right? Like they're using these things. It's answering questions. Obviously, quantitative strategies are very different and have different needs than like I'm an investor on the fundamental side in Starbucks. You can imagine everything from coffee beans to app analytics are going to be, you know, important to us. So what we can pay and what we're willing to pay depending on how big our position is, is very different than a quant fund making that one of, you know, a thousand positions.

Clara (11:12): But Social Leverage now was an early backer of Robinhood which went public at a $30 billion valuation. What did you see in Robinhood at seed stage that the rest of the market do you think might have been underestimating?

Matt (11:28): Yeah, I mean I give credit to that to my partners Howard and Tom and I think for them, you know, they saw... Howard was an investor in StockTwits, you know, like having the pulse of retail, like even to today where we're, you know, investors and he's CEO. I sit on the board. Like you get a sense of what's important and where trends are going. I think, you know, that was when it was $10 to trade on E-Trade, right? Every time you bought and sold a stock. I think they made a really beautiful, elegant app and had a vision that, you know, things would be done at zero. And I remember the first time even Howard told me, I thought it was crazy, you know, like zero fee trading. But, you know, it's a good example like my perspective was different. I worked at a hedge fund. I didn't even have a brokerage account, right? Cuz I couldn't trade anything on my own. So, I didn't have a Robinhood account until like 3, four years ago because I was never allowed to make any investments because I always worked within a, you know, a regulated institution. Whereas, you flip it to somebody like him, like he was investing all the time on his own and knew all the apps that were out there were horrible and the costs and the fees that are associated with it. And like I think at the end of the day, you're investing in people and he met Vlad and Baiju at the time and you know, that just was a great fit in terms of the skill set they have. If you know the Robinhood story, it wasn't Robinhood at the beginning, right? It was like an algorithmic trading company.

Clara (12:47): Yeah, I like that you said that you're betting on people and that's something we're going to talk a little bit more on other questions that I have for you. But I feel like at every firm, every company group, your people, your culture are what makes the company and you're taking bets on people every day. You now sit at the intersection of fintech and data as an investor. Now, how many of the startups pitching you have a real data advantage versus a data story? I mean, how do you tell the difference?

Matt (13:18): I mean, I think it's hard to have a data advantage at the very beginning because we're investing at the seed stage. These companies are just being built. So unless they have some sort of proprietary access to something or they've been building for a long time but have never taken money really investing in their vision on what they're going to get to or the system that they have that's going to capture the data. Obviously there's web scraping and everything and that can be unique and helpful. But I think this is why it's always been hard to invest in data businesses at the early stage because one, there's not much to evaluate at the beginning and then two, like most data and information services companies don't get that big, right? They do 3 to 10 million in revenue and they kind of just stick to that with small growth. And the best ones have always been like serial acquirers and rollups over time. Even if you look at AlphaSense, you know, they've made a lot of big acquisitions at this point. So it's tough. We've gotten a few at the firm now and most recently Fiscal.ai and so like I think we've got the right bets and we can continue to make them and I think obviously with AI and how things are changing right now, like the data landscape is quickly for that stealth kid that dropped out of college.

Clara (14:30): Thanks for sharing that with me. Investing at the edge. Talking a little bit about Social Leverage. You just raised its fifth fund. You've made over 125 investments since inception. At this point in the firm's history, what kind of founders do you say yes to? Like we were saying, before they even finish their pitch, what's something that you're like, okay, yes?

Matt (14:52): I mean, I think it's domain expertise and confidence. I think at this point founders really have to have a go-to-market, distribution, sales and marketing plan or moat, right? Like think about how easy it is to build product right now. At least in the areas we're in, right? We're not doing deep tech. We're not building robots. So like these founders, they need to have capability to build product and somebody technical on the team, but we're really betting on how are they going to get distribution and sales. Like the best companies are doing millions in revenue in months instead of years. Obviously, not every company is like that. Fintech takes a lot longer to build. Data takes a lot longer to build as well. You know it's having the vision that's big enough, having the confidence, having the background and experience and also for us like we have to be passionate and excited about the idea.

Clara (15:43): Now seed stage investing is fundamentally a bet on people like we were talking about, not just products. I take bets on people every day like I was saying I'm head of people at Accelerator, head of HR at Data Sharp, people people. But people change under pressure. Have you ever backed someone who completely transformed for better or worse between the seed round and what came next? What did that teach you? Can you tell me a story about that?

Matt (16:08): I mean, I think in a generic level, I would say yes, all the time it happens. I mean, I think also you got to think about like we're investing at an early stage and we're a 10-year fund, right? The best companies take... Robinhood took 9 years to go public. eToro took 13. Alpaca is one of our funds from 2018 just crossed over a billion in valuation and is really taking off. It takes a long time, right? Life events happen. Marriage, divorce, family, you know, all these different things. Founder breakups, like obviously that changes people. We try and spend time, I think, with the founders, but you know, for us, we're not doing just a Zoom and writing a check. We want to go out to dinner, have a meal with them, get to know them, talk to people that have known them, like do enough of the work so that we can kind of get a sense of who they are and then why is it that they're building what they're building. And have they had adversity in their life that they've had to deal with and what did they learn from that? I think that's kind of important because no founder journey is straight up and to the right, right? There's going to be a lot of twists and turns, like pivots. You know, I've seen even in the last couple years a whole host of different issues and situations, but also companies that were on their last month of funding and then quickly turned it around, right? So you never know, but you're betting on the people that are going to grind and are respectful with the idea that they're taking somebody's money and that they want to do right by the investors that are also putting in a lot of time and effort to help them.

Clara (17:36): And sometimes when you're, let's suppose, hiring like what I do, you don't get too many chances to get to know people before you say yes to hiring them. And then the first month in the job of this person and you're like, well, that bet didn't work out. I didn't know enough about this personally. And that happens in the corporate world a lot, right?

Matt (17:57): It's easier in the corporate world because you can fire fast and you know mistakes happen, but like you don't want to have to fire but at least you can let people go if they don't fit. Once we give our money we're kind of locked in. So we don't have that opportunity to get a "let's start over after 30 days."

Clara (18:15): Yeah. Yeah. Definitely. You stay disciplined on valuation which is a contrarian position in the world where you see deals often get bid up before due diligence. How do you hold that line when everyone around you is moving faster and paying more potentially?

Matt (18:31): Yeah. I mean, listen, I think there's a time and a place here and there to break the rules. Maybe it's a founder you've backed many times before, but I think, you know, my partner Howard always says it comes down to the math, and I think he's right. Like, our LPs have seen in our previous funds, you know, multiples of their money back, DPI. We want to see at least a 3x DPI, but like 5 to 7x is much better, right? And that's what you've seen from funds in history. So I think you got to back into how those numbers work and it's a lot easier to return the whole fund if you invest in a company at a 4 to 8 million valuation than if you invest at a 40 million valuation. Also like we kind of just think that there's so much more risk the higher the valuation is for the founder. If I give you $2 million at a $25 million valuation and you don't execute perfectly, nobody's going to do your next round at a higher number. So then you're in a down round or a flat round and like reality is 95% of investors don't do those type of rounds. So now you're struggling just to figure it out versus if you start like, instead of growing into your valuation, why don't we do something that's fair at the beginning and give yourself optionality. I think that's what we also tell founders. Like not every company needs to sell for multiple billions of dollars to be successful. You could sell your company for 300 to 500 million, life-changing money for you and your family, and our entire fund is returned. We're happy. That's a great, great exit and you don't need to raise endless amounts of capital to do that. And then maybe if it happens to be that you built the next Robinhood, well then now you can go raise hundreds of millions of dollars and go for it. So I think it's just being smart for everybody. Lower the pressure.

Clara (20:09): Yeah, I see. I'm now jumping into AI, data, and the markets. Right now, you invest in vertical AI and fintech, but we're now in a world where every startup claims to be, you know, AI powered. When you strip away the language, what does AI actually add for a financial services company that a good data strategy couldn't have done before?

Matt (20:30): I think if I look at AI for pure data and information services companies, the way I look at it is if you look at S&P and FactSet and Bloomberg, like a place that I worked at, you have tens of thousands of employees offshore collecting data, cleaning it, organizing it and pushing it out. If you look at a modern company like one of the companies we invested in, Fiscal AI, you can arguably build these humongous fundamental data companies, right? Like something like what SNL or CapIQ has built. You do that with 30, 40 people because you have AI agents doing four or five layers of collection, extraction, and QA. And then if all of the agents are in agreement, you can push it out faster and more accurately than the big guys. And if they aren't in agreement and there's some sort of issue, then you have a human in the loop, right? So you still want that human in the loop because 99% accuracy is the only thing that's going to really win over the institutions. You get what you pay for, right? Like you said earlier, sure, we could push things out and sell it for $99 a month and it's going to be inaccurate, it's going to be slow, there's going to be all these problems, but if we have full accuracy, speed, everything, then we can charge the premium pricing, still be cheaper than the incumbents because they're overpricing things. We can have better terms and conditions and we can guarantee that we're going to continue to push out these structured data points and KPIs arguably faster than others can because we are building from an AI native capability, right? So our technology is different, our ability to push updates is quicker, take feedback quicker. And so you know, if you look at the announcement yesterday, Kalshi came out with KPI prediction markets powered by Fiscal AI. They're using Fiscal's KPIs because when Uber announces earnings and you dive into how many rides did they do for the quarter, Fiscal is able to grab that data, structure it, and push it out quicker than anybody.

Clara (22:24): So, do you think that 30 to 40 people with AI agents is better than having 10,000 people offshore doing the job?

Matt (22:32): Oh, 100%. Yeah. I mean, listen, I worry about the offshore companies that were doing that before. They need to become, you know, more. If I have 10,000 people offshore and I'm not getting three times more out of them or in some cases it should be 10 times more, then we're not really embracing AI. Like if I poll my founders right now in our portfolio, the best ones are seeing 3 to 10x more efficiency out of their engineers. That's not even talking about the business teams which are also finding more efficiencies, right? Like they're doing less data entry, they're doing less prospecting, all these things. But it depends on how much they're really leaning into AI and taking it seriously.

Clara (23:11): I see. Well, I wanted to bring up a hot take for you to react to, which is that the financial industry has more data than any other sector, but it makes some of the most irrational decisions on Earth. What does that tell us about data versus behavior? What do you think about that statement?

Matt (23:30): I mean, as somebody who worked at a hedge fund where it was a fundamental shop, I don't disagree because, you know, at the end of the day, these guys have built firms and are managing money based off of their intuition, patterns they've recognized, and sometimes they don't believe in the data and at the same time the data is not always right. So I think that investing is part art, part science. If you're not in a quantitative seat, right? Put the quants aside where everything's systematic. The decision-making is not all, you know, based off of data. At the end of the day, think about an activist, like there's a lot of human behavior that's wrapped into that. And so more data is better. I think we're going to have 10x... Everybody wants to have 10 to 100 times more data. They want to pay a lot more, and we're going to over time spend more. But there's still going to be this huge value add of all of the people that were doing a lot of the grunt work in the past. Now we can spend more time in face-to-face meetings and more of that qualitative side.

Clara (24:26): We'll start wrapping up here. I wanted to ask you just, you know, that cheesy question probably comes from my interviewee side which I also hate. I hate to ask this question about where do you see yourself in 5 years, but about the industry... where is the edge going to come from in 5 years if alternative data is becoming commoditized? What's the next frontier for companies that want to use data to compete?

Matt (24:53): I think that AI is going to make all of the grunt work that everybody was doing much more efficient. I think Brett Caughran from Fundamental Edge writes a lot about this on Twitter. I think the analysts are going to be doing, you know, what used to take weeks and months for model building and presentations, all that stuff is going to come down and be commoditized to a few minutes. And so you're going to need to spend the time when you're young learning and understanding the intricacies. But the edge is going to be coming from showing up to the conferences, networking, face-to-face meetings, and really understanding people behavior, and then taking all this stuff that's been commoditized and getting more and more data and using that alongside of these qualitative things that you can't really measure. And I think that's what's exciting, and I think it's going to be interesting to see where the value swings in terms of how much people are investing in. You know, we used to pay all this money for quants and data scientists and engineers and now does the pendulum swing to the people that are better at the conversations and sales side of things?

Clara (25:57): Yeah. And I think we just went out from an era that everyone was working from home and not connecting to going back to conferences, having dinner, interacting in a more qualitative way than before. So this is definitely a huge game changer right now because yes we do have all these AI agents, Claude, ChatGPT, Gemini, they can do things for you in 30 seconds to a minute, but at the end of the day you still need to be there, you still need to show up. It's still made by people today. But um anyways that's it on my side today. Thank you for answering all my questions and joining me at Conversations in Data. Thank you so much again and we'll see you next time.

Matt (26:41): Thank you for having me on.

Clara (26:43): Yeah.