Seed to Exit
Welcome to Seed to Exit, the ultimate podcast all about startups, scaling, and venture capital. Your host is Riece Keck: Startup veteran and recruitment entrepreneur.
Join us as we dive into the journeys of startup founders and venture capitalists who share their insights, successes, and lessons learned from seed stage to successful exit.
Each episode, we bring you candid conversations with startup founders, executives, and investors. Whether you're looking for inspiration, actionable advice, or a deeper understanding of the startup ecosystem, Seed to Exit offers invaluable knowledge and real-world experiences to help you on your entrepreneurial journey.
Tune in to Seed to Exit and get ready to be inspired, educated, and connected with the exciting and ever-changing world of startups and venture capital.
Seed to Exit
Albert Azout, Venture Partner at Level Ventures | Navigating Venture Capital | Insights into Fund Management, Technological Edge, and Strategic Investment Strategies
This episode explores venture capital through the eyes of Albert Azout, Managing Partner at Level Ventures. We discuss thematic investing, the fund of funds model, and the significance of data analytics in selecting promising funds and sectors in today's market.
• Albert's journey from founder to venture partner
• Thematic investing as a strategy in venture capital
• The fund of funds model explained
• Importance of data science in identifying venture opportunities
• Sectors currently outperforming in venture investments
• The intersection of personal characteristics and investment success
• The future of crypto amidst regulatory challenges
• Balancing investment allocations between funds and direct co-investments
Don’t forget to subscribe for more insights and strategies!
All Links: linktr.ee/startup_recruiting
LinkedIn: www.linkedin.com/in/riecekeck/
Twitter/X: x.com/tech_headhunter
Recruitment: www.mindhire.ai
Youtube: https://www.youtube.com/@seedtoexitpod
We believe that in venture, there's heterogeneity in terms of where innovation is being created and where returns will be created and where the supply of later stage capital will be. We don't believe, generally speaking, in just being a generalist, and so we take a thematic bet in each fund cycle, based on where we think the market will go, and so we care about which areas our fund managers are indexing against and whether they're specialists in those areas.
Speaker 2:Today I'm excited to welcome Albert Azut, venture Partner at Level Ventures and a seasoned entrepreneur and investor. Albert's portfolio includes unicorns like Hippo Insurance and Oak North, as well as exits like Uber, quickie and DeepScale. He's also founded and sold successful ventures like Velos and Stylecaster. In this conversation, we're going to talk more about Albert's approach to building and backing businesses, his journey as a founder and his vision for the future of tech innovation. So with that, let's get into the episode.
Speaker 3:You're listening to the Seed to Exit podcast with your host, rhys Keck. Here you'll learn from startup executives, founders, investors and industry experts. You'll learn from the best about building amazing products, scaling companies, raising capital, hiring the right people and more. Subscribe and listen in for new episodes and enjoy the show.
Speaker 2:Albert, welcome on, excited to have you.
Speaker 2:Thank you for having me, albert, welcome on, excited to have you. Thank you for having me. So the fund of funds approach. I know, obviously it's, of course, it's a less common investment model that you do than you know. Call it your more traditional VC fund and we haven't had anybody on the show yet who runs a funds of funds. So I'm excited to dive into that today. Learn a little bit more about what the model is like. Before we get into that, I'd love to learn a little bit more about your background. You'd, of course, founded your first company, velos, back in I believe it was 2010. If you wouldn't mind just giving us an overview of what that looks like and how you got to ultimately founding Level.
Speaker 1:It sounds good. So actually that wasn't my first company, but I'll go back and give you sort of a history of me. So, software engineer, I went to school in Boston and then I went to work on Wall Street two and a half years as an infrastructure engineer, got more into more quantitative engineering and was starting to get really fascinated with data science and data analytics. And then I started a few companies in New York. Actually, my first company was actually a digital ad agency which scaled up. We also had a bunch of media properties and I did that for about three years and then I ended up wanting to build venture-backed companies. So my first company was a company called Stylecaster which was a large-scale ad network focused on fashion, beauty and media lifestyle. We were one of the largest platforms in the space. We raised capital and then we ended up selling to private equity and then from that business I had been thinking a lot about machine learning, because we were doing a lot of recommendation and optimization and whatnot and some of the challenges I saw with large scale data and so at that point in the market, you know, companies were starting to talk about predictive analytics and using machine learning at scale, and so we started to explore in that area.
Speaker 1:I recruited some, some folks from MIT and other places, we had a really good team and we started to work on what became a low code platform for machine learning infrastructure. And at that point in time that was not a thing, but essentially what we did was we helped companies get from raw data to at-scale machine learning, which had two parts. One was what they call feature engineering, etl and feature engineering, which is hard to do at scale, and the other was model serving, inference and so being able to to essentially, you know, deliver predictions at low latency. And so, yeah, we ended up selling the company to AOL and Verizon bought AOL and ended up, you know, going going over to them. And yeah, that's sort of the history.
Speaker 2:Cool, I love that. So how did you get from selling to AOL to starting level?
Speaker 1:So I was. You know, when I moved actually, when I sold to AOL, I worked for them for about a year and a half I ended up relocating to the Bay Area. So I lived in Palo Alto and had an office in San Francisco and about the time that I moved, I was already starting to meet with entrepreneurs and starting to invest a lot and investing in funds and other things. And then one of the investors in my company his name is Bobby Asdani was a prolific angel investor in the Bay Area and he was putting together a venture capital firm called Koda Capital and I ended up joining as a partner there in 2016. And I was involved in sort of helping build the firm. You know, the firm grew really quickly and had an interesting strategy which focused on both private investing and public investing.
Speaker 1:I was doing private investing mostly in this. You know, seed, series A, series B stages. Everything we did was enterprise technology, enterprise infrastructure, and we did some everything from like semiconductor to application level. We did. And, yeah, so that's where I actually learned the business of venture capital investing. You know, over four years I was there, you know, sitting on boards, leading investments. You know managing portfolio companies, et cetera.
Speaker 2:What motivated you to start your own firm?
Speaker 1:Yeah. So during the pandemic, I ended up wanting to move back closer to family, and so we ended up on the East Coast, back on the East Coast, and so I wanted to build my own firm, and this is like the last thing I want to do, I hope. But basically one of the challenges and I think it's just a general challenge in the ecosystem is just it's very opaque and as someone, for example, at an early stage firm, when you're trying to source, it's very difficult to understand what's going on. What are the C firms? What are they doing? What are founders doing? Where are they coming from? Where are they going? Who has access to them? What are the trends, topic areas that are evolving?
Speaker 1:It's actually a very difficult problem and usually it's done very, very manually, and so that was always something in the back of my head is like how do you actually scale that system? But the other thing that I noticed because we have a family office we were investing pretty heavily in technology was actually where we did the best from a performance perspective was when we invested in this notion of small VC funds or emerging VC funds and when we co-invested with them, and that strategy seemed to be very applicable and sort of very risk diversified, but really it's not done systematically, and so I was toying around with those two concepts when I started and that's kind of where we ended up, which is, you know, essentially reinventing the emerging VC platform and thinking about building a strategy that we think is, you know, really optimal in venture investing from an LP perspective. So that's kind of how it came together.
Speaker 2:So for those who are a little bit newer to the venture world, could you explain what the fund of funds model is and how it operates?
Speaker 1:Yeah, I mean generally speaking. In private equity and in hedge funds and whatnot, you usually use a fund of funds as an intermediary to invest in funds. So of course an LP can go directly to funds. But often it's very difficult to underwrite funds directly in some cases and or to get access to them. And so traditionally when a fund of funds existed is either for like underwriting or you know sort of underwriting alpha and or access, and you know it's a structure that exists in the market and it services.
Speaker 1:You know everything from hedge funds to private equity to venture capital, generally speaking. And so that's the model and the idea is to deploy capital for LPs into funds and to manage all of the investing activities underneath that. So that's sort of how it works generally speaking. And usually investors don't necessarily like fund-to-funds because there's typically a double layer of fees because you're paying both the underlying manager as well as the fund-to-fund funds. But in venture we believe that's kind of a different problem, just because in venture you have the ability to have outliers in performance and even with fees you can actually do very well by diversification. But we can talk about that separately.
Speaker 2:Per fees is 2 and 20 still the standard fund-to-fund model fees, or what does that typically look like?
Speaker 1:It's 1 in 10, usually as a market for fund of funds. So that's, that's usually what you see.
Speaker 2:Gotcha, you mentioned earlier the difficult problem that you're trying to solve in terms of where the seed funds are investing, where founders are going to how. How have you approached that and how are you working to solve it?
Speaker 1:Yeah, sounds good. So we we have a sort of a data angle. We have a team of seven data scientists and engineers. I think we built one of the more sophisticated intelligence engines in the market. We aggregate terabytes of data across many different kinds of data sets, so basically anything that touches venture whether it's private market transactions, people profiles, work experiences or GitHub or scientific journals, business filings, those kinds of things and what we try to really understand is unpack and reconstruct are the networks that are forming and evolving, you know, around the tech ecosystem.
Speaker 2:Yeah, sorry, sorry, please continue.
Speaker 1:That was it, and so we do that by. When it comes to networks typically it's networks of co-investors, networks of talent, founders, team members, you know those kinds of things and so we have algorithms that we've developed, as well as techniques to essentially give us some knowledge on where we should deploy capital. When it comes to GPs, For example, we have a set of models and sort of other supporting data to give us an early indicator as to this GP might outperform the rest of the market, and so, anyways, that's the kind of stuff that we do on the GP side. We also use this sort of intelligence to understand all of the companies, especially the companies in our portfolio. We use that to essentially give us sort of preempt potential co-investment opportunities, and then we also give back intelligence to the GPs themselves, the fund managers themselves, and that's usually in the form of intelligence, whether it's for sourcing intelligence or market intelligence or relationship intelligence, and so you can think of us like a technology platform focused on VC fund to funds investing.
Speaker 2:Super interesting. So I'd imagine you're probably connecting to the Crunchbase Pitchforks API, like you said, github, linkedin, scraping all these other sources, and so then aggregating all of that together and almost building like a map of what's happening where.
Speaker 1:Exactly, and we also have a lot of proprietary data sources as well, both ones that are unstructured, that we ingest as well as just because we meet with so many funds on a regular basis. We just have a lot of data. But yeah, you're right, the hardest part is tying all of this stuff together in a way that's coherent, which involves both resolving entities across datasets, which is a very difficult problem, but also just normalizing time series, returns data and a bunch of other sort of noisy stuff that needs to get resolved for us to actually have a dataset that we can use for large-scale modeling, and the modeling we do is actually based on networks. So that's also another problem that we've been solving, which is how do you build these large-scale machine learning systems on network-based data, because traditionally it's done on language data or spatial or images and things like that.
Speaker 2:Well, it's really interesting because traditionally pre-seed and seed funding is so almost feelings or intuition or conviction based right, because really what you can evaluate at that level is founder product market, but you're taking what is inherently kind of a low level data product or investment vehicle and then getting data out of it. So I'm just curious then what are some of the leading indicators that you have found when it comes to saying, okay, we're going to invest in this GP or we're not going to?
Speaker 1:Yeah, definitely. So. One thing to note about venture which is well understood, is that the underlying return distribution is power law right, so you have a very long tail of performance and sort of essentially, in the market, you know, very few companies actually return the whole market, and so it's important to be to index yourself in an area where you'll get access to to those companies Right, in order to drive returns, and if you do, then the returns are meaningful. That's sort of one thing. The other thing is that when you you know, when you choose a fund manager or a portfolio of funds, for example, each of those funds are selecting from that underlying return distribution, and so when you look at the performance, especially when it comes to like seed managers, you end up with a very similar long tail, and so it's essential, essentially, that you are able to select fund managers that are at the sort of the tail, you know, the right tail in order to have like outperformance, because it's very high volatility, right, because when you have an early manager, you know a lot of them really underperform, or most of them underperform, and then some really outperform, and so it's, you know, the whole methodology is to select a portfolio of really good small funds that have the potential for outlier performance, and if you do that well, then you'll have a really good performance as an overall basket. So that's just something that's important to understand in our time, and the point is that, first and foremost, fund strategy really matters.
Speaker 1:In venture, small funds outperform Historically they've always outperformed large funds, and it just makes sense statistically speaking, because if they have an outlier in their portfolio, it usually returns in multiples of the fund, and so that doesn't happen. As you scale AUM, you tend to have to go to either later stages or have multiple products, and so the small funds is really where we play and we try to get alpha there. Now, in terms of, like, the features that we use to select managers, I guess there's like three orthogonal vectors. I think one is network, so we have network-based algorithms that essentially what we do is we're reconstructing the networks of managers when it comes to their co-investor base or their talent base, et cetera, and we have essentially models, predictive models that are calibrated on historical performance, which, you know, essentially say this sort of structure, this sort of structure, this sort of embedding of an individual, can likely lead to outperformance with some confidence. That's sort of one piece of what we do.
Speaker 1:The other piece of what we do is more sector-based and thematic-based.
Speaker 1:We believe that in venture, there's heterogeneity in terms of where innovation is being created and where returns will be created and where the supply of later stage capital will be, and so we don't believe, generally speaking, in just being a generalist, and so we take a thematic bet in each fund cycle based on where we think the market will go, and so we care about which areas our fund managers are indexing against and whether they're specialists in those areas.
Speaker 1:To give you an example, like in our first fund in 2021, we invested in, you know, many deep tech firms which were focused on, you know, technologies, tackling industrials, software, hardware enabled solutions, but also defense, et cetera, and and there was a lot of reasons why we chose that but those ended up being like really great areas to invest in, because a lot of the capital now is going into those areas, versus, let's say, you know what is a traditionally FinTech or or just traditional vertical software, et cetera. So innovation does matter, and so thematic areas another thing we look at. And then the other thing is just is just the strategy firm strategy, you know, cause as as can the firm execute on what it's intending to execute, and is it drifting out of an initial strategy as they get larger, because the biggest danger for funds is them getting too large and the competitive dynamics change, and so we have a lot of work that we've done in that area as well.
Speaker 2:Hope that answers it does. Is there a sweet spot? So we know that smaller firms outperform larger firms. Is there a sweet spot in terms of firm size, like, are you looking as small as solo GPs? Is it up to a certain amount of AUM or a number of partners?
Speaker 1:Yeah, we tend to say less than 100 million, and there's really within that there's like a bifurcation, which are it really has to do with the game theory. You know the game mechanics, which is, you know the smaller funds can collaboratively invest in companies and they don't squeeze others out, and you get to a point where you have to lead or co-lead, and that requires more competition, more winning, and so your conviction level needs to be higher. So there's not an exact point, but it's somewhere between 40 and 50 or 55 where that happens.
Speaker 1:If you maintain the same portfolio construction strategy, which we like. Firms that have 25 to 35 companies, you know, essentially so that's kind of what happens in market, but there's no like there's no. You know you could have a firm that's larger, do very, very well, but if you want to see these really long tail kind of outcomes, then you need to have we believe you need to have a small fund.
Speaker 2:Are there particular sectors that you've seen outperforming in recent years?
Speaker 1:For sure, for sure. So what we look at is graduation rates. You know, it's sort of one thing because it's hard to know the outcomes, right, because the time is so long between you know, like the maturity of a company, its outcome. We can get some early indications from you. Know where the supply of capital is going and also the graduation rates of companies as well.
Speaker 2:And just to pause you, sorry. When you say graduation rates, are we referring to like? What number graduate from like C to series A to series B, et cetera?
Speaker 1:Yeah, typically we're looking at the bridge between C to A, which is a very meaningful bridge. Yep, there's a high failure rate especially, and also graduation rates in the markets also fluctuate. Of course, you know, in 2021 period everything was graduating and then right now it's very difficult to graduate and so a few companies do. But we sort of look at that and we index, you know, managers, sectors, et cetera, against that metric on an annualized basis to see how it's evolving. We think, like, where there's a lot of you know activity today is in this intersection of hardware, software-enabled solutions, tackling industrials and defense, which is these are traditionally businesses that a lot of VCs didn't want to touch because of the R&D intensity and capital intensity and whatnot, but these days, because of just a convergence of technologies that are maturing with diminishing cost curves, now you can tackle these problems. You know sort of very well and, additionally, there's just a lot of geopolitical and other sort of economic you know trends that are forcing us to focus on these things, whether it's like onshoring, offshoring, or whether it's the globalization, or or just generally, like you know, the political environment, um, or just war, generally speaking, so, uh, so anyway. So that's what happens is like you have these sort of macro, um, you know needs that need to get filled and um, so that's sort of one area.
Speaker 1:The other area we're seeing a lot is in in life sciences is the intersection of computation and biology applied to like therapeutics, um, orics or diagnostics and whatnot. That's an area where that's evolving because biology is becoming more of an engineering problem, right, and so that's that's a very attractive area. And then generally like data infrastructure, whether it's like AI data infrastructure, but just like how you develop, deploy software in like in the cloud and sort of multi-cloud and all that is like it's. It's always like evergreen, it's an evergreen thing. So that's what we're seeing. And then we're seeing some newer things, like in consumer, which is interesting, as well as like, I think, crypto is having like another, another resurgence, which is interesting.
Speaker 1:We're seeing a lot more now like kind of newer age crypto funds getting started um but um.
Speaker 2:But generally speaking, we're you know there's a lot happening is there still appetite for crypto, given all of the collapses of the last couple of years between Celsius, ftx and gosh? Who knows how many others?
Speaker 1:Yeah, I think the challenge in crypto is just the regulatory environment and what is the security and that. But there is a lot of you know we're seeing a lot of interest in, you know, sort of AI meets sort of crypto, like the centralized data, you know sort of data management, data sharing, those kinds of things, and then just the sort of the rethinking still the rethinking of, like, the financial infrastructure for enterprises using sort of distributed ledgers. We haven't really made a bet. We made one bet there, but yeah, we haven't been incredibly active there because we just don't see a lot of adoption in the areas that we think are important.
Speaker 2:What about on the data center side? You know Sam Altman's talking about how he needs to raise $7 trillion to run all of OpenAI's data centers, and you know, of course, that sounds like selling the shovels in the gold rush. Have you looked into that at all?
Speaker 1:I mean we tend to stay like so one of our core philosophies I think it's just generally a good rule in venture is we like to be in businesses that have, you know, sort of increasing feedback loops and positive, increasing positive returns and where you have like winner, take, go into like a larger system. You just get into this place where there's like diminishing returns and you get swapped out and so we just we don't feel like there's enough of a flywheel in some of these areas. So we're not going to be investing like in fiber optic, you know kind of equipment or those kinds of things, because you know we don't want to compete there. But but yeah, anyways, that's kind of. So we haven't looked at that area as much.
Speaker 1:I've looked at a lot of AI chips companies over time. I think it's a very grueling business to invest in If you can make it, especially when you're dealing with like the resources of like larger groups. So you know we we've stayed away from that area. We recently did do more of a quantum photonics company that we think is very exciting, which is more focused on quantum computing at scale. We did that recently.
Speaker 2:Interesting. Okay, so when we're talking about investing in funds, obviously we've talked about the technology approach. But technology aside, are you then also taking personal characteristics into account to these fund managers, or is it purely quant-driven? Or what does your overall decision-making process look like?
Speaker 1:Yeah, it's human. I mean, of course there's a lot of human qualitative work that goes into it. We think of the technology as just a guide. We use it like a funnel for sure, like sour outbound kind of activities and just initial qualification, but there's a lot of work that gets done downstream. That's very human driven. Uh, you know, I think with personality, nothing we do quantitatively, of course, you know we, we, we study the people a lot, uh, and there has to be a sort of a good chemistry and all that and we have to believe in all that. But but generally speaking, there's like a mixture of both things that come into play. And then there's a lot of like network checks, like we do a lot of what we call intelligent back channeling, like talking to the right people about the person to validate. You know, to validate both like the strategy as well as the capabilities. You know their ability to win, you know sort of their value add, those kinds of things. We just kind of we do that during the diligence process.
Speaker 2:On the, the size you mentioned I believe you said it was 25 to 35 was was the sweet spot. So does that imply then, as that's more than likely going to be an earlier stage fund like a fund one or fund two, is that is that? Am I on the right track in terms of what's appealing?
Speaker 1:yeah, we like young firms, so it's it typically funds one, two, three is what we focus on. You know, firms that are, like you know, relatively new five years, six years, something like that at most. There are situations where you know firms stay purposely small and they continue to deliver. You know there's no reason why that would not continue to be in our portfolio. But there are situations where firms get larger and more institutionalized and then you know, sort of our value add is no longer, you know, as relevant.
Speaker 2:If you're looking at. But if you're going so heavily off of data, wouldn't it make more sense to look at more experienced fund managers where you have more of a track record and more data to analyze?
Speaker 1:Yeah, our data, like our whole systems, are geared to finding early signal and seeing things that others don't see, and the nature of venture is twofold. One is that you have a high volatility of return dispersion, right, so it's. It's a hard problem, you have to select really well. But the other thing is is there's low, like relatively low, performance persistence, and so over time, sequentially, people, you know the returns aren't correlated and so it's not the case that for that firm stay good for forever, and in fact a lot of times fund two is very different than fund one performance for a lot of reasons, right.
Speaker 1:So that's sort of like that's our methodology, and it's true that, yeah, you have more information as the firms get more mature, but then also, like our alpha goes away and typically like the, since price and venture is not something that you can move, right, cause it's two and 20, no matter what, sure, and venture is not something that you can move right Because it's two and 20, no matter what, sure, what ends up happening is that what gets adjusted is fund size, and so a lot of managers, just you know, they're attracted by fees and they, you know, and believe in the belief that they can execute on that strategy, on a strategy that's larger, and we just rather not take that kind of bet sometimes.
Speaker 2:They're more focused on the two than they are on the 20.
Speaker 1:Exactly, exactly right.
Speaker 2:Exactly right, and you're also co-investing directly in companies as well. Along with the vesting and funds, how are you deciding on the allocation and capital between the two things?
Speaker 1:We keep them separate. In the next fund it'll be around 80-20. 80 in the fund of funds and 20 in the co-invest. We think it's a good strategy for LPs because, of course, the assumption is you're selecting well. But if you're selecting well on both, then you'll benefit from some earlier liquidity. On the co-invest side, because we're investing typically at Series B plus companies, Series B-ish and so like the hybrid strategy works really well. But yeah, so it's about 820, I would say.
Speaker 2:When you raise your next fund. I'm just curious, if you don't mind me asking, how much are you planning on that being? What does that look like, and what is the plan for the next fund look like?
Speaker 1:Yeah, it'll be around 230 across two vehicles, 230 million across two vehicles, and the strategy is similar. We don't think our strategy scales too much right. So there's a point at which you can't deploy that much capital into a small fund, and so the flagship product will always be sort of constrained. We're very intellectually honest about what we can deploy and we care a lot about our investor returns. So, anyways, that's the next vehicle that we're building, and it'll probably be 20 or so firms, 20 to 22, you know, funds in the fund of funds and about 10 to 12 companies in the Co-Invest site.
Speaker 2:How do you square the fact that you're growing beyond what your own models show the best returns at in terms of that 25 to 35 or 50 level AUM?
Speaker 1:Well, we're a fund of funds, so it's a different statistical model, right, because we're a portfolio of potential outliers, and so, you know, as long as we can get allocation into those potential outliers, right. So the constraint that we have is how much capital can we commit to a fund? That's a certain size, right, and so that breaks down right, like you can't put $15 million, and so that breaks down right, like, if you can't, you can't put $15 million into a $21 fund, right, it's just, it doesn't work. And so that's that's where the math breaks down. And so there's, there's a limit as to which, like, our strategy will, will work, um, you know, without having to compensate by investing in bigger funds.
Speaker 2:Gotcha Okay cool. Well, this was super interesting, much more, I would say, probably math and finance heavy, than the normal conversations we have, but I love this. I love this component of it and really grateful and excited for you to come on the show.
Speaker 1:Thank you so much for your time, Rhys, and we'll be in touch.
Speaker 3:Thanks for listening to See to Exit. If you enjoyed the episode, don't forget to subscribe and we'll see you next time.
