VCP 100: Nominate the Top Utah Founders VCs Must Watch! Nominate Now

Jon Bradshaw & Peter Harris

Jon asks Peter everything about data science. What data sources are they currently using? And how much will it disrupt the functioning of Venture Capital firms.

How Data Science is Improving Venture Capital

How is Data Science improving Venture Capital?

Jon asks Peter everything about data science. What data sources are they currently using? And how much will it disrupt the functioning of Venture Capital firms.

Some Lies/ Points  Covered in This Episode Include:

  1. Would you say data science is changing things?
  2. What Data Sources are University Growth Fund are using at this point?
  3. For seed and early-stage firms, when data is almost non existent, will it be a barrier
  4. A venture capital deal pipeline has three key elements: sourcing, benchmarking and value-add. Can Data science help on all three fronts?
  5. What are some of the other challenges you foresee?
  6. It’s a very intuitive field, do you think backing it with numbers and data will give VCs clarity while making decisions or make them second guess their intuition on                                                things that really matter- word of mouth, integrity, past personal or peer experiences.Things that data can’t support.
  7. Often, data from sources such as Twitter, LinkedIn, Pitchbook, Crunchbase, and AngelList are obtained and then pooled and organized. The organization and manipulation of third-party data can be time and labor-intensive. Pooled third-party data that is improved and arranged in a customized manner can eventually become proprietary in nature. Your thoughts?
  8. Once VC ventures become data backed, they will need to hire or re-organise teams that can collate and work with such data. How soon do you see that shift coming

PS-  it may call for a different talent sourcing model and organizational structure, with resulting implications for the structuring of compensation and incentives.

  1. For example, the venture capital firm Social Capital has built an automated system to invest in startups without meeting them. Companies upload data about themselves, and if the firm’s algorithms score the companies well, the firm backs them with an investment. The process was designed to keep bias from entering the equation. By mid-2018, the firm had assessed over 5,000 startups and invested in 60. Most of the investments were in companies based outside the major venture capital markets of the Bay Area and New York, and many were based overseas. About 80% of the companies featured non-white founders and 30% featured female founders. Do you see them as outliers or can that be the upcoming trend?

Let us know your thoughts on data science/ AI changing the Venture capital? What should we talk about next? Give us a follow and leave us feedback.

Hosted By

Episode Transcript

Jon: All right. So today, Peter, let's talk about how is data changing venture capital?
 
Jon: And the first question is, is is it or is it not changing the way venture capital works today, 2022?
 
Peter: When you say data, what do you mean?
 
Jon: Data science.
 
Peter: Okay. So like using data.
 
Jon: To make investment decisions as opposed to be like. Like, so for example, I don't know if this is true or not, but some VCs claim they can get access to credit card information from larger later stage companies to an idea of where someone's really at. You could go to comScore, maybe different metrics online. There are things like Pitchbook.
 
Jon: Ultimately, you're trying to figure out like.
 
Peter: Trying to get an edge.
 
Jon: An edge. Some have talked about and I'm going braindead, but there's different like DNA data analytics platforms, and it could be the VC says, just authenticate, share your data. And we're making a decision.
 
Peter: So like, you know, like data and data analytics, AI, machine learning, like those types of technologies are touching every facet of life, right? Period. So is it impacting venture capital? Of course it is. That said, there was a little bit of like, hey, there's going to be these funds that are going to come out. They're going to use like pure data analytics, and they're going to totally disrupt all VCs because they're going to make better decisions with their algorithms and stuff.
 
Peter: And like, frankly, we just haven't seen that hasn't happened yet.
 
Jon: Who's tried doing that so far?
 
Peter: So I'm really good friends with the folks over at Correlation Ventures. I don't think that they would say like, we're trying to disrupt all venture funding out there. In fact, they would argue that they they're very much not that's not the case because they co-invest alongside other venture funds, but they have an algorithm. Right. And they they gather a bunch of data points.
 
Peter: They put them into the machine, and then the machine pumps out a yes no investment decision and they follow through with it. Right. So you've got you've got funds like that to that level, right. Where they it's really just completely reliant on the algorithm and then you've got other funds which are pure old school and don't have any sort of like real technology.
 
Peter: I think the reality is that the majority are kind of in the middle and they have some sort of technology data solution that helps them source deals that might help them with hiring, that might help them with vetting, deals that might help them with supporting their portfolio companies over time with like new hires and so forth. So, and some funds have all of that, some funds have just certain pieces of it.
 
Peter: but you know, at a minimum, like a lot of fun, pretty much all funds use a CRM. So like you could consider that like a data, but.
 
Jon: That's not true data science just because you have data like.
 
Peter: That's true. It's not true data science, but it is a data play, right? It's you're tapping into Pittsburgh, you're using CRM products like Affinity that shows you like all your different connections and the different companies.
 
Jon: What data sources is the university Growth Fund currently using?
 
Peter: So our approach is I mean, we use Pitchbook and then we use a bunch of like online data, other online databases and sources. But you know, we're not an algorithmic driven fund for us. You know, I think one of the reasons why data analytics hasn't like come in and had a massive, massive impact on venture capital is because at the earliest stages it's it's art and science.
 
Peter: Right. And certainly like the science piece of it is of like, hey, does this like idea have merit? Because it's a really big market. Could you do some, you know, data science around that? Yeah, potentially, probably. But like the more intangible things of like building a relationship with the entrepreneur, vetting the entrepreneur, but then also convincing the entrepreneur to work with you, right?
 
Peter: Like, those are all like two way streets, right? Those are all things that are more qualitative than they are quantitative. And those are really hard for like a machine to replace somebody. And so ultimately, like the best funds, in my opinion, are going to be the ones that have like this really nice marriage of like great investors that are highly interpersonal, that understand how to vet things qualitatively, but then also are leveraging data science to their advantage when it comes to sourcing and supporting and and vetting deals to give them, you know, more leverage, if you will.
 
Jon: Okay. How how is data science changing the way you source deals or like your how your friends so like let's look at the different the different stages because we have sourcing and benchmarking.
 
Peter: You know so I'll give you an example. Like I just flipped through a pitch deck for a venture fund recently and they're spending, you know, a lot of money annually to identify great talent. And that's that's their whole approach, right? They're like, we're going to find the very best entrepreneurial talent out there. And we've got this whole, you know, algorithm and data set and we plow tons of money in every year to make it as good as it possibly can to help us identify like the next great entrepreneurs.
 
Peter: Right. And then once we know who we are, who they are, we spend a ton of time focusing on them, supporting them, and kind of winning them over to us to be their capital provider. Right. So that's an example of how, you know, one firm is trying to do it and there are more nuances to what they're doing.
 
Peter: But I mean, at its core, that's kind of it. You've got other funds that are saying, hey, no, like we're going to identify like what are the early indicators that a company's about to take off, Right? And so we spend a lot of time analyzing things like, you know, employee data from LinkedIn and website hits like from score, maybe they're pulling in credit card information to see like where people are spending money, you know, all those types of things.
 
Peter: And they're pulling all that and feeding the algorithm and then spinning out like, Hey, this company is really interesting. It's about to take off. You should go get in front of it. So, you know, they're just different, different approaches in that about.
 
Jon: What's the next frontier of data science and venture capital.
 
Peter: So look, I think the the big thing is that like adoption is not super widespread yet. And I think over time, funds are going to need to have an increasing amount of data science in terms of how they're sourcing and underwriting and those that don't are likely to get left behind, especially as they move into later stage investing.
 
Peter: I think maybe another place that would be really interesting is, you know, can data science say something meaningful about valuation? Because if you think about like one of the challenges that everything's going through right now is that like there's this there's this disconnect between public valuations and public and private valuations. And to me, what that saying is that, like we're not very we're still not very good at like valuing things.
 
Peter: And maybe there's an opportunity there for data science and AI to, to play a role in terms of saying like, hey, this is the right valuation or an appropriate valuation given the risk for this type of company, given all the different data factors that can play. Because I don't think that on the private side that that is playing as strong a role as it does already exist on the public side.
 
Jon: Do you think data science is going to destroy like the intuitive gut feel that VCs have, or do you think it'll add or correct that as time goes on?
 
Peter: I think it can add to it. I don't think it'll it'll completely destroy it because again, I think I don't know, you know, data science and AI engineers and experts would disagree with me and say, no, I can do anything that humans can do better. but I think there is something that's hard to replicate in terms of like building relationships because part of it is like, are you a good entrepreneur?
 
Peter: And like evaluating that through interactions. But there's also like this interpersonal connection that has to occur, right? Like when, when an entrepreneur in a V.C. work together, the VC funds the entrepreneur. Like there's a certain level of trust and partnership that needs to occur there for for the company to ultimately be successful. And at the end of the day, companies are still, for the most part, run by human beings, right?
 
Peter: That's not going to change anytime soon. And they're going to want to work with other human beings, especially when things are unknown, risky, unreliable. Right. Like, they're going to want somebody that can empathize with them and work with them. When VCs get get hit with the most like criticism, it's when VCs act like robots, right? When they're not understanding, when they're not empathetic.
 
Peter: Right. That's when VCs get the most criticism. And alternatively, that's when also VCs get the most compliments is when a VC like steps up for the entrepreneur and offends them or her. They back them. When things are tough, right? They spend their time supporting and helping them, right? That's what establishes like great VCs and their reputations and a lot of cases, not just when things are going really well and I don't know what are the days, but I don't think that, like.
 
Jon: The data could tell you.
 
Peter: Software is ever going to like, Hey, Spencer.
 
Jon: It could say, Hey, spend more time with John because something happening. Or it could say like in those stories could still happen, just data science, right?
 
Peter: And that would help them do it better, but it wouldn't replace them.
 
Jon: Yeah.
 
Peter: Right. Can I beat the very best chess players? Yes, but, you know, it beats the AI, the very best chess player, plus A.I.. And I think that's going to be the same for investors across the board. Regardless of his venture hedge funds, real estate that, you know, whatever, like humans leveraging AI and data science are going to be able to outperform just data science or just AI or just human beings.
 
Jon: I think that the one place where AI and data science will have a difficult time will be when new data sets and unknown patterns are presented to the AI because it would need enough time to learn and to map. And a lot of this A.I. modeling, when you look at like games, they're just doing like a trillion repetitive iterations to come up and eventually it'll do something and a much better outcome will come forward.
 
Jon: And I think one challenge that it may or may not be in the space with AI and data science is that my next question is, is as data science becomes a much bigger thing, will smaller funds have a much harder time competing with funds like BlackRock that have infinitely larger budgets to spend on data? Scientists and on data science?
 
Peter: Maybe. But, you know, I like smaller funds. I always start doing, for the most part, starting with seed stage companies. Right? And those seed stage companies, your earlier point like they don't the data set doesn't always exist. Imagine you're you're you're trying to use a AI to determine if Airbnb is going to be successful at the seed stage.
 
Jon: Okay.
 
Peter: Right. Like good luck because all the data says absolutely not.
 
Jon: But at a certain.
 
Peter: Point they're like massive liability issues. Nobody's doing it zero like market, right? Nobody's willing to pay for it.
 
Jon: But if data was ubiquitous of all startups, it would start picking up on these are emerging leaders.
 
Peter: Yeah, over time. But you guys actually have to start like developing data before you can actually point back to that, right? So like, yeah, sure. By the time Airbnb becomes like a series B company like Sur, some software company, you know, some software could figure it out, like, Hey, this company's really growing fast, you should take a look at it.
 
Peter: But by then, kind of everybody knew that already. So, you know, I don't think it's particularly helpful then. And to your earlier point, like we'll see will smaller funds struggle? I think really small funds that are focused on the seed stage. I think ultimately they'll probably be okay. I think larger funds, if they are not deploying some sort of data strategy.
 
Peter: Yeah, I do think that they'll probably struggle against those that don't.
 
Jon: Ask what's the next big data strategy play you'll get, you'll be making or is there one on the on the horizon that you think about?
 
Peter: You know, our our approach is a little bit different because we're very much an education focused fund and we're also core investor. So we're co-investing alongside other top tier venture and private equity funds. And to a certain extent we're leveraging what they're doing around like data science, to identify deals and support companies and so forth. So, you know, for us we're a small fund and a lot of our resources are focused on educating students.
 
Peter: But that said, like I'm always thinking about these types of things and trying to figure out like, where do we, you know, where do we invest to get an edge? And sometimes that's, you know, through things like our Sierra, through databases like Pitchbook, and then using that data to to make better informed decisions.
 
Jon: Okay. I think maybe as one last question, is there is a venture capital firm, social capital, Chamath led it and their big player, Angel, was we are going to make very data driven decisions. So I think over 5000 startups, the firm assessed over 5000 startups, but only invested in 60. And most of these companies were based out of typical markets of the Bay Area outside of New York City.
 
Jon: They were typically featured by nonwhite founders, things like that. Sure. What is your take on on their approach to this?
 
Peter: Yeah, I think I think it's a great approach, I think but I think it's worked out fairly well with them now. There's been a lot of like controversy at that firm, but that's probably driven most by Chamath. But no, I think that approach makes makes a lot of sense. And I don't think that they're alone. I think they're probably just one of the more vocal about it.
 
Peter: Right.
 
Jon: Okay.
 
Peter: And, you know, look, it's fantastic that they're they're backing more diverse founders across the board, both, you know, by race and geography and so forth, because you could easily see it going the opposite direction. Right. Which is like, well, if I look at all the companies been really successful, like they've all been founded by, you know, predominantly white men, right?
 
Peter: So let's go back more white men, because that's what the data says. Right. And I think you have to be kind of intentional about building your your algorithms and your AI to not fall into those those traps. So, you know, kudos to them for for building something that that gives them that can really like strip out some of those biases.
 
Jon: Yeah, I'm really interested to see what the future of data science, what happens when it comes to venture capital investment decisions right now with all my recent conversations, I feel like data science isn't really like I feel like at least at the first age, most of it's intuition, your gut feel. What do I currently know are using the market?
 
Jon: Yeah.
 
Peter: Do you feel like as an entrepreneur that it would be to your advantage or disadvantage if funds used a more data driven approach to sourcing or making investment decisions?
 
Jon: I don't. I mean, I don't know. I think like.
 
Peter: Imagine you just.
 
Jon: Like I think if.
 
Peter: You just submit like an application form and then it makes a decision whether or not to fund your, your company and gives you a valuation. Will that be more interesting than like building a relationship with a VC and kind of hashing it back and forth?
 
Jon: I, I don't know. So I'm trying to look for something really fast. You're now you're taking me off as the answer is, I don't know. I think I like the idea of a combo. I think I like the idea that I could come and hit you with real numbers that you might think is interesting. And then it's been validated.
 
Jon: But I don't know if that's that hard. I don't know. It's a good question. I think to consider this an additive X where you just come through when you connect your data sources, and then from that that's when they decide, do we talk to you or not? And that's kind of the and it's the data that's the gatekeeper.
 
Jon: And if you pass the initial thing, hey, we're looking for a certain amount of growth, We're looking for certain, you know, whether you're a freemium model or non freemium model, because you can plug in because we used to use bare metrics at tiny talk. Yeah. And it was amazing because I always felt like I was I was going blind, like I saw numbers going up.
 
Jon: But what do the numbers actually mean? What's the lifetime value of a customer? How much is it going to cost me to hire an accounting firm to actually give me this number? And is it changing in real time? Yeah. And Josh, paying for it from their metrics came through and had better metrics be I just plug stripe in and it was phenomenal.
 
Jon: I feel like I had a lot of clarity and I think if I was a VC and I could say, Hey, plug in to one of these data sources and then we'll decide if we go to the next step. I think a lot.
 
Peter: Of us do that already, right? So like we just looked at a deal in the ecommerce space and I was like, Yeah, I want to see your whole Shopify metrics, right?
 
Jon: But you're not plugging. Yeah, but you're, you're asking, you were asking for access to their data, not plugging in to having them plug into your data model like that.
 
Peter: That's true. But ultimately it's, it's not a huge difference, right? This is like show me the data and whether I'm going to run through an algorithm in my head or I'm going to have it programmatically, like set up, you know, yeah, some combination or something along the spectrum. Ultimately, it's going to come to a similar conclusion, right?
 
Jon: Yeah. So I think I don't know. I think I'm fine with it.
 
Peter: All right. You're right. With like an AI sitting on your board making decisions on whether or not that's.
 
Jon: An interesting question, and I've never heard that before.
 
Peter: Maybe are they going to let you hire Are they going to let you issue more, more stock options.
 
Jon: And a board member? That'd be interesting.
 
Peter: Yeah. It's something to think of.
 
Jon: That's also an interesting business idea.
 
Peter: There you go. A board member.
 
Jon: All right. Well, thanks for watching, everyone. This is the venture capital podcast. Go to venture capital firm if you want more resources or things like that. Thanks for joining us, Peter.
 
Peter: Yeah. Tell us in the comments where you think data science is going to take venture capital.
 
Jon: And if you're on YouTube like or comment below about additional questions you'd like us to cover. Thanks, guys.
 
Peter: Thanks.