Tech Brew Ride Home - (Bonus) a16z's Andrew Chen On The Cold Start Problem

Episode Date: December 11, 2021

The great Andrew Chen of a16z joins us to discuss his new book: The Cold Start Problem: How to Start and Scale Network Effects. This an indispensable guide to get that "rocket-ship" action going for whatever product or company you're working with. Sponsors: RadPowerBikes.com AltoIra.com/techmeme Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Andrew, thank you so much for coming on the Tech Meme Ride Home. Awesome. Thank you for having me. Yeah, I've been seeing your tweets all week where you were super excited.

You're refreshing your browser and the Amazon author portal and things like that. Believe me, I know how that feels. So first of all, congratulations on the book. And it must feel great. It does feel great. After three years of research and interviews and writing and more writing and even more writing and locking myself in a closet to record the audiobook, it feels good to finally be, have it out there this week. So can I start with the dumbest possible question, which is, you know, where the idea for the book came from and why you wanted to do it, but that'll lead us right into the thesis of the, the book, which I think is really important to get to. Yeah, absolutely. Well, so I've spent nearly 15 years in Silicon Valley, and one of the things that really

jumped out at me, as soon as I moved, I grew up in Seattle and I moved to Silicon Valley in 2007, is just how special and how interesting the companies are that get built in the Bay Area. And they are among the most successful, most impacted. full products in the world. These are products like Airbnb and Instagram and Dropbox and Slack and all these amazing products that that I explore in in the book. And and the question is what is the secret? What is the propelling force that unifies all of those things? And is there one? And and and so what I argue and what I got excited about putting together is that at the core of all of these products, what unifies them is that these are, are new products that are meant to connect people in different ways. Sometimes it's because of work and collaboration.

Sometimes it's because somebody's trying to buy something, the other person is trying to sell something. Sometimes it's for social media. One side is creating, the other side is consuming. And so all of these products have that special quality. They connect people. And then the secondary part of it that we as an industry called network effects is that the more users that use these products, the more valuable they get,

And that allows them to sort of spin up their momentum so that they, so that the winners become even bigger winners and it kind of continues and continues until they become really this thing that takes over the world, that takes over entire industries. And the book is really an attempt to provide the underlying theory and structure for how and why this happens. Right, because what's inherent in the name, you know, a cold start is. Yes, getting network effects sort of can turbocharge any product or any business or whatever. But just because you have a good idea, just because you launch a product doesn't mean it's the chicken and the egg. Doesn't mean that anyone's going to notice it. It doesn't mean whatever. But even when you get to sort of product market fit, even when you get to that sort of rocket launch sort of thing,

there's all sorts of problems inherent in keeping that going or even planting the seeds for that. So one of the key terms that you have in the book that maybe you could define for us at the beginning is the idea of an atomic network. Do you want to start there? What's an atomic network? Yeah, well, I want to just respond to one other quick thing that I just said. for one second and then I'll transition over it, which is, yeah, so the title, the Cold Start problem addresses that this is a, there are two sides to this coin, right? On one hand, if you have a product that gets more valuable as more users use it on the other side of the coin, the

problem is that if no one's using the product, it's not valuable at all. And if you go back to, I actually have a quote in the book, which is from the chairman of, the American telephone and telegraph company, aka AT&T, back in 1905, who says that a telephone without a connection at the other end of the line is not even a toy or a scientific instrument. It is one of the most useless things in the world. Its value depends on the connection with the other telephone and increases with the number of connections. And I think, you know, what that tells you is if you jump into like app world in contemporary, in the contemporary tech ecosystem, you could build a product like Tinder with all the swiping and all the profiles and everything and all the messaging.

But if you don't have Tinder's users, what's going to happen is it's not going to be a valuable servers. And you could say the same thing about Airbnb and you could say the same thing about Instagram. These are products where it's easy to copy the features, but it is very hard to copy the network, which is why these guys are so defensible over a long period of time. And so the question becomes if you want to start and build one of these companies, how do you do it? And to use the Tinder example, actually, the Tinder example is really interesting because when that team led by Sean Rad built the very first version of Tinder, it actually got almost all the features right. They actually had the like a way to, they had the swipes

very early. They had big profile photos. They had Facebook mutual friends. They had a lot of those really early. But in the early days, they actually were just telling you. their friends, hey, we just built a new app. Can you just come use this new app? And they would just dribble in these users over time. And the problem is like, they didn't have enough users using the product at the same time in the same way in order for it to be valuable, because you can swipe through five profiles super fast, right? And then be done. Right. And then you're done. It's the phone example where if you don't have anyone to call, if you if you can't go on to a dating network and find, hopefully,

lots of attractive people to peak your interest, then it's a useless. That's right. And it's useful even if you've built all the right features. And so even though they had built the right features, they needed to actually figure out that they should throw a party at the USC campus. They should put bouncers in front of this amazing house they had rented out and tell the bouncers to turn away people unless they'd install Tinder. And then they invited 500 people. And with those 500 USC students, they were able to then take over the USC campus. And once they figured that out, it was much, much easier for them to then go to more colleges, more campuses, and do the same thing over and over and over again to take over more schools.

And in that way, that number, that 500-person number is a special threshold, which I call the atomic network, which is what is the smallest size of network that you need in order for your product, for your, for your, for your, for your product to become high retention and high engagement. And for Tinder, that numbers, they decided it was 500. For a product like Zoom, two people is enough. Two or three people is enough. For a product like Slack, it's much better to use Slack when it's a team of five or ten people. And the whole point of calling an atomic network is that if you really, really focus your energy as a founder or as somebody working at a larger company to introduce a new product,

if you really put your energy into building your first couple atomic networks, then, and some of those networks, sometimes that's college campus, sometimes that's a team at a company, sometimes that's a city like Airbnb and Uber, like really focused on neighborhoods and cities. As soon as you can know how to do that, then you should be able to build your second atomic network, your third one, then your fifth one, and your 10th one,

then your 20th one. And then you can use that to basically build the momentum so that you can begin to see the first glimmers of a positive network effect that you can leverage through the course of your product launch. So to be clear, the 500 number is not anything set in stone. Like it could be different depending on the product. And as you're describing, like, you know, for say a SaaS product, it would be, you know, getting traction at a hot, startup that's growing really quickly so that when you get their team of 50, six months later, when that's a team of 500, that's your 500 or something. So it's not a hard and fast number.

It's more about finding traction and something that's meaningful for that sort of community of the ringword. I don't know. Yeah. Yeah. And I think you need to make sure. I mean, when I interviewed Seward Butterfield from Slack on this, he talked about the idea that three was enough, but getting a whole team, that was when it really would start to

And so I think the important thing is if you start one of these networks and you do not land enough users, such that it's below that threshold for your atomic network, then the whole thing will probably collapse because there won't be enough people. And so you have to have that theory from the early stages, how many users do you need to get this off the ground? And then you need to quickly go there and establish a lot of density and a lot of interconnectedness by targeting teams, campuses, cities, and neighborhoods. And that's how some of the greatest companies have been taken off the ground. There's literally a tipping point where you can see it in the analytics, you can see it in the uptake, where it's almost like once you pass this threshold, sure, you can goose it or whatever, but it starts to take on sort of a life of its own that's inherent in the network effects? Yeah, I think that you can literally make a graph. And when I ran many of the growth teams at Uber during their heyday, we would make these graphs that would show how many drivers do we have in a city that would be kind of the x-axis.

And then the y-axis would be what is the conversion rate such that people would be able to take a trip successfully? And what you saw is, of course, when you have very, very few drivers, the conversion rates are very low because the EPA's, The amount of time it takes for a driver to get to you is going to be like 30 minutes, right? So you're not going to take Uber on that case. What happens over time is once you get enough Uber drivers, and it depends city on city by city, once you can start to get a car in 10 or 15 minutes, then people start finding it very useful. Then your network has really been filled in. But then on the other side, once you have so many drivers that you can get a ride in one minute or two minutes,

well, you're into the diminishing returns at that point. then it's then your conversion rate starts to level off because it doesn't matter if you can get it in one minute versus two that's fast enough and I think that that's that that that a that diminishing return on the size of the network is something that you see everywhere as well because if you look at booking dot com or you look at slack or you look at any of these once you have too many people too many participants on your network it just actually starts to get a little bit cluttered you need to start then investing in discovery and algorithms and all these other things in order to increase relevance. And that's when your network, you need to start maintaining it. And that's on the other side of the cold strap problem. Right, because it's not like you hit this tipping point and then it's all gravy. You don't have to do it. It's almost like that's when the work starts because then if you do get lucky enough to

get that sort of growth trajectory from your network effects, then it's a matter of managing the network so that that growth effect isn't being killed by the fact that you're having that growth, right? That's right. That's right. And I think we see that in every single one of these types of products that there's an inevitable set of forces like there's market saturation. Once you acquire, once Uber when I was there, acquired all the city dwellers in San Francisco and New York and places like that. Then we started acquiring more like people in the suburbs. They just had less of a use for Uber. You start dealing with fraudsters and trolls and trolls and spammers who decide that your new amazing social media platform is one that they're going to spend all the time trying to take money

from your users. Even Dropbox, which is a B2B product, found that a lot of users were putting like pirated movies into their folders and then sharing out these pirated movies, in Southeast Asia. And, you know, these are not good use cases for these underlying products, but you have to figure out how to deal with these. Because once you have hundreds of millions of users, these kinds of bad actors are just going to show up. You also talk about, and I think you use Twitch as an example of this,

that a lot of times, again, when you are lucky enough, fortunate enough, to get these network effects and hit that growth curve, that you'll often see companies hit a ceiling or products hit a ceiling. So talk to me about that in terms of, is this similar to what we're talking about, where it's almost like you're the victim of your own success, where it's about managing your success? That's right. I think behind the scenes of all of these iconic companies,

when you ask folks, hey, have there been periods where you had plateaus, you know, slow periods, things where nothing seemed to work. Everyone was panicking. And you just hear these amazing stories. I talked to some early Facebook people as part of writing the book. And even Facebook, which has billions of daily active users, had six distinct points before their IPO, where they were flat month over month. And they had to form these growth teams and figure out how to innovate their way out of these plateaus.

And the way that the innovate is, is, Brian, exactly what you're talking about, which is, in many cases, they need to figure out, well, in the Twitch case, for example, they started out as Justin TV. You could stream anything on Justin TV that you wanted. A lot of it was like pirated movies, pirated football games. And they decided that even though they had millions of users for Justin TV, that the most engaged kind of five or 10% of their user base were the folks that were playing video games and watching other people play video games. And so they narrowed the product down to that. And then they reinforced all the incentives for why you would stream on Twitch, which was you would pay the streamers. You would give them tools so that their streaming capability was even better. And by doing all of those things, the team was able to then build what we know now as as Twitch,

a company that if it were an independent standalone company would be, worth many, many billions, maybe many tens of billions of dollars. You also, I mean, towards the end of the book, putting my history hat on for a second, again, coming to the idea that it's not just, once you pass this threshold, it's all gravy. Like, network collapse is possible. You talk about, like, you know, MySpace, Bebo. There's a whole raft. There's a long list.

There's a long list. Yeah. So is there anything that you've seen in your research for how to avoid that? Or even maybe even better than avoiding it, are there signals that you can see where even if we've gotten a million users where we had 10 before or 10 million users, 100 million users, where you can see, uh-oh, we might be in danger of collapsing under our own weight? Yeah. Well, I think that there are a couple of things to, to keep.

keep in mind. I'll start with one of the most exciting versions of all this, which is when there is a big new shift in computing platforms. When you are in the desktop world, and I have a bunch of sections in the book about Microsoft and what it was like to be part of the Microsoft ecosystem in the 1990s, I mean, they were just dominant. They were just ferocious. And you had such a big opportunity to do. disrupt their ecosystem when the internet came out and the web came out. And you could build a lot of new properties, whether that was eBay, whether that was Yahoo, that maybe Microsoft would have liked to have owned when they had pure dominance. And I think the same thing happened in the mobile era where WhatsApp and Instagram became the most important new social products on the planet and became natural to fold that into the

Facebook ecosystem. And I think we're going to see the same thing play out actually because of Web3. Web3 is causing a lot of energy and creativity. A lot of new founders and startups are emerging, and they're blending a lot of new technologies combined with a lot of what we've learned in the past. And so I think we're going to start to see new games platforms, new social networks, new collaboration tools all powered by Web3. And it's going to provide a new opportunity for the next Google or Apple or or any of these companies to emerge. Well, you you got the jump on me because my next question was going to be about how this is applicable to Web 3 in the sense that, you know, not that it was invented by Web 2.0, but you know, one of the defining things of Web 2.0

was the idea of network effects, especially with social media and things like that. So maybe the way I want to ask this is because we're so early in Web 3, and you can either answer like how do you apply the Cold Start problem to Web 3, or because we're so early in Web 3 right now, and we're still in the phase of educating people about it, converting people to the ideas. If you were starting a Web 3 company right now and you wanted to get these sort of network effects, where again, you're still having to educate. It's not like you flip a switch and everyone knows what to do with your product. What are some key insights maybe that if I was starting a Web 3 company

right now, you would use to sort of jump ahead towards getting that sort of traction and stuff? Yeah. Well, I think the excitement level in Web3 is just incredible. And I like to compare it to the period of mobile where, if you remember, there was like flashlight apps, there was like fart apps. There was like all sorts of stuff. Anytime there was a cool app, we'd all go rush out and install it and just check out what is this thing. Well, but not only that, I remember when my brother first got a phone where he could get sports scores on it. And I literally said to him, like, well, why do I know? need that when I can just go upstairs and go on my desktop to check the towards my name? Why do I need to pull it out of my pocket frame? That's right. We're still in that era.

And what that means is whenever there's a cool new project, everyone tries it out. Everyone is doing its own thing. And I think that entire energy is going to settle. It's going to settle and it's going to mature. And I think what we're going to find is that what I think everyone's going to realize is that Web3 has network effects built into its core. And you need to actually master network

network effects in a web three context in order to be six. So what do I mean by that? Well, take Bitcoin, for example. Brian, I value Bitcoin, partly because I think you'll value Bitcoin and because all these other people are going to value Bitcoin. And if you took all the same source code and took all the same blockchain, all that history that's been saved and you forked it and you created, you know, DIT coin, which had all the same properties, but there was no participants.

that be valuable or not? The answer is probably not, right? Everyone still thinks Bitcoin's valuable because everyone else thinks Bitcoin is valuable. It's just circular in that way. And that's true whether we're talking about Bitcoin or we're talking about board apes or we're talking about crypto punks. And so I think the first thing is anyone that's a Web3 entrepreneur right now actually does have to solve a Cold Star problem. They have to think about what Discord community, what Reddit community, what set of Twitter influencers do I engage with, so that the right people will get excited to buy and hold this new set of NFTs I'm about to drop. And if you don't know how to figure that out, then no one's going to find your thing valuable.

And in the same way, you may have seen that the Board Apes team has added dogs, they've added mutants, they've added kind of all these add-ons. And they are doing a masterful job doing the equivalent of one of these companies adding on new products, new product line extensions. This is like Facebook adding Messenger. This is like Facebook adding marketplace. They're leveraging their network of the board apes holders, and they're adding in new things for them to get excited about. And then I think the other thing that we're going to see is that there's a really exciting new area of crypto gaming,

led by companies like Axi Infinity and Mythical and all of these, and Raleigh and Forte and so on. And the really exciting thing about that whole area, is that ultimately the way that I think that whole industry is going to evolve, it's going to evolve the way that Fortnite and Call of Duty Warzone and all of that has evolved over time, which is it's not like a DVD where you go to Best Buy and you're buying this DVD and plugging it into your Xbox. That's not what it is. These are free to play games.

You're going to play them with your friends. You're going to benefit from an ownership, from an economic ownership perspective, of being some of the early players in the game. you're going to buy and sell virtual goods that maybe you can only earn by being amazing at the game. And because of all those things, it's going to make the game a lot more fun because there's going to be an economic game in addition to the actual core gameplay that draws us all there. And I think a lot of this activity that's happening in Web 3 really does have network effects at its core. And somebody that masters it is going to figure out how to turn these things into basically what is going to be the next new social network. The next new social network is not going to just have feeds and profiles and all these things that we've been building for years and years and years.

It's going to be brand new. You've got to look at what the 10-year-old kids are doing. They're spending all their time playing Minecraft and Roblox. And so if that's true, what that means is very likely their social network is not going to look like ours. It's going to look like what they've been used to growing up and playing for many, many years. Well, and you made me think of what Packy said this week about combining money with fun. and things like that. Let me give you, let me ask you one more question and then I'll let you go. But we've been talking about this in terms of advice for founders, advice for product managers,

people with ideas and things like that. But is there, if there are investors, if there's other VC folks listening, how would you say that the cold start problem would apply to finding new ideas and making decisions on investments. Yeah. Well, I think in the same way that for many years, anytime somebody created a social app or more recently, like a creator app or a Metaverse app that everyone gets excited about it, to me, the Cold Star problem, I think,

is a theory that underlies all of those different sectors. And so I think by understanding the theories that I propose that I present, I think it will allow us to extrapolate and then figure out what the next new sectors are. Because fundamentally investing is about what's next. It's about being optimistic about technology. It's about finding the next set of founders. And so when I hear about Metaverse, for example, I'm excited about it because I know embedded within the metaverse idea is network effects.

You're going to want to hang out in the metaverse with your frets, which means that if it works, it's going to be a huge, huge idea, as opposed to there have been product categories and things that I'm less excited about, because even if they succeed, it's not going to be something that's going to be impactful to connecting billions of people. So I think that's one thing. The other thing I would mention is, look, I mean, I think, you know, In our industry, so much of what we do is picking and trying to filter out what's going to work, what's not going to work.

And I think by having this point of view that's developed from these 30 plus case studies that are laid out in the book, hopefully it creates a lot of the patterns that people can look at and get excited about so that they know that, for example, early, early, early on, if you're building one of these collaboration tools and you're not growing very fast and you're manually onboarding customers that you're like, you know what?

Slack did do that. And it's not like you should bang on the entrepreneur and tell them like, you're not growing fast enough. Like you can Slack did that. And same with same with what we were talking about with Dropboxes. There's a period where they were growing very, very fast.

But they didn't know where their growth was coming from. They didn't realize that they were really a B2B app. And so helping people figure out these key strategic decisions in the context of network effects, hopefully will be really valuable to investors as they shepherd their portfolio towards the next phase. Well, and let me just underline that because this is not a book of just cold theory. Like, these literally are case studies. I mean, from examples that people will know about, about like Zoom, Clubhouse, Instagram, Reddit. So it's tangible examples. It's fantastic. The book is called The Cold Start Problem. Andrew Chen, thank you so much for coming on to talk about it.

And again, congratulations. Absolutely. Thank you, Brian.