The Exchanges

Q of that built out discipline, we spoke about kind of on the latest data, the pre IPO, but very data centric firms that do price also very intelligently with data in terms of pricing. Does data at the maybe A or B stage where you guys really operate, does that really play into effect for you in terms of being able to kind of intelligently price these assets using data?

A Not really. Not, not at all. And then that's sort of the beauty, I think, of being in the early, early stage, right? In the series A and in the late C stage, even in the series B to some extent, you're, you're largely not going to make a lot of gains by being better on pricing. And that's because, you know, like these things don't get priced very precisely. You know, the, the, the sort of opposite end of the spectrum here is sort of like quantitative hedge funds. That really do a lot of data analysis, but the specific data that they stare at is market data, price data, and they use that. They try to price the asset at any given time very precisely to within very small margins, but that's not what we're trying to do at the earliest stage, and it's more our nature of the outcome distribution. The probability distributions that we're dealing with at the very earliest stage are such that if you're plus or -20, 30, even 40, 50%, it's not going to make or break the result. Either you're going to lose all your money, or you're going to make multiple times your money. There's not a lot in between.

AI assessment note: “Not really. Not, not at all.”

Q of that built out discipline, we spoke about kind of on the latest data, the pre IPO, but very data centric firms that do price also very intelligently with data in terms of pricing. Does data at the maybe A or B stage where you guys really operate, does that really play into effect for you in terms of being able to kind of intelligently price these assets using data?

A Not really. Not, not at all. And then that's sort of the beauty, I think, of being in the early, early stage, right? In the series A and in the late C stage, even in the series B to some extent, you're, you're largely not going to make a lot of gains by being better on pricing. And that's because, you know, like these things don't get priced very precisely. You know, the, the, the sort of opposite end of the spectrum here is sort of like quantitative hedge funds. That really do a lot of data analysis, but the specific data that they stare at is market data, price data, and they use that. They try to price the asset at any given time very precisely to within very small margins, but that's not what we're trying to do at the earliest stage, and it's more our nature of the outcome distribution. The probability distributions that we're dealing with at the very earliest stage are such that if you're plus or -20, 30, even 40, 50%, it's not going to make or break the result. Either you're going to lose all your money, or you're going to make multiple times your money. There's not a lot in between.

AI assessment note: “Not really. Not, not at all.”

Q the network, because it's about the only thing that I do have, in lack of idea. But, um, I do have to ask, if we move kind of one stage down the funnel to the evaluation phase, many suggest that maybe potential downfall of data in evaluation and the picking phase. But what are your thoughts on this, and kind of how data deals with the picking and evaluating phase?

A Yeah, this is really where we spend, you know, the vast bulk of our time Sort of, um, data science in evaluating. And I think that this is largely a reflection of sort of where we focus. I think, you know, evaluating means different things at different points, sort of in the life cycle of a company. When you're talking about sort of a seed company, very early stage companies that have no product or where the product is just, just hitting the market. I mean, there's not much data to look at, so it's sensible to not, to not really rely as much on data. There might be some data in the sense of like attempting to score, you know, sort of objectively score, uh, Founder profiles or attempting to score the, the founding team somehow attempting to score the market. And I think that's totally reasonable, but it's true that there's not a ton of data there. Right. Um, and then if you look at the other end of the scale, right, if you look at sort of, once you get to series C, series D, you know, at that point, right, the investor is really valuing the business. They're about, they're buying the business. They're definitely buying the founder, right? But I mean, if you think about the seed stage, right, they're really buying the founder when the investor invests. And when you think about the late stage, they're buying mostly a business, right? And then it's definitely some founder, but most…

AI assessment note: “evaluating means different things at different points sort of in the life cycle”

Q you look back at the companies that you've worked with, and some of them are different in many respects, when you look back at the companies you've worked with, are there commonalities in terms of where they maybe potentially struggle to find that core, unbiased truth, or struggle in certain respects with regards to kind of their growth and discovery of product market fit? Are there commonalities with their discovery?

A Yeah. One thing that is common is that early stage companies, they're never great in execution. And when I say that, what I mean is that, you know, when you look at Facebook in the late 2000, they were really remarkable execution wise. They were able to just do a lot of things very quickly and be, and be able to understand what was going on as they were doing it. And early stage teams, of course, they're not going to be at that level yet, but it's all about sort of striving to get to that next level, striving to execute at the next level and executing at the next level doesn't necessarily mean just doing more It means doing the right stuff and knowing that you did the right stuff. And it's that aspect of knowing that you did the right thing, which is often a bit lacking in these early stage companies. I think anecdote works really well, you know, when you're dealing with the first couple of customers, when you're dealing with only maybe a dozen users, maybe two dozen users. But once things start pushing into dozens of customers, possibly hundreds of users, maybe several hundreds of users, that's where like data is there to really check the anecdote, to make sure that people are really doing what you think they're doing. And that's really the place where we try to help provide that articulation.

AI assessment note: “that's where like data is there to really check the anecdote”

Q Can I ask, do you think that a lot of firms that are kind of positing themselves as data first Yes, in terms of sourcing, but actually kind of using LinkedIn Navigator a little bit intelligently. Do you think that is prominent, or do you think actually there is this new wave of fundamentally transformative sourcing techniques that many, many venture firms are applying to sourcing?

A So definitely, I would say there's one end of the category, which is, as you described, basically using LinkedIn and behaving like recruiters, right? And then in terms of really transformative, you know, there are some things out there that are interesting, I think. I don't know if it's, it's probably early right now to call them transformative purely in terms of sourcing. I think that's because You know, you really get down to like what it means to be good traditional venture capital. You know, if you talk about these great old firms, you know, you talk about benchmark, you think, talk about Sequoia. I mean, what is sourcing for them? Well, it's their brand and their network, right? But the traditional thing has to work. And then if you can layer on these data things, that's great. But if you start from a place where you don't have any of the traditional thing and you try to go with the machine alone, it might be helpful, but you know, you're going to be this massive disadvantage because fundamentally the early stage, you know, investing ecosystem is really well connected. So that networking thing works Pretty well. It's not really broken, per se.

AI assessment note: “it's probably early right now to call them transformative purely in terms of sourcing”

Q Can I ask, do you think that a lot of firms that are kind of positing themselves as data first Yes, in terms of sourcing, but actually kind of using LinkedIn Navigator a little bit intelligently. Do you think that is prominent, or do you think actually there is this new wave of fundamentally transformative sourcing techniques that many, many venture firms are applying to sourcing?

A So definitely, I would say there's one end of the category, which is, as you described, basically using LinkedIn and behaving like recruiters, right? And then in terms of really transformative, you know, there are some things out there that are interesting, I think. I don't know if it's, it's probably early right now to call them transformative purely in terms of sourcing. I think that's because You know, you really get down to like what it means to be good traditional venture capital. You know, if you talk about these great old firms, you know, you talk about benchmark, you think, talk about Sequoia. I mean, what is sourcing for them? Well, it's their brand and their network, right? But the traditional thing has to work. And then if you can layer on these data things, that's great. But if you start from a place where you don't have any of the traditional thing and you try to go with the machine alone, it might be helpful, but you know, you're going to be this massive disadvantage because fundamentally the early stage, you know, investing ecosystem is really well connected. So that networking thing works Pretty well. It's not really broken, per se.

AI assessment note: “probably early right now to call them transformative purely in terms of sourcing”

Q and seeing that in the data, once one does see that and kind of you want to physically want to move on it, it comes to the kind of elements of winning the deal, so to speak. How do you think about the data's ability and your process in terms of being able to win deals over, you know, multiple other term sheets that will often be present in market?

A For us at Tribe, we really think of data as just a form of truth, right? Like really our goal is to just find the truth and articulate it clearly. That's the goal. And data happens to be sort of the purest form of truth. Data is actually totally unbiased. When you infer something from data, then you do something biased, right? That's where you introduce bias. Um, so a lot of our work is around sort of just articulating very clearly. This is what we see, and this is how it sits in the world of things we've seen before. You know, over here, you're kind of top quintile. Over here, you're kind of bottom quintile, and be able to articulate many facets of product market fit through that lens and give it back to the founder. And for us, we find that founders are By and large, find that feedback extremely valuable, right? They understand that a lot of times founders are like, okay, you know, I had an intuition that this piece was strong, this piece was weak, but I didn't realize that this was median and that this was top quintile. I thought it was maybe both of them were median. It can sort of kick off a very high common context conversation between us and the founder. And it's really through that back and forth where we develop that relationship and understanding of what we think is important, what they think is important, what we think we can help with. Namely amplifying this type of…

AI assessment note: “we find that founders are By and large, find that feedback extremely valuable”

Q you look back at the companies that you've worked with, and some of them are different in many respects, when you look back at the companies you've worked with, are there commonalities in terms of where they maybe potentially struggle to find that core, unbiased truth, or struggle in certain respects with regards to kind of their growth and discovery of product market fit? Are there commonalities with their discovery?

A Yeah. One thing that is common is that early stage companies, they're never great in execution. And when I say that, what I mean is that, you know, when you look at Facebook in the late 2000, they were really remarkable execution wise. They were able to just do a lot of things very quickly and be, and be able to understand what was going on as they were doing it. And early stage teams, of course, they're not going to be at that level yet, but it's all about sort of striving to get to that next level, striving to execute at the next level and executing at the next level doesn't necessarily mean just doing more It means doing the right stuff and knowing that you did the right stuff. And it's that aspect of knowing that you did the right thing, which is often a bit lacking in these early stage companies. I think anecdote works really well, you know, when you're dealing with the first couple of customers, when you're dealing with only maybe a dozen users, maybe two dozen users. But once things start pushing into dozens of customers, possibly hundreds of users, maybe several hundreds of users, that's where like data is there to really check the anecdote, to make sure that people are really doing what you think they're doing. And that's really the place where we try to help provide that articulation.

AI assessment note: “One thing that is common is that early stage companies, they're never great in execution.”

Q venture as you do today with Tribe. And I wanted to start on the four key components of success in venture being kind of sourcing, evaluation, winning, managing, and I want to break them down one by one. So if we start on sourcing, Jonathan, tell me, how do you think about the ability for data to actively bring to the surface the best opportunities, maybe before others see them?

A Yeah, you know, we really think of sourcing as sort of a multi-edged prong. You know, I know that there exist firms out there that really rely a lot on data science on sourcing, and we did that type of work early on in social capital also. And, you know, we ended up finding, by and large, that there's no one silver bullet in sourcing, right? Sourcing is one of these things where good old ground game, just raw networking and And being out there is a really important piece of it. Building brand is an important piece of it and having an outbound strategy is an important piece of it. And then with regards to how you power your outbound strategy, you can do all sorts of things with data sets out there that can help you do that, but it's no one of them that sort of wins all the time. There are a bunch of public signals that you can get your hands on these days by effectively, you know, running algorithms on top of crunch base. You know, you acquire that data or possibly a better version of it. There exists panel data that you can get both sort of app usage panel as well as credit card panel data. And you can, you know, if you can, you can get possibly get a hold of LinkedIn or something like that. Talent data, which is also difficult to get. But if you manage to get your hands on some of these data sets, there are things you can do on top of them to help you find signals. But really,…

AI assessment note: “there are things you can do on top of them to help you find signals”

Q the network, because it's about the only thing that I do have, in lack of idea. But, um, I do have to ask, if we move kind of one stage down the funnel to the evaluation phase, many suggest that maybe potential downfall of data in evaluation and the picking phase. But what are your thoughts on this, and kind of how data deals with the picking and evaluating phase?

A Yeah, this is really where we spend, you know, the vast bulk of our time Sort of, um, data science in evaluating. And I think that this is largely a reflection of sort of where we focus. I think, you know, evaluating means different things at different points, sort of in the life cycle of a company. When you're talking about sort of a seed company, very early stage companies that have no product or where the product is just, just hitting the market. I mean, there's not much data to look at, so it's sensible to not, to not really rely as much on data. There might be some data in the sense of like attempting to score, you know, sort of objectively score, uh, Founder profiles or attempting to score the, the founding team somehow attempting to score the market. And I think that's totally reasonable, but it's true that there's not a ton of data there. Right. Um, and then if you look at the other end of the scale, right, if you look at sort of, once you get to series C, series D, you know, at that point, right, the investor is really valuing the business. They're about, they're buying the business. They're definitely buying the founder, right? But I mean, if you think about the seed stage, right, they're really buying the founder when the investor invests. And when you think about the late stage, they're buying mostly a business, right? And then it's definitely some founder, but most…

AI assessment note: “evaluating means different things at different points, sort of in the life cycle”

Q from doing this, and how's this defensible? Because to me, nothing's really defensible. In the world of Facebook, Amazon, Google, Apple, really, defensibility has died in the resources, talent, and energy that they can provide to any single problem they choose. How do you think about this when you think about N of one and one of N, and am I right, or am I too pessimistic in saying this?

A I think in the long term, you're absolutely right. Nothing is defensible in the long term, right? The question is like, are you in an era where you can defend it for a while and give yourself enough breathing room to possibly innovate something else, build something else that can give you an edge elsewhere? So, I mean, I think that's how we tend to think about that. You know, when we think about defensibility right now, the most obvious pattern that we've seen in the last 15 years now is really the concept of a network effect, right? The concept of a network effect seems to really be able to put in a chunk of defensibility that gives you Or at least medium term defensibility that's very hard to assail. And in some sense, marketplaces are a reflection of that, right? Marketplaces are sort of the situation where there is, in theory, some strong network effect on both sides that helps to propel the business forward. Other things, other than network effects, obviously a very strong brand can be something that's very hard to replicate, but it oftentimes takes longer. It's not so clear that you can build those things completely, um, methodically the way that you can sometimes build the marketplace. But then again, or network, network effects, excuse me. But then again, you know, even network effects Usually when they occur, it's not because you, you set out to engineer it. There are …

AI assessment note: “I think in the long term, you're absolutely right. Nothing is defensible”

Q I would, though, love to start today with a little bit on you. So not many of my guests have done string theory PhDs. So how did you go from that to the wonderful world of venture today?

A Yeah, sure thing. Yeah. When I was younger, I wanted to be a physicist. So I, you know, I went to Berkeley to study physics and then ended up doing my PhD PhD at Stanford studying black holes and cosmological inflation and string theory, and this was in the sort of early 2000. It's really sort of the most useless topic you can probably get your PhD in. Towards the end of my PhD, it was just clear I didn't want to be an academic, and in the, you know, this is in 2006, so back then everybody was trying to go to investment banks, but, you know, I wanted to go into technology somehow, and ended up landing at Microsoft to be a product manager in the web search group. This was in the very narrow amount of time between MSN, And live and Bing. Right now, it's called Bing. In the middle, it had a little bit of time where it was called Live Search, so I worked on Live Search for a little bit. While I was there, the Facebook platform opened up, and me and a couple of friends built one of the very first sort of social networking games directly on the Facebook platform. This was Superpoke. I don't know if you remember Superpoke from 2006.

AI assessment note: “Towards the end of my PhD, it was just clear I didn't want to be an academic”

Q I would, though, love to start today with a little bit on you. So not many of my guests have done string theory PhDs. So how did you go from that to the wonderful world of venture today?

A Yeah, sure thing. Yeah. When I was younger, I wanted to be a physicist. So I, you know, I went to Berkeley to study physics and then ended up doing my PhD PhD at Stanford studying black holes and cosmological inflation and string theory, and this was in the sort of early 2000. It's really sort of the most useless topic you can probably get your PhD in. Towards the end of my PhD, it was just clear I didn't want to be an academic, and in the, you know, this is in 2006, so back then everybody was trying to go to investment banks, but, you know, I wanted to go into technology somehow, and ended up landing at Microsoft to be a product manager in the web search group. This was in the very narrow amount of time between MSN, And live and Bing. Right now, it's called Bing. In the middle, it had a little bit of time where it was called Live Search, so I worked on Live Search for a little bit. While I was there, the Facebook platform opened up, and me and a couple of friends built one of the very first sort of social networking games directly on the Facebook platform. This was Superpoke. I don't know if you remember Superpoke from 2006.

AI assessment note: “Towards the end of my PhD, it was just clear I didn't want”

Q from doing this, and how's this defensible? Because to me, nothing's really defensible. In the world of Facebook, Amazon, Google, Apple, really, defensibility has died in the resources, talent, and energy that they can provide to any single problem they choose. How do you think about this when you think about N of one and one of N, and am I right, or am I too pessimistic in saying this?

A I think in the long term, you're absolutely right. Nothing is defensible in the long term, right? The question is like, are you in an era where you can defend it for a while and give yourself enough breathing room to possibly innovate something else, build something else that can give you an edge elsewhere? So, I mean, I think that's how we tend to think about that. You know, when we think about defensibility right now, the most obvious pattern that we've seen in the last 15 years now is really the concept of a network effect, right? The concept of a network effect seems to really be able to put in a chunk of defensibility that gives you Or at least medium term defensibility that's very hard to assail. And in some sense, marketplaces are a reflection of that, right? Marketplaces are sort of the situation where there is, in theory, some strong network effect on both sides that helps to propel the business forward. Other things, other than network effects, obviously a very strong brand can be something that's very hard to replicate, but it oftentimes takes longer. It's not so clear that you can build those things completely, um, methodically the way that you can sometimes build the marketplace. But then again, or network, network effects, excuse me. But then again, you know, even network effects Usually when they occur, it's not because you, you set out to engineer it. There are …

AI assessment note: “I think in the long term, you're absolutely right. Nothing is defensible in the long term”

Q venture as you do today with Tribe. And I wanted to start on the four key components of success in venture being kind of sourcing, evaluation, winning, managing, and I want to break them down one by one. So if we start on sourcing, Jonathan, tell me, how do you think about the ability for data to actively bring to the surface the best opportunities, maybe before others see them?

A Yeah, you know, we really think of sourcing as sort of a multi-edged prong. You know, I know that there exist firms out there that really rely a lot on data science on sourcing, and we did that type of work early on in social capital also. And, you know, we ended up finding, by and large, that there's no one silver bullet in sourcing, right? Sourcing is one of these things where good old ground game, just raw networking and And being out there is a really important piece of it. Building brand is an important piece of it and having an outbound strategy is an important piece of it. And then with regards to how you power your outbound strategy, you can do all sorts of things with data sets out there that can help you do that, but it's no one of them that sort of wins all the time. There are a bunch of public signals that you can get your hands on these days by effectively, you know, running algorithms on top of crunch base. You know, you acquire that data or possibly a better version of it. There exists panel data that you can get both sort of app usage panel as well as credit card panel data. And you can, you know, if you can, you can get possibly get a hold of LinkedIn or something like that. Talent data, which is also difficult to get. But if you manage to get your hands on some of these data sets, there are things you can do on top of them to help you find signals. But really,…

AI assessment note: “you can do all sorts of things with data sets out there that can help”

Q having kind of access to that extended data points, but I do have to ask Jonathan, then we spoke about kind of product market fit within the companies within the actual model that you guys pursue in terms of data usage and data application. Is there a product market fit phase where you feel in terms of insertion point with companies, there is a product market fit for your model?

A Oh, absolutely. And, you know, I don't think we don't do this stuff. You know, not every company needs the same thing in terms of where they're at. A lot of times we'll invest in a company and we'll give them sort of some viewpoint on, on where we think their data is. Should be what we think we could see in their data. And oftentimes we will go and just do the measurements to go in and do some of the analysis with the raw data ourselves, because, you know, the folks on our team, you know, we have like massive expertise in this area, so we can do it very quickly. Um, and it doesn't necessarily make sense for them to spend time on it because it's their, their priority. One at that point is usually continuing to build the product, but at some point becomes the right thing to do. I just want to think back to Slack and Carter, two companies that we've worked with, you know, in the last couple of years, you know, both of them didn't really start investing in this Full bore, you know, well into about the series C, if I remember correctly. Um, you know, they may have maybe had one or two folks beforehand, but they didn't really, but those one or two folks tended to be sort of jack of all trades all across the company in terms of being able to do the data, but being able to do many other things. And they didn't really start building out, building out as a discipline until a little bit l…

AI assessment note: “both of them didn't really start investing in this Full bore, you know, well into about the series C”

Q Absolutely love that kind of entrance intervention that I do have to also, and many of our mutual friends said I should ask you about this. And it was you enjoyed an incredible inhale time at Facebook. How did you create the data science practice at Facebook? And maybe what were the biggest takeaways for you from that experience?

A Yeah, well, I wasn't the first one for sure, right? There was a data science team. I actually originally was brought into the marketing team. Because they needed data help. Part of the, part of what was going on at Facebook in that era was that it was actually very hard to just count things, right? Like doing data science was actually difficult because the technology was hard back then. I were sort of relatively early in their life cycles. And so there were only a few people who were sort of doing it every day. And a lot of teams were not getting the amount of support they wanted. And so there ended up existing a bunch of these little teams and many other groups such as marketing, you know, myself. And then over time, you know, we ended up pulling those teams together. We created the data science and analytics team. Sort of one full bore team sort of in 2011, I think. And then that team was sort of the core thing that, that grew into the ginormous data science and analytics organization at Facebook today.

AI assessment note: “we ended up pulling those teams together. We created the data science and analytics team.”

Q Absolutely love that kind of entrance intervention that I do have to also, and many of our mutual friends said I should ask you about this. And it was you enjoyed an incredible inhale time at Facebook. How did you create the data science practice at Facebook? And maybe what were the biggest takeaways for you from that experience?

A Yeah, well, I wasn't the first one for sure, right? There was a data science team. I actually originally was brought into the marketing team. Because they needed data help. Part of the, part of what was going on at Facebook in that era was that it was actually very hard to just count things, right? Like doing data science was actually difficult because the technology was hard back then. I were sort of relatively early in their life cycles. And so there were only a few people who were sort of doing it every day. And a lot of teams were not getting the amount of support they wanted. And so there ended up existing a bunch of these little teams and many other groups such as marketing, you know, myself. And then over time, you know, we ended up pulling those teams together. We created the data science and analytics team. Sort of one full bore team sort of in 2011, I think. And then that team was sort of the core thing that, that grew into the ginormous data science and analytics organization at Facebook today.

AI assessment note: “we ended up pulling those teams together. We created the data science and analytics team”

Q that, I mean, obviously that's kind of Connoted to ownership over time. In terms of maintaining that ownership, how do you guys think about reserves? So often it's kind of, you know, the best investments are rarely made on one check. So how do you think about intelligently allocating and concentrating capital for reserves? And does one use data as kind of one of the primary tools in this case?

A Yeah, I mean, for us, you know, we continually do this type of data work. And of course, the interesting thing there is that in many ways it becomes more relevant later on. It's not that it's more relevant, it's that it becomes more the only thing. When you're in the earliest stage, Data is one of many things. You know, the interesting thing about a series A is that there is no one most important thing. There are many, many important things. And then the further and further you get along, the more and more that data becomes kind of the most important thing. And so, you know, we continue to do that work and we, we do that work both for our own reserve allocation, but also for our work with other co-investors. You know, we don't necessarily think that we're the best growth investors out there. So part of our job is to help our portfolios get the best possible growth investors in their later rounds. So even if that means that we have to give up some ownership for it, because, you know, we're there to ensure the long-term success of the company. Not necessarily defend our ownership all the way up.

AI assessment note: “we do that work both for our own reserve allocation, but also for our”

Q having kind of access to that extended data points, but I do have to ask Jonathan, then we spoke about kind of product market fit within the companies within the actual model that you guys pursue in terms of data usage and data application. Is there a product market fit phase where you feel in terms of insertion point with companies, there is a product market fit for your model?

A Oh, absolutely. And, you know, I don't think we don't do this stuff. You know, not every company needs the same thing in terms of where they're at. A lot of times we'll invest in a company and we'll give them sort of some viewpoint on, on where we think their data is. Should be what we think we could see in their data. And oftentimes we will go and just do the measurements to go in and do some of the analysis with the raw data ourselves, because, you know, the folks on our team, you know, we have like massive expertise in this area, so we can do it very quickly. Um, and it doesn't necessarily make sense for them to spend time on it because it's their, their priority. One at that point is usually continuing to build the product, but at some point becomes the right thing to do. I just want to think back to Slack and Carter, two companies that we've worked with, you know, in the last couple of years, you know, both of them didn't really start investing in this Full bore, you know, well into about the series C, if I remember correctly. Um, you know, they may have maybe had one or two folks beforehand, but they didn't really, but those one or two folks tended to be sort of jack of all trades all across the company in terms of being able to do the data, but being able to do many other things. And they didn't really start building out, building out as a discipline until a little bit l…

AI assessment note: “both of them didn't really start investing in this Full bore... well into about the series C”

Q It is quite a binary one. Taking this to the next step for me as a venture nerd, I love portfolio construction, probably one of the many reasons I'm still single, Jonathan. But I do have to ask, with this approach, is there a right portfolio size, and how do you think about kind of analysis of optimal portfolio size?

A Well, I don't think there is necessarily a right portfolio size, because the thing is that the sort of outcome distribution of the assets you buy, you're in control of that to some extent. One of the things that we have found through using data for many years in the context of venture now is that We believe that historical loss ratios are too high in venture to some extent. So historical loss ratios for venture at large tends to be in this 30 to 40% range, which on the one hand is good because, you know, we as investors, sort of, we as a technology sector are exploring many ideas, which I think is fantastic. But if you think about it from the point of view of an investor, it's like, well, that's kind of high. So using, you know, data is one of these things that can help you sort of mitigate this type of thing. We don't optimize for this. You know, we don't optimize for downside protection. But it is one of those things that sort of naturally comes out of the byproduct of investing this way. And so this is an example where you're actually manipulating part of the outcome distribution. Well, maybe not manipulating, but you're selecting for certain characteristics of the outcome distribution of the assets that you're buying. And if you do that, if you're changing the outcome distribution of the assets that you're buying by selecting them differently, then how you construct the por…

AI assessment note: “Well, I don't think there is necessarily a right portfolio size, because the thing”

Q and seeing that in the data, once one does see that and kind of you want to physically want to move on it, it comes to the kind of elements of winning the deal, so to speak. How do you think about the data's ability and your process in terms of being able to win deals over, you know, multiple other term sheets that will often be present in market?

A For us at Tribe, we really think of data as just a form of truth, right? Like really our goal is to just find the truth and articulate it clearly. That's the goal. And data happens to be sort of the purest form of truth. Data is actually totally unbiased. When you infer something from data, then you do something biased, right? That's where you introduce bias. Um, so a lot of our work is around sort of just articulating very clearly. This is what we see, and this is how it sits in the world of things we've seen before. You know, over here, you're kind of top quintile. Over here, you're kind of bottom quintile, and be able to articulate many facets of product market fit through that lens and give it back to the founder. And for us, we find that founders are By and large, find that feedback extremely valuable, right? They understand that a lot of times founders are like, okay, you know, I had an intuition that this piece was strong, this piece was weak, but I didn't realize that this was median and that this was top quintile. I thought it was maybe both of them were median. It can sort of kick off a very high common context conversation between us and the founder. And it's really through that back and forth where we develop that relationship and understanding of what we think is important, what they think is important, what we think we can help with. Namely amplifying this type of…

AI assessment note: “winning is sort of that last step. It's where we've developed that context”

Q It is quite a binary one. Taking this to the next step for me as a venture nerd, I love portfolio construction, probably one of the many reasons I'm still single, Jonathan. But I do have to ask, with this approach, is there a right portfolio size, and how do you think about kind of analysis of optimal portfolio size?

A Well, I don't think there is necessarily a right portfolio size, because the thing is that the sort of outcome distribution of the assets you buy, you're in control of that to some extent. One of the things that we have found through using data for many years in the context of venture now is that We believe that historical loss ratios are too high in venture to some extent. So historical loss ratios for venture at large tends to be in this 30 to 40% range, which on the one hand is good because, you know, we as investors, sort of, we as a technology sector are exploring many ideas, which I think is fantastic. But if you think about it from the point of view of an investor, it's like, well, that's kind of high. So using, you know, data is one of these things that can help you sort of mitigate this type of thing. We don't optimize for this. You know, we don't optimize for downside protection. But it is one of those things that sort of naturally comes out of the byproduct of investing this way. And so this is an example where you're actually manipulating part of the outcome distribution. Well, maybe not manipulating, but you're selecting for certain characteristics of the outcome distribution of the assets that you're buying. And if you do that, if you're changing the outcome distribution of the assets that you're buying by selecting them differently, then how you construct the por…

AI assessment note: “I don't think there is necessarily a right portfolio size”

Q that, I mean, obviously that's kind of Connoted to ownership over time. In terms of maintaining that ownership, how do you guys think about reserves? So often it's kind of, you know, the best investments are rarely made on one check. So how do you think about intelligently allocating and concentrating capital for reserves? And does one use data as kind of one of the primary tools in this case?

A Yeah, I mean, for us, you know, we continually do this type of data work. And of course, the interesting thing there is that in many ways it becomes more relevant later on. It's not that it's more relevant, it's that it becomes more the only thing. When you're in the earliest stage, Data is one of many things. You know, the interesting thing about a series A is that there is no one most important thing. There are many, many important things. And then the further and further you get along, the more and more that data becomes kind of the most important thing. And so, you know, we continue to do that work and we, we do that work both for our own reserve allocation, but also for our work with other co-investors. You know, we don't necessarily think that we're the best growth investors out there. So part of our job is to help our portfolios get the best possible growth investors in their later rounds. So even if that means that we have to give up some ownership for it, because, you know, we're there to ensure the long-term success of the company. Not necessarily defend our ownership all the way up.

AI assessment note: “we do that work both for our own reserve allocation”