The Exchanges

Q Everything else, if I'm a YC company and I'm coming to you, is there anything else to say?

A Um, well, I would say like another meta problem is over reliance on the graphs and the numbers. Like you can look at graphs and you can slice and dice and make some like very interesting observation. Um, but it doesn't tell you why it just tells you what is true. Right. Um, and there are many reasons behind the same truth in the world. And they might lead you in completely different directions about what to do next in your product. Uh, and so I think it's very naive to look at some graph and then decide your product roadmap based on some graph that you make. Um, it might inform areas of exploration or like, why is it that we have one second sessions on bump? We should go figure that out. Um, but it should not tell you what to do. Um, instead you should go talk to users and hear, like ask them, Hey, have you ever opened bump for like a second and not used it? And they may tell you like, oh yeah, sometimes I accidentally open it. Whoops. Sorry. And then, aha, you've like learned something real in the world as opposed to the effect of something real in the world.

AI assessment note: “another meta problem is over reliance on the graphs and the numbers”

Q Everything else, if I'm a YC company and I'm coming to you, is there anything else to say?

A Um, well, I would say like another meta problem is over reliance on the graphs and the numbers. Like you can look at graphs and you can slice and dice and make some like very interesting observation. Um, but it doesn't tell you why it just tells you what is true. Right. Um, and there are many reasons behind the same truth in the world. And they might lead you in completely different directions about what to do next in your product. Uh, and so I think it's very naive to look at some graph and then decide your product roadmap based on some graph that you make. Um, it might inform areas of exploration or like, why is it that we have one second sessions on bump? We should go figure that out. Um, but it should not tell you what to do. Um, instead you should go talk to users and hear, like ask them, Hey, have you ever opened bump for like a second and not used it? And they may tell you like, oh yeah, sometimes I accidentally open it. Whoops. Sorry. And then, aha, you've like learned something real in the world as opposed to the effect of something real in the world.

AI assessment note: “another meta problem is over reliance on the graphs and the numbers.”

Q Did you really get fired from Google twice?

A This is the first question to the quick fire round. My God. I didn't get fired from Google. Um, I did get asked to leave the team twice. Um, and so this is a, I'll give the shortest possible version of this story as I can. Um, in the early days, right after we got acquired, um, there was a vigorous internal debate about whether Google should build a private non-social photo management tool, which is what we got acquired to do, or should we continue on the Google plus social, like building a social network track? Uh, there was vigorous debate. The, it turns out the person that we reported up through cited on the, yeah, we're doing the social thing. That's what we're doing. And, um, I felt so much conviction about Google photos and that it would be successful. The Google photos that we ended up ultimately building that I just decided to not take no for an answer. Like in big companies, there's this phrase disagree and commit when like your bosses tell you to do something you don't want to do. Um, if you talk to anybody I've ever worked with, I'm very bad at doing that. Um, it's very hard for me to do something that I don't believe in. And so I just decided I'm not going to do that, and I basically just told everyone, like, no, I'm going to keep building this with these people on the team, so sorry, and I did that because I, like, had a unique situation. We got acquired, like, I d…

AI assessment note: “I didn't get fired from Google. Um, I did get asked to leave the team twice.”

Q We spoke about kind of obsession and sweating the detail. Again, so many people said this about you, even to the fact that like on your last day at lunch, apparently you were like moaning about a feature that you still hate. I heard this. I did my research. Um, and my question to you is how do you create a culture of product obsession and granularity within your team?

A Um, I think you just have to show it yourself. It has to be authentic. Like you have to personally, um, be obsessed with the success of your product, be obsessed with the product experience, be obsessed with the customer. Um, and then people will see that it's true. And what's great and hard is that it's really hard to fake real obsession, right? Obsession is a form of love. Um, and love is a very hard thing to fake. Uh, it's, I would argue like it's probably the hardest thing to fake in the world. And So you can't just act the part. You can't just pretend to be obsessed about a thing that you're not actually obsessed about. Um, so step one, it's very simple. Work on something that you're obsessed with, um, or you could become obsessed with. Then, just let it show. Be yourself. Be authentic. Like, I remember so many times in the early days of Google Photos, I would, like, walk over with my phone to the engineer who was building the main grid of the iPhone app, and I'd sit next to him and I'd be like, This doesn't feel right to me. Does it feel right to you? Like, see how when I scroll, there's like a little glitch and I saw it. Did you see it? I would record things with my, um, slow-mo camera, like on my other phone, I'd record the grid scrolling and then I play it back in slow-mo for these people. And I'd be like, did you see that little glitch there? And, um, by me just doing…

AI assessment note: “I think you just have to show it yourself. It has to be authentic.”

Q touch on product reviews, because I think they're quite misunderstood, and there's, there's a lot of confusion around them, especially for founders. Um, and so when we think about product reviews for you today, how do you do product reviews? How often do you do them? Who's invited? Again, think of me as like a early stage YC company. How would you advise me on how to do them right?

A Yeah. Again, if you're an early stage YC company, like you probably don't have enough people in your company to even invite to a product review. It's probably just a, you know, turn your head to the person next to you and like talk about what you're going to do, right? And see what they think. Uh, In a, in a later stage company, like I'll, I'll pick more like Google photos, which is like a, you know, larger company, um, a product review. So first of all, like often people have these standing product reviews, like every two weeks, I want to hear what that team's doing. Um, I don't like doing it that way. I think it's personal preference. Um, but I, I prefer to have it be about specific things or plans or projects that are being done or decisions that need to get made where the team, um, Has, like, spent a bunch of time on it, they have an opinion on what they want to do, and they either want, like, feedback on that plan, or they want somebody to make a decision that they don't feel capable of making themselves, and they bring it to a group of people that are the, like, experts on this project, whether, so it's basically the leads of every, like, important part of the project, design, engineering, product, sales, whatever it is, um, and you sit in a room, and you don't spend all your time, like, talking about the plan, because you probably made a deck or a dock or something about…

AI assessment note: “it's basically the leads of every, like, important part of the project, design, engineering”

Q I see is the operators. And what I mean by the operators is the Series C professional CPOs who come in and put a lot of process in place and actually do very little. And they like talking to other CPOs and they like talking to the board of feeling important. Do you agree with that? How do you prevent getting that operator, which actually just brings process and delays?

A Right. Yeah. A hundred percent agree. I would say like the number one failure mode when people hire product leaders is they over index on previous scale. How big of a team have you previously managed or tenure? Like how long have you been doing this? What experience do you have? Had you have experience in our domain? It turns out that if you're the median PM in the world, You're like middle of the stack. Not good. Not bad. You're just median. If you stay in the game long enough, you will be that person by definition almost, right? If you last long enough, you will eventually lead a team of 25 PMs, but you're not actually good necessarily. You're median. And it's very dangerous to hire those people. You've heard this saying before, but like A players hire A players and B players hire C players. It's totally true. It happens like literally all the time. I watch it happen. Once you hire that B player, To be your lead. Like, bye-bye. Your team is gone. Like, the best people on your team are not going to want to work with that person. So, how do you avoid it? I think you've got to retain this ability to be either a craftsman or a visionary. Hopefully both. Because those are the people that people want to follow. They want to listen to what they're saying. They want to work with the people who want to work with that person. I think it's incredibly valuable. Yet, founders think, I bet…

AI assessment note: “Yeah. A hundred percent agree. I would say like the number one failure mode”

Q I see is the operators. And what I mean by the operators is the Series C professional CPOs who come in and put a lot of process in place and actually do very little. And they like talking to other CPOs and they like talking to the board of feeling important. Do you agree with that? How do you prevent getting that operator, which actually just brings process and delays?

A Right. Yeah. A hundred percent agree. I would say like the number one failure mode when people hire product leaders is they over index on previous scale. How big of a team have you previously managed or tenure? Like how long have you been doing this? What experience do you have? Had you have experience in our domain? It turns out that if you're the median PM in the world, You're like middle of the stack. Not good. Not bad. You're just median. If you stay in the game long enough, you will be that person by definition almost, right? If you last long enough, you will eventually lead a team of 25 PMs, but you're not actually good necessarily. You're median. And it's very dangerous to hire those people. You've heard this saying before, but like A players hire A players and B players hire C players. It's totally true. It happens like literally all the time. I watch it happen. Once you hire that B player, To be your lead. Like, bye-bye. Your team is gone. Like, the best people on your team are not going to want to work with that person. So, how do you avoid it? I think you've got to retain this ability to be either a craftsman or a visionary. Hopefully both. Because those are the people that people want to follow. They want to listen to what they're saying. They want to work with the people who want to work with that person. I think it's incredibly valuable. Yet, founders think, I bet…

AI assessment note: “A hundred percent agree... So, how do you avoid it?”

Q Now I want to start, you know, you've had the most incredible tenure with Google. Tell me, how did you make your way into the world of startups and products and come to, you know, lead Google Photos for the last seven or eight years?

A Yeah, I would say it's largely accidental is the true answer. Um, I was an engineer by training in school, and I worked as an engineer at Texas Instruments early in my career, and then I think like a lot of smart engineers of my era, the thing you do next is you leave and you go to business school to learn, quote, business, uh, which I did, and, uh, the first week of business school, I was faced with a personal problem, which was I was meeting all of my new classmates, and we were trying to share our phone numbers with each other, and I just would, like, Like, five times a day, I would, like, type in somebody's phone number, and then ask them how to spell their name, and then I would call them, and they would, like, dismiss the call and add my number to their address book, and it was right when the iPhone had come out, and they had just opened up the app store, and I thought, like, oh, why hasn't anybody solved this problem? Like, this should be a solved problem in the world. Like, maybe I can make a side app that would, like, make it work. Uh, so I started building that as kind of, like, a side project in school with the aspiration of, like, maybe making enough money to pay for my business school, which would, like, be, like, On the order of tens of thousands of dollars. And, uh, we launched it and people started using it like way more than we thought. And so we decided like, …

AI assessment note: “Yeah, I would say it's largely accidental is the true answer.”

Q to ask, you know, when you, you mentioned obviously the insight there with the Google photos and the kind of the explosion of data and photos that one would have on, on a phone, you know, uh, again, Juan cracking questions from Juan. He said, what were the top one or two decisions went on in hindsight that you think made Google photos the success that it was? What is?

A I would say The, probably the number one, the meta, the meta decision that we made early on, um, was around what our mission would be. Um, and I think this is a thing that often gets overlooked or not enough time is spent on, on bigger, in bigger companies. Uh, it's probably more critical at small companies that you figure this out, but, uh, we decided that we would have a very clear mission, which was to be the home of the world's memories, right? And that was like the only thing we were going to work on. And we were quite explicit at the beginning about What things we would not go into, right? And we said internally at Google, like, we're not going to build a social tool. This isn't a social network. Like, that's not what we're doing. Uh, we're not going to build some crazy, like, photo editing product. Like, it's not about being, quote, creative with your photos. It's about storing your life's memories. And I think by being really clear about that mission early on, it did a few things for us that I think propagate to this day now, 10 years later. Um, One is it really helped us select who should be on the team and who shouldn't be on the team. There were a bunch of people that were on the existing, we kind of spun the team out of the Google Plus team, and there were a bunch of folks who signed up to build social, right? They were very excited about social, and that's awesome,…

AI assessment note: “meta decision that we made early on, um, was around what our mission would be”

Q next question, but you mentioned some notion of product market fit. There's so many different notions and I have so many founders I was saying like, I don't fucking know if I've got it or not. Some people are somewhat disappointed, some are not disappointed. I don't know. How do you advise a founder to know when something is hitting and there is the first signs of product market fit?

A If you don't know if you've got it or not, you don't, um, is the simplest answer, right? It is a thing where something clicks and people just use your product or demand your product so much more than you could have ever fathomed. That is generally when you've got some initial glimmers of product market fit. You have no time in your day to do anything but satisfy those customers. Whether it's keeping your servers up, if you're a software company, whether it's like, Fielding inbound queries from customers who really have heard about your thing and they really, really need it. Those are the signs that you probably have some glimmers of product market fit. But I think most of the time people find I built something that has some resonance with some small set of people and oh, I must have product market fit now. And generally you might be on the right track, but you probably don't have it if it's unclear whether you've got it or not.

AI assessment note: “If you don't know if you've got it or not, you don't”

Q the team being brilliantly self-sustaining without you. I spoke to so many of them before the show, and they mentioned that you are a product visionary. Literally everyone said you're a product visionary, and I thought that was a good entry point, because there's often this tension between art and science in product. So as a starting point, Is product more art and science? How do you answer that one?

A What a question. My opinion is, the meta answer is it really depends, but my, my opinion is it's more art than science. Um, and so let me explain maybe why I believe that. Um, if you look at like truly great products, pick any product that you use that you're like, that product is just really well done. Um, generally there are a very small set of people behind that product That really love it. That obsess over every single thing about it. Um, it could be one person. It could be a team of five people, but usually it's a quite small number of people who make that thing great. Um, and how do they make it great? Why is it so great? It's because they love it. They love that product. They love the customer. Maybe they are the customer. Maybe the customer is their friend or their family member, and they just want to make something great for that person. Um, I think that is the core of almost all great products. And I think in order to do that, you have to treat it as more of an art. You have to put yourself into it. You have to obsess. Um, and you can't do that with science. Like you can't just do experiments and see, oh, experiment arm one, a was the best performing one. Like, let's go with that. Uh, you have to understand the customer and really feel what they feel. And I think that is much more art than science.

AI assessment note: “my opinion is it's more art than science”

Q In terms of like learning and making it better. I think the really hard thing is often the people who you hear feedback from are the most vocal and they may not be representative of the entire customer base. How do you think about the feedback to ingest and act on versus the feedback to always listen to, but not necessarily act on?

A Yeah. Um, I think there's two bits in there. Um, one bit is who are you talking to and are they representative of your desired user base in the future? Um, it's really great if the answer is you are representative of your desired user base. So I would say in the case of Google photos, like that was Pretty much true, and I would say it still is largely true today. Like, me personally, I think I'm, like, quite close to the center of the distribution in terms of, like, what type of people want to do what with their photos. The second bit though is like, what are they saying to you and what are you taking away from it? Um, I think too often, and this is like something people talk about all the time, so it's not new, but too often people listen to what customers say and they think like, oh, we better just go build the solution that they said, right? They, I don't, I'm confused by the screen. Why don't you put this button there? And they just go do that. Um, and I think that's the total wrong way to do it. You should listen to all the words that they say and all the expressions they make on their face and every, every piece of data you can when you're talking to a customer. And then use that to understand why, why did they say that to me? Are they really confused by this? Or maybe this thing like doesn't matter to them and they haven't like put in any effort to understand it. Like ma…

AI assessment note: “understand the why behind what they're saying, not what they're saying”

Q So I guess my question to you is how do I assess that? I'm literally a sales CEO. What's the process for assessing if they're good at product?

A Um, So I would find some products that you think are good, and some products that you think are bad, and I would ask them about them. Like, hey, what are your favorite products? Like, why are they your favorite products? See if they've ever gotten to the point of, um, thinking about why a product they love is great, versus why a product that they don't love is bad. And can they explain it to you? Do they have an opinion on it? Um, these are, like, things that are very squishy. It's very hard to like, how, how would you like write a multiple choice test that would test that? It's very difficult. Um, but these are the things you're looking for because they show that the person has like thought from first principles about why a thing that they see is true in the world is true. Um, and the why is what matters. And like the skill of asking why and learning why or figuring out why I would say that is like the number one skill for building early stage companies is because that's what you're doing all day long. Why is this working? Why is this not working? How do we do more of the thing that is working?

AI assessment note: “find some products that you think are good, and some products that you think are bad”

Q Can I ask, what do you think are the biggest mistakes that founders make when they construct and then examine their cohorts and retentions?

A Number one mistake is they don't do it, to be honest. Like so many companies I talk to, I ask like, what's your cohort retention curve look like? And they don't know. Now, do I blame them? No, because I did exactly the same thing when I was a founder. Uh, we were raising money for Bump. Uh, we were like sitting in the Sequoia office, and somebody asked like, what's your cohort retention look like? And I'm like, It's good. And then after the meeting, I like went and looked it up and I'm like, oh shit, I have no idea what they were talking about. Like I should, I should know these things. Um, but yeah, so I get it. Like, it's not a normal thing. It's not a thing you do in life except for this.

AI assessment note: “Number one mistake is they don't do it, to be honest.”

Q We spoke about kind of obsession and sweating the detail. Again, so many people said this about you, even to the fact that like on your last day at lunch, apparently you were like moaning about a feature that you still hate. I heard this. I did my research. Um, and my question to you is how do you create a culture of product obsession and granularity within your team?

A Um, I think you just have to show it yourself. It has to be authentic. Like you have to personally, um, be obsessed with the success of your product, be obsessed with the product experience, be obsessed with the customer. Um, and then people will see that it's true. And what's great and hard is that it's really hard to fake real obsession, right? Obsession is a form of love. Um, and love is a very hard thing to fake. Uh, it's, I would argue like it's probably the hardest thing to fake in the world. And So you can't just act the part. You can't just pretend to be obsessed about a thing that you're not actually obsessed about. Um, so step one, it's very simple. Work on something that you're obsessed with, um, or you could become obsessed with. Then, just let it show. Be yourself. Be authentic. Like, I remember so many times in the early days of Google Photos, I would, like, walk over with my phone to the engineer who was building the main grid of the iPhone app, and I'd sit next to him and I'd be like, This doesn't feel right to me. Does it feel right to you? Like, see how when I scroll, there's like a little glitch and I saw it. Did you see it? I would record things with my, um, slow-mo camera, like on my other phone, I'd record the grid scrolling and then I play it back in slow-mo for these people. And I'd be like, did you see that little glitch there? And, um, by me just doing…

AI assessment note: “you just have to show it yourself. It has to be authentic.”

Q What is good retention numbers? Is it 30%? I remember Jeremy Liu told me, I think it was 36% snap pad on 30 day retention. Um, what is good?

A It really depends on your product and your user acquisition channel. Um, you know, all that matters to me is that you're Graph gets flat. It can get flat at five percent of your initial cohort of users, so long as that five percent is flat and consistent and valuable. Like, so, so I think I, I don't have the actual data, but I remember at one point seeing a curve of Airbnb's, um, I think either 30 or maybe 90 day, uh, period cohort retention. And it, it's pretty steep and it gets pretty, pretty low, but it's flat. And each time they use it, Airbnb makes a hundred dollars, right? So that's totally great. It's a hundred billion dollar company. Uh, so, so you've got to understand the context of like, what are those people doing? How frequently are they doing it? And how valuable is it to you and to them to use your product?

AI assessment note: “It really depends on your product and your user acquisition channel.”

Q touch on product reviews, because I think they're quite misunderstood, and there's, there's a lot of confusion around them, especially for founders. Um, and so when we think about product reviews for you today, how do you do product reviews? How often do you do them? Who's invited? Again, think of me as like a early stage YC company. How would you advise me on how to do them right?

A Yeah. Again, if you're an early stage YC company, like you probably don't have enough people in your company to even invite to a product review. It's probably just a, you know, turn your head to the person next to you and like talk about what you're going to do, right? And see what they think. Uh, In a, in a later stage company, like I'll, I'll pick more like Google photos, which is like a, you know, larger company, um, a product review. So first of all, like often people have these standing product reviews, like every two weeks, I want to hear what that team's doing. Um, I don't like doing it that way. I think it's personal preference. Um, but I, I prefer to have it be about specific things or plans or projects that are being done or decisions that need to get made where the team, um, Has, like, spent a bunch of time on it, they have an opinion on what they want to do, and they either want, like, feedback on that plan, or they want somebody to make a decision that they don't feel capable of making themselves, and they bring it to a group of people that are the, like, experts on this project, whether, so it's basically the leads of every, like, important part of the project, design, engineering, product, sales, whatever it is, um, and you sit in a room, and you don't spend all your time, like, talking about the plan, because you probably made a deck or a dock or something about…

AI assessment note: “I prefer to have it be about specific things or plans or projects”

Q is the operators. And what I mean by the operators is the Series C professional CPOs. Who come in and put a lot of process in place and actually do very little. Um, and they like talking to other CPOs and they like talking to the board and feeling important. Do you agree with that? And how do you prevent getting that operator, which actually just brings process and delays?

A Right. Right. Yeah. A hundred percent agree. I would say like the number one failure mode when people hire product leaders is they over index on Either like, um, previous scale, like how many, how big of a team have you previously managed? Or, um, tenure, like how long have you been doing this? What experience do you have? Do you have experience in our domain? Um, it turns out that if you're the median PM in the world, you're like middle of the stack, not, not good, not bad. You're just median. Um, if you stay in the game long enough, you will be that person, like by definition almost, right? If you last long enough, um, you will eventually lead a team of 25 PMs. Uh, but you're not actually good, necessarily. You're median. Um, and it's very dangerous to hire those people, because, um, you know, you've heard this saying before, but, like, A players hire A players, and B players hire C players. It's totally true. It happens, like, literally all the time. I watch it happen, and once you hire that B player to be your lead, like, bye-bye. Your team is gone. Like, the best people on your team are not gonna wanna work with that person. Um, so, how do you avoid it? I think you gotta retain This ability to be either a craftsman or a visionary, hopefully both, uh, because those are the people that people want to follow. They want to listen to what they're saying. They want to work with …

AI assessment note: “A hundred percent agree. I would say like the number one failure mode”

Q Can I ask, how do you ensure everyone feels safe to speak? You are, Dave, a Product OG. You're very young, Dave, but you know, you've been doing this for a while. Um, you can't just say what you think often. How do you ensure that everyone has equal share of voice and it's a free discussion?

A Yeah, yeah. Um, my approach, um, is to try to just be my authentic self all the time, and then people in the room will realize, like, even though maybe I look, like, um, important on paper or whatever, I'm just a regular person, right? I'm just like you. Um, and I think if leaders who approach leadership that way, with, like, a real authenticity, they're just being their normal self. It's the same person that you saw in that meeting that you would have seen on a Saturday morning at my house. Um, I think that is the Best tool that a leader can use to make everyone feel comfortable and able to say what they think. Because again, when have I been most scared in a meeting? It's when I think all the people around me are better and smarter and more important than me. And then I'm like really careful what I say. And that's dangerous for products. You want people to say like all the things that they're feeling. So how do you avoid that? You make sure the room Doesn't feel like a bunch of important people. It's just normal people, right? Uh, and so I think that's, that's like the tool I use.

AI assessment note: “my approach, um, is to try to just be my authentic self all the time”

Q I mean, if you had any Brits on the, talk about the weather from the first minutes, but that's valuable, right? That's right. Totally. You should hire me. I can chat shit For days. Um, final one on product reviews. Where do you think founders, or not founders, but product leads most go wrong with product reviews?

A I think mostly it is spending time on things that you could have done outside of the room or aren't actually even relevant at all. Uh, so it's, it's really cutting to the meat of what, what, like, why are we in this room? Um, And again, it's really hard. Like you've got some person there presenting. They've spent obviously like so much time getting ready for this big review. It's hard as a leader to say like, stop talking. I don't, I've already read your doc. I don't need to hear this. Like what I do want to know is what do you think we should do about this question? Right? Like talk to me about it. Uh, that is a skill that's very hard to master because you're, again, you're like kind of being rude to a person and that's difficult for you. Um, but it's the best way to make it an efficient use of time. So I like product reviews where it's really a debate about what we should do, and people bring, like, different perspectives on what, on that debate, and you fight it out, right? The best product reviews I can, I can remember on Google Photos were ones where I had, like, heated debate, and if you talk to some of the folks on the team, like, my boss in the early days, he and I, like, had heated debates about what to do, um, but we often got to a better answer as a result. Um, if everybody just is the yes man and says, yep, we'll do the thing you said we should do, like, You will mi…

AI assessment note: “spending time on things that you could have done outside of the room”

Q of founders, often building developer tools, API economy, respectfully, they have the charisma of a lemon, um, or a pineapple, I mean, they are the most uncharismatic people in the world, and, uh, It's not the storyteller. So how do you think about that? Because I worry that I'm going to miss the next generation of specifically developer and engineering centric founders because they can't tell a story as well.

A I would say charisma is in the eye of the beholder, right? It's not a uniform thing. Um, you put me in front of a group of people who are going to be the best in the world at building dev tools. They might think that I'm very non-credible. They might look at him and be me and say like, yeah, he like seems like a slick talker, but like, I don't believe him about any of these things. So I think it really is in the eye of the beholder. Um, what, why does charisma matter? It's so you can hire the best people in the world at this problem. And so that you can get the customer to like, listen to you, um, or, or like figure out a way to message to the customer such that they will listen to you. Um, that's very specific to various domains. So I think you can be, What some people might call like very non-charismatic or uncharismatic, but be like very charismatic to the group of people you want to hire.

AI assessment note: “charisma is in the eye of the beholder... very charismatic to the group of people”

Q And that, I think. Do you have any advice on that for me?

A Yeah. You certainly need to make that trade off at some point, but I think most founders do it too early. They think they, they think like, oh, we have a product that like shows some glimmers of, you know, product market fit. Let's go hire a sales team. And that will be the thing that makes us succeed. Um, usually that's the wrong answer. Like usually you are the person who knows the product and the customer the best. Um, and so you should probably be the one doing all of those important jobs for, for longer than you think it's reasonable. Like there should be a point I think where you're like, I can't believe I'm spending all of my time doing this. If only I could spend my time on these other really important things, but I can't, cause this is still the most important thing. Then it's the time to like, go bring somebody in, um, who really knows what they're doing and can shadow you and kind of take where, take it from where you are right now and take it further.

AI assessment note: “most founders do it too early. They think they, they think like, oh”

Q I love the way I have these articulate schedules and then I just get too interested. Why did you sell bump? It was a great product. I remember it back in the day. It was great. Why did you sell?

A Well, the simple answer is we, uh, need, we were gonna need to raise some more money and we weren't sure that that was the right idea for ourselves or for the team or for investors. Um, so usually when you sell a company, it's because you either really can and want to or you need to. For us, we kind of needed to. Um, I think the bigger answer is we had stumbled upon an insight that I think a lot of people in the world hadn't yet had, which is with these mobile phones that now everyone has, um, we are soon going to be inundated with this repository of photos and videos that we really care about, but we have no idea what to do with. Where will they live? How will it keep them? How do you keep them organized? Like what in the world? Um, and I think we found that insight before most people in the world because we were kind of living at the cutting edge of mobile phones. We were taking a lot of photos ourselves to test how bump would work. Um, and I remember like back in the day, you were probably too young for this, but, um, back in the olden days, you would take photos on your iPhone and they were just on your iPhone and they filled up your, your disc on your iPhone and there was no solution. So you had to like plug your iPhone into your computer and then go into the file explorer and drag the photos off of your phone into some folder on your computer. Then you had to figure out l…

AI assessment note: “we were gonna need to raise some more money and we weren't sure”

Q Do you find serial founders or older founders more aligned to that gut instinct, willingness to have the higher variables? Versus maybe more younger insecure who don't want the risk.

A You know, um, I, I think repeat founders often, they, they know more about how the game works. And so they, um, I think they alter their behavior to some extent, both for the good, but also for the bad. They, they don't let themselves run crazy on some wild idea that everyone around them thinks is stupid. Uh, they generally feel some peer pressure to not do that. Whereas very young founders, like naive founders, to some extent, they will go off on crazy pursuits because they think it's the right thing to do, and everyone around them will be like, I don't know what that kid is doing. That's insane. And then they build something that people really love. And so I think, you know, to some extent, as a startup founder, you want to be a little irrational when it comes to how you solve your problem. But the thing you got to be very rational about is Is there a person that actually has some problem that I can solve? And understanding that customer very deeply.

AI assessment note: “repeat founders often... alter their behavior... Whereas very young founders... will go off on crazy pursuits”

Q In terms of like learning and making it better. I think the really hard thing is often the people who you hear feedback from are the most vocal and they may not be representative of the entire customer base. How do you think about the feedback to ingest and act on versus the feedback to always listen to, but not necessarily act on?

A Yeah. Um, I think there's two bits in there. Um, one bit is who are you talking to and are they representative of your desired user base in the future? Um, it's really great if the answer is you are representative of your desired user base. So I would say in the case of Google photos, like that was Pretty much true, and I would say it still is largely true today. Like, me personally, I think I'm, like, quite close to the center of the distribution in terms of, like, what type of people want to do what with their photos. The second bit though is like, what are they saying to you and what are you taking away from it? Um, I think too often, and this is like something people talk about all the time, so it's not new, but too often people listen to what customers say and they think like, oh, we better just go build the solution that they said, right? They, I don't, I'm confused by the screen. Why don't you put this button there? And they just go do that. Um, and I think that's the total wrong way to do it. You should listen to all the words that they say and all the expressions they make on their face and every, every piece of data you can when you're talking to a customer. And then use that to understand why, why did they say that to me? Are they really confused by this? Or maybe this thing like doesn't matter to them and they haven't like put in any effort to understand it. Like ma…

AI assessment note: “you have to understand the why behind what they're saying, not what they're saying.”

Q Can I ask, what do you think are the biggest mistakes that founders make when they construct and then examine their cohorts and retentions?

A Number one mistake is they don't do it, to be honest. Like so many companies I talk to, I ask like, what's your cohort retention curve look like? And they don't know. Now, do I blame them? No, because I did exactly the same thing when I was a founder. Uh, we were raising money for Bump. Uh, we were like sitting in the Sequoia office, and somebody asked like, what's your cohort retention look like? And I'm like, It's good. And then after the meeting, I like went and looked it up and I'm like, oh shit, I have no idea what they were talking about. Like I should, I should know these things. Um, but yeah, so I get it. Like, it's not a normal thing. It's not a thing you do in life except for this.

AI assessment note: “Number one mistake is they don't do it, to be honest.”

Q What is good retention numbers? Is it 30%? I remember Jeremy Liu told me, I think it was 36% snap pad on 30 day retention. Um, what is good?

A It really depends on your product and your user acquisition channel. Um, you know, all that matters to me is that you're Graph gets flat. It can get flat at five percent of your initial cohort of users, so long as that five percent is flat and consistent and valuable. Like, so, so I think I, I don't have the actual data, but I remember at one point seeing a curve of Airbnb's, um, I think either 30 or maybe 90 day, uh, period cohort retention. And it, it's pretty steep and it gets pretty, pretty low, but it's flat. And each time they use it, Airbnb makes a hundred dollars, right? So that's totally great. It's a hundred billion dollar company. Uh, so, so you've got to understand the context of like, what are those people doing? How frequently are they doing it? And how valuable is it to you and to them to use your product?

AI assessment note: “all that matters to me is that you're Graph gets flat.”

Q So I guess my question to you is how do I assess that? I'm literally a sales CEO. What's the process for assessing if they're good at product?

A Um, So I would find some products that you think are good, and some products that you think are bad, and I would ask them about them. Like, hey, what are your favorite products? Like, why are they your favorite products? See if they've ever gotten to the point of, um, thinking about why a product they love is great, versus why a product that they don't love is bad. And can they explain it to you? Do they have an opinion on it? Um, these are, like, things that are very squishy. It's very hard to like, how, how would you like write a multiple choice test that would test that? It's very difficult. Um, but these are the things you're looking for because they show that the person has like thought from first principles about why a thing that they see is true in the world is true. Um, and the why is what matters. And like the skill of asking why and learning why or figuring out why I would say that is like the number one skill for building early stage companies is because that's what you're doing all day long. Why is this working? Why is this not working? How do we do more of the thing that is working?

AI assessment note: “I would find some products that you think are good... and ask them about them”

Q is the operators. And what I mean by the operators is the Series C professional CPOs. Who come in and put a lot of process in place and actually do very little. Um, and they like talking to other CPOs and they like talking to the board and feeling important. Do you agree with that? And how do you prevent getting that operator, which actually just brings process and delays?

A Right. Right. Yeah. A hundred percent agree. I would say like the number one failure mode when people hire product leaders is they over index on Either like, um, previous scale, like how many, how big of a team have you previously managed? Or, um, tenure, like how long have you been doing this? What experience do you have? Do you have experience in our domain? Um, it turns out that if you're the median PM in the world, you're like middle of the stack, not, not good, not bad. You're just median. Um, if you stay in the game long enough, you will be that person, like by definition almost, right? If you last long enough, um, you will eventually lead a team of 25 PMs. Uh, but you're not actually good, necessarily. You're median. Um, and it's very dangerous to hire those people, because, um, you know, you've heard this saying before, but, like, A players hire A players, and B players hire C players. It's totally true. It happens, like, literally all the time. I watch it happen, and once you hire that B player to be your lead, like, bye-bye. Your team is gone. Like, the best people on your team are not gonna wanna work with that person. Um, so, how do you avoid it? I think you gotta retain This ability to be either a craftsman or a visionary, hopefully both, uh, because those are the people that people want to follow. They want to listen to what they're saying. They want to work with …

AI assessment note: “A hundred percent agree. I would say like the number one failure mode”

Q Okay. Okay. And I send it out. I share it three days beforehand, a week beforehand. Do I ask for feedback? Do I ask for commentary?

A Yeah. Um, I prefer doing it in that form where you send a doc that's very like, Succinct. Um, and you ask for either general feedback or you ask for directed feedback on a specific set of questions that you want to get people's inputs on. Um, and people can do it all largely asynchronously. Um, and then I do, I do like getting the group together in person to like talk about the most important bits. Um, but it should be talking about what hasn't been said yet and like bouncing ideas off of each other as opposed to just Rereading the doc or presenting the deck, which I would say, 90% of the time, that's what happens. Um, because humans want to feel validated in their work, they want to feel good, their boss is in the room, yada, yada. Um, that's a big waste of time.

AI assessment note: “you ask for either general feedback or you ask for directed feedback”