How to build breakout products

Note: Transcripts are automatically generated from episode audio, and are not fully corrected for spelling, grammar, and formatting.

MARK PINCUS: It does feel very analogous to internet nuclear winter, as we called it in ’02 to ’03, right when you started LinkedIn and I started Tribe. This feels like a similar moment. It’s not figured out, but you know it’s going to be.

JEFF BERMAN: This is Masters of Scale.

[THEME MUSIC]

This week on the show, Reid Hoffman sits down with Mark Pincus. You should know Mark, but if you don’t, he’s a serial tech entrepreneur and investor best known as the founder of Zynga. Thanks to games like FarmVille, Zynga blitzscaled to 300 million monthly active users on Facebook at its peak. They were also huge on MySpace, but we don’t remember MySpace so much these days. Now Mark has a new book out, Life at the Speed of Play. The godfather of mobile gaming talks with Reid about how to build products people love, how to learn fast from your mistakes, and much more.

HOFFMAN: Mark, it’s always an honor and a pleasure. Welcome back to Masters of Scale.

PINCUS: Happy to be invited back.

HOFFMAN: Part of the reason we decided to do this is that you have a new book coming out, and I wrote the foreword. I’m a huge fan of the book. But okay, internet boy, why did you write a book? What led you to writing a book? Very old-school media.

PINCUS: Well, coming from someone who’s been prolific with his books, internet boy.

HOFFMAN: Yes.

PINCUS: I would say, after writing the book, I think everyone should write a book once. I think it’s one of those things that’s such a labor of love. But I wanted to capture a few things in the book. It’s one part life motivation, one part stories, and one part product management bible, and we’ll see if they all work. Originally, you may remember, the first few years I was writing this book, I called it Proven, Better, New because I was like, “That’s the core value, that’s the core value.” But I started to realize that that’s in service of something much bigger. To me, Life at the Speed of Play has meaning on a bunch of levels. One is the vibe. It’s the way that I think we should approach making products. I think it should be fun and playful and light. Someone asked me at a party, “Is this book also targeted to children?” And I said, “No, actually, they already know how to play.” So I think remembering the level of fun and imagination that we have in play as kids, and giving ourselves permission to play, is important.

I also think that in this AI era, which you told me about years before we got to it… I wish I’d listened harder. In this AI era, we are hurtling toward living our lives at the speed of play. And I think that’s becoming more and more accessible to all of us. This idea that you don’t have to quit your job to try out and build an idea that you’ve been thinking about. You don’t have to mortgage your house, raise a lot of venture capital, and do all these things before you can experiment. That wasn’t true when you and I were starting out, and even a couple of years ago that wasn’t true. And now we’re at this place where we have the one-person founder team. So that’s the other part of it: it’s the era that we’re entering.

Copy LinkMark Pincus on his Proven, Better, New framework

HOFFMAN: This Proven, Better, New — what’s the framework? What’s the baseline for understanding it? We’ll do that, and then we’ll move to talking about that in the age of AI.

PINCUS: In order to explain Proven, Better, New, I’ve got to start a chapter earlier, on instincts versus ideas.

HOFFMAN: Okay. Great. Yes.

PINCUS: Literally a chapter earlier, I guess. The foundational point for me as a product maker is that we have winning instincts, these things in our gut. You know it, and you tune into it, and you see it and think, “I know that’s right.” So if you accept this premise that your instincts are right and your ideas are usually wrong, then what do you do with that? What I mean by that is, you could say we all had the idea for Raya, the dating app, or whatever, but we didn’t. We had the instinct that something was broken there or that there was a better way, but not that idea.

So Proven, Better, New is this framework that we got to out of the failure of Tribe. Then came the commitment to winning at Zynga and, with that, this framework of Proven, Better, New. The idea was to say, “Okay, we’ve got to be more scientific about this.” It’s a way to isolate and box in your innovation in one place and to say, “This is where I’m innovating, and this isn’t.” So if you’re Zuck and you’re building Threads, which is a copy, a fast follow of another very successful product, he did a very good job of Proven, Better, New. I’d give him an A-minus. And the things that are proven are things that, on that platform, for those users, work and that you’re not planning to make better. And you should copy those legally. If you do it artfully, people don’t realize it’s a copy, but you should not change a single thing.

HOFFMAN: Stay as close to it as possible without doing anything illegal.

PINCUS: Yeah. The real art of it is that the end user doesn’t feel like it’s a blatant copy. That’s the real magic trick.

HOFFMAN: By the way, there are funny things in copies because, in 2004, when I went to Shenzhen, there were three copies of LinkedIn. They were such exact copies in Chinese that I could fill out a profile, and I don’t speak any Chinese — speak, read, write, et cetera. But I’d just click on this button, type in this field. But that’s proven: no, we know what the game mechanic is. We know what the product mechanic is. We see the demand, we see the signal for how people use it, why they value it, et cetera. Stay close to that within what you’re doing in your product.

PINCUS: Yeah. And the best product makers … The irony of this, the paradox of this, is that the best masters of the craft are the best at copying, and they have the least ego around it. And the most junior, amateur people feel the need to change everything, and they’re the worst at copying, and they fail for the wrong reasons. And that’s the thing: if you don’t do the proven right, you’re usually going to get false-negative signal, very rarely false-positive signal, but your product will fail for the wrong reasons. Then better is — what you believe is better is called new. What’s actually better is what 10 out of 10 of the existing users of that product would say. So better is usually very simple and basic things like it’s now free, no download. It’s something they really care about. It’s like one inch better.

HOFFMAN: Where you already have the consumer experience of what this thing is, and what would be obvious would be immediate, like, yes, this is it.

PINCUS: Yes. It’s clearly better. And it’s clearly better because it’s also something I really care about, and I care so much about this product that I noticed that you polished it more.

HOFFMAN: Yes.

PINCUS: Okay. And then new is your novel idea that will probably be the reason people will try this product, and it will probably also fail. So we called it the back of the box: new and improved. My favorite bad example was the game Risk, right? You and I love board games.

HOFFMAN: Yes.

PINCUS: I remember when I said, “Some MBA at Microsoft decided to change my Risk game.” What did they do? It was like proven, worse, and bad new. They took the wooden pieces that I loved and made them plastic. They made them Roman numerals. I want to show my force. I don’t want Roman numerals. That’s not better. And then they had another game on the back called Castle Risk, which I never even learned. I’m like, “What happened to classic Risk?” So that happens too often, that people try to make it better. But if we realize our new idea is probably wrong, then we can plan ahead for it and say, “Well, what are four other ones?” And am I looking around at other products, other versions of what I’m doing, so I can learn faster? So what you try to get to with Proven, Better, New is we want to be like scientists with white lab coats, and we want to prosecute this instinct and say, “Okay, I’m not attached to this idea, and my team isn’t, and I’m going to give myself and my team permission to change every Monday.”

Copy LinkWhat AI changes when there is no clear proven playbook

HOFFMAN: Now take Proven, Better, New and use that lens to talk about the age of AI, because one of the specific strangenesses is there’s very little proven.

PINCUS: Yeah.

HOFFMAN: So you’re suddenly in a place where proven is … okay, ChatGPT is proven in some ways. Claude Code is proven in some ways.

PINCUS: But they’re pretty expensive proven. They’re like, “I can point to something proven.” And all you need … it used to be 10 billion. Now all you need is 100 billion for that game.

HOFFMAN: And the compute cluster.

PINCUS: People say, “Does Proven, Better, New still apply in the AI world when you’re going to a new platform?” And we can come back to whether this even is a new platform. What do you do when there’s no proven? It’s much harder. I tried to deal with it in the book by talking about the fuzzy proven you’re looking for. Is there something that was proven somewhere else that I could try to apply here? That’s a better well to start with. That’s a better vein than just saying, “I’m going to come up with something the world has never seen before.” That’s much less likely because you and I know that the more things change, the more they stay the same. The other hard part of this is when we think about whether there is a platform. I’m curious: Do you believe that AI is a platform?

HOFFMAN: Well, there are several senses in which we talk about things as platforms. In some senses, at least, it is a platform, but platform is—

PINCUS: How about in the consumer sense?

HOFFMAN: AI certainly is platform technology for consumer applications, although the model underlying a number of different agents could be the technological platform for this. There’s also the question of the front door, right? The fact that people will generally go to one, maybe a couple more, that’s the front door for everything. That’s why we use Google and search synonymously.

PINCUS: As the portal.

HOFFMAN: As the portal. So there’s a platform sense in which I think there will be people who have an agent, maybe a couple, but certainly an agent that’s their general front door.

PINCUS: But they don’t yet, right?

HOFFMAN: Well, I think that’s one of the benefits of the ChatGPT position. I think ChatGPT is the closest to that right now.

PINCUS: Totally agree. I would argue it is not yet a platform at a consumer level.

HOFFMAN: Despite the fact that it’s huge, it’s not yet there.

PINCUS: Even in contrast, Google in the early days was a platform for web distribution because there was SEO and SEM through Google. So that was a platform. The web, obviously, and the browser were, but Google was itself a platform enabler and magnifier, and we’re not yet there. I think we could see it. We hope for it.

Copy LinkWhy consumer AI may be bigger than most people expect

HOFFMAN: By the way, this is funny because this is echoing back to our very earliest conversations as friends. Both you and I went through this internet winter, when the internet was considered to be concluded. It was Yahoo, if you remember that. It was Amazon, et cetera, et cetera. All of Silicon Valley broke in two directions. It broke either to enterprise — “Oh, we’re just going to go down enterprise” — or cleantech. And we were like, “You guys, the consumer internet had just started.” There was a reason you were among the earliest Web 2.0 guys, because it was like, guys. Then we formed a little group of people who were doing angel investing. We started doing companies and all the rest because we were like, “Hey, the consumer’s coming.” So part of what you’re saying is that with AI, people haven’t realized the whole consumer revolution that’s going to come with AI. And I agree with you.

As we know, and as we see it in Silicon Valley, everyone’s talking, “Well, OK, enterprise is where the money is.” That was the same thing back in the internet winter, et cetera. And it was like, no, consumer is going to be big. Now, enterprise will be big too. And by the way, they were right. The cleantech generally failed, and the shift to enterprise drove the cloud revolution.

PINCUS: SaaS.

HOFFMAN: SaaS. So there will be a whole bunch of enterprise stuff, but it’s funny how few people are thinking about the consumer thing right now. It’s a very good illustration. That was one of the things you and I started having dinners about, going, the world is here and no one realizes it.

PINCUS: In fact, I remember when we both wrote checks for Friendster, it was not because we thought it had any chance. We were like, “This is an experiment that needs to happen because this is important for the consumer internet.”

HOFFMAN: Yes, exactly. It was later that venture came in around it because it was like, “Oh, look, the growth curve is beginning to happen.” One of the funny things is I hadn’t realized there was such a demand on the consumer side that Friendster didn’t understand virality or growth hacking. Literally, the wave picked up the boat and the tsunami moved it along.

PINCUS: Yes. That’s a good way of putting it.

HOFFMAN: So I agree with you about the AI thing. It’s not yet a platform in those regards.

PINCUS: It does feel very analogous to internet nuclear winter, as we called it, in ’02 to ’03, right when you started LinkedIn and I started Tribe. This feels like a similar moment. It’s not figured out, but you know it’s going to be.

BERMAN: Still ahead, more with Reid Hoffman and Mark Pincus on why great founders need to ignore the fear of embarrassment.

Welcome back to Masters of Scale. You can find this conversation and much more on our YouTube channel, and be sure to check out the link in the show notes to subscribe to our newsletter.

Copy LinkHow game design principles shape breakout products

HOFFMAN: Part of the reason why I think the book is important is because when you were doing Zynga and I was doing LinkedIn, we had a bunch of things that we would trade as friends, but it wasn’t like Settlers of Catan, like two sheep for a rock. It was more, “Hey, I’m seeing this. Hey, I’m seeing this.” It’s knowledge and intel. Part of it was Zynga more or less wrote the modern book on game dynamics and thinking about game dynamics for everything — for product design, for interaction, et cetera. It frequently would get criticism because it was so oriented to game dynamics. It’s like, “Well, wait a minute, are you guys just a game dynamics company? Where are the games?” I was like, “Well, look, we’re very professional at the game dynamics, and we build the games that work with the game dynamics.” And that’s part of the reason why FarmVille, household name, Words With Friends, household name, et cetera, became such big successes. I think this lens in Life at the Speed of Play — the dynamics of Proven, Better, New, instincts and ideas, kill the bad ideas, keep the good ideas, et cetera — is a general thing for everyone.

PINCUS: Games represent a perfect canvas to learn these concepts of Proven, Better, New, because you can break it down into mechanics. But I do think, and I try to point out in the book, that it really does apply to any consumer product.

HOFFMAN: I think the book is important right now because one of the things I think we’re going to see with AI is a massive generation of products.

PINCUS: Yeah.

HOFFMAN: This is just about to go bonkers large.

PINCUS: I believe that. I even see it with my partner, Hillary. She started getting into vibe coding with Claude Code and got way ahead of me. I was like, “Wait a second, this is what I do.” She made this little mobile app for all of us and our nanny. It was pretty amazing. Based on the location, it would recommend activities for five kids, with each kid by age. And I was like, “Wow, there’s a great design aesthetic.” The point to me was, “Oh, that’s where we’re headed. We’re all going to become product makers because it’s democratized, because it’s accessible.” My one daughter, Georgia, is making a poker app. I’m like, “Wait, that’s what I do.” So I do think that everyone becomes full stack when the stack’s automated and it’s just promptable, and then it gets down to taste. Hillary has amazing taste, so I was so impressed with the subtle design of her app.

Copy LinkBuilding fast-testing cultures in the age of AI

HOFFMAN: Which of the key things from the Zynga playbook would you recommend to AI-native companies? For me, it would be launching early in order to learn, because that creates speed of engagement and learning. You obviously practice that. It could also be things like, “We’ll build levels one to 10 and see if it works,” which is something I learned from Zynga, and it’s a classic.

PINCUS: All of that is more possible and at your fingertips today than ever before with AI. I like to say, “Don’t build it right. Build a culture of building it wrong, then build it right,” presuming some level of success. I love Eric Ries. We’ve had great debates. I’m in the same spirit, but that word “viable” is actually hurtful, because viable also means there’s hope of viability.

HOFFMAN: Yes.

PINCUS: What you want is not viable. You want conviction. Viable is a very weak word to me. It’s not strong. Conviction is strong. What I would say from the Zynga playbook that people should ruthlessly copy today, because it was proven, is this: Test more ideas in a week than your industry tests in a year. Build a testing and failure machine, usually at the top of the funnel. Especially in consumer, it starts with, “Is anyone even going to click on this?” Don’t be embarrassed. Don’t build the rest of it. Just build the landing page, just the ad link, and nothing else. It’s amazing to me, even today. I think we’re in this weird paradox with AI. I would’ve thought by now that people would be all over the place with these testing machines, building prototypes in a day and testing them, and they’re not. That’s not what I see. Because you can almost get to that minimum viable product in three months instead of three years, people want to do that, but then they go native in the wrong way. They get too committed to that product, and then they can’t kill it because they’re like, “Well, we have a lot in this, so I’m going to keep iterating on it.” Or, so often, teams say it’s not ready yet.

It’s so true that if you’re not embarrassed, you’re not launching early enough. Launch is the wrong word. You should turn it on at the level where you learn. But I always go to this question with founders: How ambitious are you? And everyone says a 10. Then, when you unpack that 10, you realize, no, you’re not actually. If you’re really ambitious, you’re willing to be wrong for a long time.

You can be wrong for three years and then really right, or a little bit right every day. Most people take a little bit right. Like you, I’m an extreme AI optimist, and I believe that we are on the verge of the greatest gains in not just human productivity, but in the proliferation of new kinds of services for humans, by humans. I don’t want to sound tone-deaf about it, but my belief, just like in every other era of the internet — the web, mobile — is that there are so many industries we can’t imagine today that are going to pop up.

Copy LinkWhat founders and parents can learn about screen habits

HOFFMAN: In addition to being a great entrepreneur and creating stuff, you’re also one of the world’s most dedicated dads. Your kids are amazing, you put a huge amount of time into it, you’re always paying attention, and you’re doing things with them. Obviously, with games and social networks and everything else, people are worried about addictiveness to phones, addictiveness to games, and addictiveness to some apps. As a parent, what do you think? What do you do? What would you recommend the industry do the same or differently?

PINCUS: That’s a great question. The irony is that so many of us who are spending our time building apps and working in this industry shield our kids from it, the same way any parents could, and I’d argue should. With my twin girls, who are 15 now, I made it pretty far. I told them, “No smartphone until you’re 16.” They eventually had to convince me to get them a flip phone. So I said, “You’re going to be the coolest kid. You’re going to be the only one with a flip phone.” They didn’t think they were cool. That got me a couple of years. Then, when they were 14, they did an internship at Niantic working on Pokemon Go. They built a new first-time user experience. I was very proud of it. They convinced me they needed phones for that, and then I wasn’t able to take them away again. But I said, “You’re permanently on probation. If I see you on your phone at dinner or not paying attention to your little brother or sister, your phone’s gone forever.”

So I put the fear of God in them. It’s a good question, though. We enjoy being together. We enjoy playing games together. We have really big debates. I try to meet them where they are, and for the most part, they’re on their phones way less than other kids, and I think that’s really healthy. I have noticed that when one of them is on their screen for long periods of time, they get more isolated and just flatter, and sometimes sad, so I take it away and I’m like, “Go outside.” But I think the reality is it’s a parenting issue for all of us.

Maybe it sounds self-serving, but I don’t blame Facebook for this. I don’t think it’s like cigarettes or that it’s a drug, and I don’t blame games. I think it is a part of our culture and society. What are we modeling for them? We’re on our phones constantly. We don’t even want to look anymore at Apple showing us the amount of time we spend on our phones. I think most of us are. So it seems hypocritical for us to be doing this. It’s like we’re smoking cigarettes and then we blame the cigarette companies for our kids smoking. Who do they see doing that? Now, you say, “Well, how do we get addicted?” But I think it’s a real challenge. It’s giving all of us ADD, and it makes it harder than ever. We have a 1-year-old, we have a 4-year-old, and my partner is very vigilant about no screens ever. A lot of my friends who are new parents are the same way.

I’ll tell you the other thing. The second I had my twins, when they were born in 2010, I had two thoughts immediately. One was, how do I implant GPS devices so I never lose them?

HOFFMAN: Classic dad.

PINCUS: I was like, “Well, that’s a good business model. You’ll never stop paying for that.” Someone told me it was not OK to do that. People still don’t do that. The other one was, could we make an app for parents that would gamify screen time for our kids, that would let us limit it and incentivize them? I had one of those job wheels, an analog job wheel, that they would spin, and they really enjoyed it. So I don’t have any magic answers. What I’ve done with my kids is keep them off as long as I could. They went through being into games, and now my daughter Georgia is obsessed with chess and chess.com. I feel like that’s a more OK addiction than Roblox.

HOFFMAN: Obsession and commitment are part of what can lead to a lot of success. It has to be healthy versions versus unhealthy ones.

PINCUS: Yeah. The game that I’m really trying to get them into, and I tell every parent this, is Claude Code. Whether it’s Claude Code, Codex, or whatever it is, I basically think my kids should be using that. It pains me to see what high school is like today. You don’t go through this, you don’t witness it the way I do, but high school feels like it’s the end of a 100-year cycle, and my poor kids are stuck in it. It just doesn’t feel efficient. They go there all day, then they come home and they get so much homework. My kids have been homeschooled for the last semester, and they were missing the social part, but they had so much more time. They told me they learned more, so the learning efficiency went way up, and they had way more time for other things.

I was like, OK, in an ideal world, they’d be doing much more efficient learning, and then if they were on a screen, it would be for Claude Code, and otherwise they’d be outside doing stuff, with nothing in between.

Copy LinkWhy the metaverse may still arrive

HOFFMAN: There was a huge amount of focus on the metaverse, and now it’s completely pivoted to AI. What are your prognostications on the future of the metaverse?

PINCUS: I think the metaverse is starting to happen, and I think it’s going to happen, and it’s going to be as big as Zuck thought. It’s just not going to happen inside an immersive game goggle that is painful and hot to wear on your head, even if the form factor gets better. Whether we start to wear devices that listen to us, whether we have AR Ray-Bans, or whether we’re just on the phone we’ve always been on, like I was saying before, I think everything is going to get more metaverse-like. I think what’s going to be the big unlock for games is the more we put the real world, real humans, and real social interaction into our games, because basically games should be a good use of your time. It’s this big industry, and it’s a total waste of time.

Imagine if games improved your life skills. Imagine if games were doing LinkedIn for you. Imagine if you were getting a job by playing a game. How much bigger would it be? So I think that’s an unlock. At the same time, I think all of these Duolingos of the world, all of these apps, are going to be more and more gamified. I think we’re going to see more playful experiences. The winning services are going to become more fun and more rewarding, which is, to me, part of the promise of the metaverse. I also think that once the unlock happens for consumer agentic AI, what’s that going to be? That’s the metaverse, right? When we have avatars that are roaming the world while we’re not there, networking and finding us dates and jobs and new ways to make money, that’s the metaverse, right?

It just doesn’t have to be immersive 3D. It could all happen in your brain because you’re getting a message back, a text back, that says, “Reid AI just gave an amazing talk in Bombay. You crushed it.”

Copy LinkIgnoring bad advice and betting on pattern breaks

HOFFMAN: Last question: What’s a piece of advice you’re glad you ignored?

PINCUS: I had a junior partner at Accel once tell me — I think it was when I was trying to raise money for Zynga, and Accel had backed Support.com with a very successful outcome — “I just worry that you’re not coachable.” I said, “That’s a tricky one. I actually have a lot of coaches, but I’ve chosen all of them, and I’m not coachable by a VC who’s never held this job, who’s never sat in this seat.” So I guess it’s not exactly advice, but it was implied advice that I should pretend that I’m going to learn something from the person with money or the adult, and I’ve always said no to that.

HOFFMAN: By the way, I agree with you. The classic VC stupidity is, “I’m a pattern recognizer, and I see patterns.” Look, patterns are useful. Also know how to play the game.

PINCUS: Yeah. And if you just look at patterns, you might miss the pattern interrupt, which is usually AI.

HOFFMAN: Which is also usually the thing that is the discontinuous investment.

PINCUS: Right, which usually leads to the best outcomes. Everyone else is looking at patterns and playing the same game, and you know from your brilliant Settlers of Catan that you have to have some different strategy in your back pocket, or you can’t expect a different outcome.

HOFFMAN: Mark, always a pleasure.

PINCUS: It’s been 10 years, but it was worth it to get to talk again.

HOFFMAN: And we’ll do it again soon.

BERMAN: Thanks again to Mark Pincus for joining us on Masters of Scale. His new book is Life at the Speed of Play: Launch Products People Love. It’s a fantastic read, full of actionable insights like the ones he shared in his conversation here with Reid. We’ll put a link to it in the show notes. Thank you for listening.