The Exchanges

Q Speaking of scalable models now, I'd love to start On the theme of network effects, a term that very commonly bandied around today, but you've said before that many people maybe get them wrong, and they're maybe misunderstood, so I'd love to start first on a bit of nomenclature, and how do you truly define network effects?

A Yeah, so, I mean, there's a really well-accepted definition of network effects. I think, put it simply, the value of a product or a service increases as more users use it. So, All it means is if you have 10 users today and if 20 users tomorrow, having 10 extra users helps the existing 10 users. That is, they get incremental value from the product. So, you know, it's very simply defined, but I think the common confusion arises from the fact that people get confused with systems, nodes, technology, platforms, you know, and I think they forget that it is about measuring the incremental value to the existing user. As more users join it, and if that doesn't exist, then there is no network effect.

AI assessment note: “the value of a product or a service increases as more users use it”

Q Can I ask, what are the true signs that founders fundamentally get the underlying mechanics of network effects? And is there any examples where you've really seen this and you've gone, wow, this founder really gets

A Yeah, I think that, you know, they may not truly understand the technical terms behind network effect, but some founders intuitively pay attention to the right elements that are required for a network effect. So what is that? So number one is retention. I think Facebook actually did this really, really well, and I applaud the founding team to have paid attention that early on these metrics. So very simply, you They first launched in Harvard, and there was a lot of craze about the app. A lot of schools demanded that Facebook be available in those schools, yet they did not launch it for almost 18 months, and the simple reason being, the one metric they measured was how many of the Harvard students are using Signed into Facebook, and how many of them use it daily? So they had a goal that they wanted at least 80% of Harvard on Facebook, and more than 60 to 70% to use it daily. And they didn't want to roll it out to the second university until that happened. That was really smart, because A, they were trying to figure out what is the true utility of the product? How do I incorporate features to make sure there's strong engagement? Because engagement is a sign of network effect, not growth. This is where the confusion comes. Most people think every fast growing has a network effect. It doesn't. The retention is a strong sign of network effect. And so, having that level of attention t…

AI assessment note: “The retention is a strong sign of network effect.”

Q Speaking of scalable models now, I'd love to start On the theme of network effects, a term that very commonly bandied around today, but you've said before that many people maybe get them wrong, and they're maybe misunderstood, so I'd love to start first on a bit of nomenclature, and how do you truly define network effects?

A Yeah, so, I mean, there's a really well-accepted definition of network effects. I think, put it simply, the value of a product or a service increases as more users use it. So, All it means is if you have 10 users today and if 20 users tomorrow, having 10 extra users helps the existing 10 users. That is, they get incremental value from the product. So, you know, it's very simply defined, but I think the common confusion arises from the fact that people get confused with systems, nodes, technology, platforms, you know, and I think they forget that it is about measuring the incremental value to the existing user. As more users join it, and if that doesn't exist, then there is no network effect.

AI assessment note: “the value of a product or a service increases as more users use it”

Q Can I ask, what are the true signs that founders fundamentally get the underlying mechanics of network effects? And is there any examples where you've really seen this and you've gone, wow, this founder really gets

A Yeah, I think that, you know, they may not truly understand the technical terms behind network effect, but some founders intuitively pay attention to the right elements that are required for a network effect. So what is that? So number one is retention. I think Facebook actually did this really, really well, and I applaud the founding team to have paid attention that early on these metrics. So very simply, you They first launched in Harvard, and there was a lot of craze about the app. A lot of schools demanded that Facebook be available in those schools, yet they did not launch it for almost 18 months, and the simple reason being, the one metric they measured was how many of the Harvard students are using Signed into Facebook, and how many of them use it daily? So they had a goal that they wanted at least 80% of Harvard on Facebook, and more than 60 to 70% to use it daily. And they didn't want to roll it out to the second university until that happened. That was really smart, because A, they were trying to figure out what is the true utility of the product? How do I incorporate features to make sure there's strong engagement? Because engagement is a sign of network effect, not growth. This is where the confusion comes. Most people think every fast growing has a network effect. It doesn't. The retention is a strong sign of network effect. And so, having that level of attention t…

AI assessment note: “So number one is retention. I think Facebook actually did this really, really well”

Q I'm intrigued. Where do you see most scaling startups go wrong in establishing their growth teams, and what does that look like?

A Yes, I think there are two camps of mistakes I've seen. One camp is they have a growth team too early when the retention is really bad, or it's not even strong enough, and so the growth team is working really hard to get all these users, but there's a big leaky bucket, and so it's not helping anyone, because you're trying to acquire, you know, incremental users that are not that In terms of retention, and so all the effort goes wasted. So too often, one They have a growth team before they fix their retention, and then wonder, why is it not working? The second camp is waiting too long before having a growth team, and this really comes down to the CEO. I think that most CEOs and product CEOs would tell you that they don't like to be driven by data to make product decisions, and I think the two are not mutually exclusive. Having a growth team doesn't mean they're going to drive all your product decisions, and so the CEOs who have figured that out know to balance Some things customers don't know what they want, and truly you have to come up with your product testing, and that's what the product team does. But if you have a core product that's working to help scale, it is essential or, you know, very beneficial to have a growth team that's focused on using data to help inform how to drive that. And sometimes it may help inform the product decisions too. There's a very good example i…

AI assessment note: “I think there are two camps of mistakes I've seen.”

Q You mentioned retention there. You mentioned stickiness earlier. I'm intrigued by kind of a core property to the benefits of network effects being the defensibility element. How does one stress test and determine the strength of defensibility of a network effect? We've seen several on-demand companies that maybe leverage network effects. Maybe lose theirs. How do you think about sustainability of network effect?

A Yeah, I think there are really two levels of defensibility that you have to use to test. One is what is the barrier to entry? So what does it take for another player to come and offer the same thing that you do? And two is the barrier to exit for the user. So often I've seen a lot of people pay attention to the first one, right? Almost every founder is paranoid about who are the competitors? What are they offering? How does my product compare to that? Are my users using both or are they using me? And I think they do a very good job on number one. What they forget about is number two. What is the barrier to exit? But it has a lot of things tied behind it. So in the case of communication products like Facebook, WhatsApp, it was very clear, right? So if I joined Facebook and I had all these connections and messages exchanged and a history with all my users, it was really difficult for me to switch as a user, even though Google Plus was launched. I have all this history. I have all these connections. I know how to reach them. I also have my newsfeed, whatever. And now all of a sudden to switch all of that as a user, you will find it incredibly hard. A product that's not like Facebook, but it has that same network effect is Airbnb, right? Because it's a marketplace and obviously it's a matching of hosts and guests, but over time they've built so many reviews and ratings, both about …

AI assessment note: “there are really two levels of defensibility that you have to use to test”

Q I absolutely love that focus on kind of barrier to exit. I would love to ask a question from a friend, James Currier at NFX Guild, focusing on networking And he asked, will network effects be more or less important in the coming years, do you think, and why?

A I think that network effects will always be very important simply because it helps create winner-take-all markets. So therefore, to the extent people can pay attention as they build their product on how to build a moat, I do think it will become more and more important, not less. Having said that, Network effects is not the only way to scale a company and have So scale is a great way to do that. I mean, Amazon is the prime example of a large company that started off as a first party business, which had no network effects, and they've done phenomenally well as they've scaled it. So there are certain types of products which naturally lend towards network effects, and I think it's prudent for founders who are working on those types of products or services to pay attention to that. But that's not the only way to scale.

AI assessment note: “I do think it will become more and more important, not less.”

Q I'm intrigued. Where do you see most scaling startups go wrong in establishing their growth teams, and what does that look like?

A Yes, I think there are two camps of mistakes I've seen. One camp is they have a growth team too early when the retention is really bad, or it's not even strong enough, and so the growth team is working really hard to get all these users, but there's a big leaky bucket, and so it's not helping anyone, because you're trying to acquire, you know, incremental users that are not that In terms of retention, and so all the effort goes wasted. So too often, one They have a growth team before they fix their retention, and then wonder, why is it not working? The second camp is waiting too long before having a growth team, and this really comes down to the CEO. I think that most CEOs and product CEOs would tell you that they don't like to be driven by data to make product decisions, and I think the two are not mutually exclusive. Having a growth team doesn't mean they're going to drive all your product decisions, and so the CEOs who have figured that out know to balance Some things customers don't know what they want, and truly you have to come up with your product testing, and that's what the product team does. But if you have a core product that's working to help scale, it is essential or, you know, very beneficial to have a growth team that's focused on using data to help inform how to drive that. And sometimes it may help inform the product decisions too. There's a very good example i…

AI assessment note: “I think there are two camps of mistakes I've seen. One camp is”

Q Fundraising best practices from Andreessen to YC, what's your take?

A I think that the two most important things that are like critical, no matter which round you're raising, one is storytelling. It's super important to invest time to really articulate the vision, the mission of the company, and why you're excited about it as a founder, because it not only demonstrates why the investor should really invest in the company, but they're evaluating you as a CEO on your ability to attract really awesome execs and senior leader Two is really having a vision for the future broken down into two-year goals. Because I think vision is important, more important in seed and series A. It's also important in the growth stage. But the more capital you raise, people want to figure out what do you plan to do with the money? How do you track the milestones? And how have you done that in the past? You know, sort of displaying your tracker card.

AI assessment note: “two most important things that are like critical, no matter which round you're raising”

Q You mentioned retention there. You mentioned stickiness earlier. I'm intrigued by kind of a core property to the benefits of network effects being the defensibility element. How does one stress test and determine the strength of defensibility of a network effect? We've seen several on-demand companies that maybe leverage network effects. Maybe lose theirs. How do you think about sustainability of network effect?

A Yeah, I think there are really two levels of defensibility that you have to use to test. One is what is the barrier to entry? So what does it take for another player to come and offer the same thing that you do? And two is the barrier to exit for the user. So often I've seen a lot of people pay attention to the first one, right? Almost every founder is paranoid about who are the competitors? What are they offering? How does my product compare to that? Are my users using both or are they using me? And I think they do a very good job on number one. What they forget about is number two. What is the barrier to exit? But it has a lot of things tied behind it. So in the case of communication products like Facebook, WhatsApp, it was very clear, right? So if I joined Facebook and I had all these connections and messages exchanged and a history with all my users, it was really difficult for me to switch as a user, even though Google Plus was launched. I have all this history. I have all these connections. I know how to reach them. I also have my newsfeed, whatever. And now all of a sudden to switch all of that as a user, you will find it incredibly hard. A product that's not like Facebook, but it has that same network effect is Airbnb, right? Because it's a marketplace and obviously it's a matching of hosts and guests, but over time they've built so many reviews and ratings, both about …

AI assessment note: “two levels of defensibility that you have to use to test”

Q What a fantastic way to start, and I love that intro story. I would love to hear, you said about your time with BCG there, and the patterns being common for the successful startups. What were those patterns? I'm too intrigued to not ask.

A Absolutely. So I think that it really boils down to two things. One is the management team. We both in late stage investing like private equity and venture capital. This is the reason investors spend a lot of time figuring out the quality of the management team and really testing whether the CEO has the ability to attract really good senior execs to come work with them behind their mission to build a team. Because Ultimately, scaling is all about the team, and it's not just one or two people at the company. The second is how much attention they really pay to building a sustainable You can do hacks, and you are, you tend to be scrappy, which is important when you start a company, but at scale, really paying attention to the business levers, and paying attention to unit economics, and figuring out if you're truly building a scalable model is really important, and probably more important in the venture world than the private equity world, simply because you're dependent on private funding to help scale the company for a long time.

AI assessment note: “So I think that it really boils down to two things. One is the management team.”

Q I absolutely love that focus on kind of barrier to exit. I would love to ask a question from a friend, James Currier at NFX Guild, focusing on networking And he asked, will network effects be more or less important in the coming years, do you think, and why?

A I think that network effects will always be very important simply because it helps create winner-take-all markets. So therefore, to the extent people can pay attention as they build their product on how to build a moat, I do think it will become more and more important, not less. Having said that, Network effects is not the only way to scale a company and have So scale is a great way to do that. I mean, Amazon is the prime example of a large company that started off as a first party business, which had no network effects, and they've done phenomenally well as they've scaled it. So there are certain types of products which naturally lend towards network effects, and I think it's prudent for founders who are working on those types of products or services to pay attention to that. But that's not the only way to scale.

AI assessment note: “I do think it will become more and more important, not less.”

Q Fundraising best practices from Andreessen to YC, what's your take?

A I think that the two most important things that are like critical, no matter which round you're raising, one is storytelling. It's super important to invest time to really articulate the vision, the mission of the company, and why you're excited about it as a founder, because it not only demonstrates why the investor should really invest in the company, but they're evaluating you as a CEO on your ability to attract really awesome execs and senior leader Two is really having a vision for the future broken down into two-year goals. Because I think vision is important, more important in seed and series A. It's also important in the growth stage. But the more capital you raise, people want to figure out what do you plan to do with the money? How do you track the milestones? And how have you done that in the past? You know, sort of displaying your tracker card.

AI assessment note: “two most important things that are like critical, no matter which round you're raising”

Q Speaking of the network effects themselves, there are clearly many different types. A friend of mine, James Curry, has illustrated 13 before. I'm intrigued, when analyzing opportunities today with YC, are there certain forms that maybe get you more excited than others due to, it could be scalability, unit economics, whatever the reason, and why is that?

A Yeah, so I think that I classify the direct Network effect, which is the product virality itself that causes the network effect is probably like the ultimate form of network effect. So if you look at the telephone when they first launched, it was useless if just one user had the telephone. You know, you wanted to be able to call someone on the other end, and so as a result, more users started adopting the telephone because then I could reach everyone via phone versus sending an in-mail letter or a written letter at that time. So that is what I call product virality, which is the product itself spreads because of the organic consequence of its use. The other product that has It has a similar network effect as Facebook, because if you went on Facebook on your own without any friends, like, what are you going to do? You're not, it has some utility, but not a lot. The most common use case of Facebook is really because you want to look at your other friends' lives and be connected with them. So the product spread because the connection was what made it worth it. So that is, I think, the ultimate form of network effect, which is like really sticky, also has the advantage of viral growth, and If the company and the product really executed well, they can actually scale much faster, and they accumulate more value. Having said that, most companies, you know, especially the ones that appl…

AI assessment note: “direct Network effect, which is the product virality itself that causes the network effect”

Q And you can ask, when is that inflection point with the need to segment a team specifically for growth? What are your thoughts on this and kind of advising investing in companies now? When do you see that now we need a dedicated growth team?

A Yeah, I think the first step is really to Strong retention. There is no point in forming a growth team because you have a leaky bucket. And so the first step is to really pay attention to their attention. And so that really comes down to what is the product you're offering and how do you truly benchmark your attention to see whether it's good or not? So let's take an example, right? So Airbnb, when they first launched, I mean, they didn't have vital growth. I mean, they really struggled for the first three years to build liquidity in their marketplace, but One of the things they paid attention to was, what is my retention? And, you know, Airbnb is travel, so people use it once or twice a year. So you can only measure retention once or twice a year in terms of bookings, right? Because you have to wait an entire year to see if they're going to come back. However, one of the things they paid attention to in the early days was, people actually spent a bit of time researching. You usually come back from a vacation, and you want to plan your next vacation, and you start doing some research. So, You can actually track the number of visitors that are coming back to research on your site, the number of visitors that are actually looking at listings or a particular city, and, you know, people tend to book, even if they book their travel only a few weeks before, they tend to research a fe…

AI assessment note: “There is no point in forming a growth team because you have a leaky bucket.”

Q What a fantastic way to start, and I love that intro story. I would love to hear, you said about your time with BCG there, and the patterns being common for the successful startups. What were those patterns? I'm too intrigued to not ask.

A Absolutely. So I think that it really boils down to two things. One is the management team. We both in late stage investing like private equity and venture capital. This is the reason investors spend a lot of time figuring out the quality of the management team and really testing whether the CEO has the ability to attract really good senior execs to come work with them behind their mission to build a team. Because Ultimately, scaling is all about the team, and it's not just one or two people at the company. The second is how much attention they really pay to building a sustainable You can do hacks, and you are, you tend to be scrappy, which is important when you start a company, but at scale, really paying attention to the business levers, and paying attention to unit economics, and figuring out if you're truly building a scalable model is really important, and probably more important in the venture world than the private equity world, simply because you're dependent on private funding to help scale the company for a long time.

AI assessment note: “So I think that it really boils down to two things. One is the management team.”

Q Speaking of the network effects themselves, there are clearly many different types. A friend of mine, James Curry, has illustrated 13 before. I'm intrigued, when analyzing opportunities today with YC, are there certain forms that maybe get you more excited than others due to, it could be scalability, unit economics, whatever the reason, and why is that?

A Yeah, so I think that I classify the direct Network effect, which is the product virality itself that causes the network effect is probably like the ultimate form of network effect. So if you look at the telephone when they first launched, it was useless if just one user had the telephone. You know, you wanted to be able to call someone on the other end, and so as a result, more users started adopting the telephone because then I could reach everyone via phone versus sending an in-mail letter or a written letter at that time. So that is what I call product virality, which is the product itself spreads because of the organic consequence of its use. The other product that has It has a similar network effect as Facebook, because if you went on Facebook on your own without any friends, like, what are you going to do? You're not, it has some utility, but not a lot. The most common use case of Facebook is really because you want to look at your other friends' lives and be connected with them. So the product spread because the connection was what made it worth it. So that is, I think, the ultimate form of network effect, which is like really sticky, also has the advantage of viral growth, and If the company and the product really executed well, they can actually scale much faster, and they accumulate more value. Having said that, most companies, you know, especially the ones that appl…

AI assessment note: “direct Network effect, which is the product virality itself... is probably like the ultimate form”

Q And you can ask, when is that inflection point with the need to segment a team specifically for growth? What are your thoughts on this and kind of advising investing in companies now? When do you see that now we need a dedicated growth team?

A Yeah, I think the first step is really to Strong retention. There is no point in forming a growth team because you have a leaky bucket. And so the first step is to really pay attention to their attention. And so that really comes down to what is the product you're offering and how do you truly benchmark your attention to see whether it's good or not? So let's take an example, right? So Airbnb, when they first launched, I mean, they didn't have vital growth. I mean, they really struggled for the first three years to build liquidity in their marketplace, but One of the things they paid attention to was, what is my retention? And, you know, Airbnb is travel, so people use it once or twice a year. So you can only measure retention once or twice a year in terms of bookings, right? Because you have to wait an entire year to see if they're going to come back. However, one of the things they paid attention to in the early days was, people actually spent a bit of time researching. You usually come back from a vacation, and you want to plan your next vacation, and you start doing some research. So, You can actually track the number of visitors that are coming back to research on your site, the number of visitors that are actually looking at listings or a particular city, and, you know, people tend to book, even if they book their travel only a few weeks before, they tend to research a fe…

AI assessment note: “There is no point in forming a growth team because you have a leaky bucket.”