Twitter’s ex-Head of Product on Elon, consumer products, culture, more | Kayvon Beykpour

1. 0:00 – 4:31

   Kayvon’s background

   1. KBKayvon Beykpour0:00

      (instrumental music) The first time I ever met Elon was over FaceTime. He was just like, "Do you want to just, like, come, like hang out? You can swipe left or swipe right."

   2. LRLenny Rachitsky0:05

      You're kind of known for, at Twitter, someone that turned the culture of the product team, and Twitter in general, from a very stagnant, nothing is changing product, to shipping all the time.

   3. KBKayvon Beykpour0:15

      We wanted to change the lack of ambition, the lack of creativity, the lack of customers feeling that the product had changed at all.

   4. LRLenny Rachitsky0:21

      So here's a list of stuff that you, or your team shipped while you were there. Super Follows, communities, newsletters, topics, fleets, testing reactions, edge to edge photos, Twitter Blue, Spaces, and obviously live video.

   5. KBKayvon Beykpour0:31

      The sacred cows are like their own roadmap. What are all the things that you think we're not allowed to change? Let's start there.

   6. LRLenny Rachitsky0:37

      And this was all relatively quickly.

   7. KBKayvon Beykpour0:39

      I was like, "I might flame out completely, but hell if I don't try."

   8. LRLenny Rachitsky0:45

      (instrumental music) Today my guest is Kayvon Beykpour. Kayvon was the beloved and longest tenured head of product at Twitter, and also GM of the consumer business at Twitter, up until the day that it was sold to Elon Musk. He landed at Twitter through an acquisition of his company Periscope, which was the world's largest live streaming platform, which ended up inspiring Instagram Live, TikTok Live, Facebook Live, and basically every other social network getting into live video. He sold the company to Twitter in 2015, continued leading Periscope for a number of years, and then moved into leading product, and then the entire consumer business. In our wide ranging conversation, Kayvon shares what it was like getting Elon up to speed at Twitter, what it was like to be fired from Twitter, which actually happened during his pat leave. He also shares all kinds of lessons and stories from transforming Twitter's internal culture from a risk averse, stagnant product org, to one that was shipping major features regularly. We talk about how they used acqui-hires and up and coming hungry product leaders to lead new initiatives and break through many of their sacred cows. We also get into jobs to be done, Elon's layoffs of most of Twitter's staff after the acquisition, his lessons from building and shutting down Periscope, and also building consumer products in general, and so much more. This episode is full of stories and lessons, and a bunch of stuff that you haven't heard anywhere else. With that, I bring you Kayvon Beykpour after a short word from our sponsors. And if you enjoy this podcast, don't forget to subscribe and follow it in your favorite podcasting app or YouTube. It's the best way to avoid missing future episodes, and it helps the podcast tremendously. This episode is brought to you by Interpret. Interpret unifies all of your customer interactions from Gong calls to Zendesk tickets to Twitter threads to App Store reviews, and makes it available for your product team. It's used by leading product orgs like Canva, Notion, Loom, Linear and Descript to accurately integrate the voice of the customer into your product development process, helping you build best in class products. What makes Interpret special is its ability to build customer specific adaptive AI models that provide the most granular and accurate categorization of all your customer feedback, and also connect customer feedback to revenue impact to help product leaders confidently prioritize things that will actually move the needle for your business. If you want a custom model built for your organization so that you can automate your feedback loops and prioritize your roadmap with confidence, get in touch with the team at interpret.com/lenny. That's E-N-T-E-R-P-R-E-T.com/lenny. Today's episode is brought to you by OneSchema, the embeddable CSV importer for SaaS. Customers always seem to want to give you their data in the messiest possible CSV file, and building a spreadsheet importer becomes a never-ending sink for your engineering and support resources. You keep adding features to your spreadsheet importer, but customers keep running into issues. Six months later you're fixing yet another date conversion edge case bug. Most tools aren't built for handling messy data, but OneSchema is. Companies like Scale AI and Pave are using OneSchema to make it fast and easy to launch delightful spreadsheet import experiences, from embeddable CSV import to importing CSVs from an SFTP folder on a recurring basis. Spreadsheet import is such an awful experience in so many products. Customers get frustrated by useless messages like error on line 53 and never end up getting started with your product. OneSchema intelligently corrects messy data so that your customers don't have to spend hours in Excel just to get started with your product. For listeners of this podcast, OneSchema is offering a $1000 discount. Learn more at oneschema.co/lenny.

2. 4:31 – 11:34

   Getting Elon up to speed at Twitter

   1. LRLenny Rachitsky4:32

      Kayvon, thank you so much for being here and welcome to the podcast.

   2. KBKayvon Beykpour4:35

      Thanks so much for having me, Lenny. Great to meet you finally.

   3. LRLenny Rachitsky4:38

      It's amazing to meet you. I think this is going to be quite a unique and interesting podcast. A big thank you to Scott Belsky, illustrious former podcast guest, for introducing us. So when he introduced us, the one thing that he told me is that hopefully Kayvon will share the story about our time getting Elon up to speed at Twitter. I would love to hear that story. I bet other people would too. Are you able to share that story?

   4. KBKayvon Beykpour5:03

      You know, when all the drama went down with, with Twitter and, uh, Elon ended up buying the company, and, um, after the eight month saga of legal back and forth ended up actually taking control of the company, um, there was that first, like, two-day period where it was complete chaos at Twitter with the sink and, like, Elon spreading his tentacles trying to find out, like, you know, who are the people that he wants to keep and what are the projects that are interesting. Um, in the midst of all that, Scott ended up getting contacted and being asked, like, you know, who should I, who should, you know, Elon talk to? And Scott recommended that, um, Elon chat with me, and so I, I... The first time I ever met Elon was over FaceTime, um, where, uh, Elon was just very curious to ask, like, "Hey, you know, you were at Twitter for a while. You seem to, you know, have, have done some things. Like, what, what should I be digging into and who should I be talking to?" At the end of that conversation, um, we ended up arranging an in-person meeting where Scott and I-... um, went to Twitter HQ to actually meet Elon. I think this was like, day two of Elon having, you know, brought the sink in. And so we had this really bizarre, wild, but really fun experience of walking into Twitter HQ and, in my case, walking back to Twitter HQ for the first time after having been fired, which was-

   5. LRLenny Rachitsky6:24

      Hmm.

   6. KBKayvon Beykpour6:24

      ... a very strange experience for me. And, you know, we, we walked into the building. I was sort of like, like scurried through the back door 'cause it, I didn't wanna make a scene and, and make it, you know... Um, there was a lot of, like, rumors around like, "Is Keyvan coming back?" And I just wanted to avoid all of that, um, and so it was just like, a very weird experience of being like, scurried through the elevator and through the back door and go to this like, massive conference room which we had on the second floor of the 110th building and in that massive conference room, it was me, Scott, uh, Elon, and then at the very, very end of the, the room, Walter Isaacson, who by the way, like, I had a hard time... I knew I recognized... I'd never met Walter in person but I was like, "Is that Walter Isaacson?" (laughs) Um, but he said nothing the entire time and we had this very, you know, we had probably like, a two-hour conversation, you know, talking about the past, the future of Twitter, um, the good, the bad, the ugly, and at the very end of the conversation, Walter came up and introduced himself and was like, "Hi Keyvan, I'm Walter. Like, can I get your information just in case I need to ask any follow-up questions?" And I was like, "Oh, shit." (laughs) Like, "I guess that whole conversation was on the record? I don't know." So it was a very surreal experience, um, for a bunch of reasons, including just being weird f- for me since I was very conflicted about coming back to Twitter even physically in the office, um, but it was... And I must say, it was like, it was really fun. It was fun talking to someone... I mean, obviously Elon, I'd never met him before and he's one of the most, you know, um, successful entrepreneurs of our time, and so that was exciting, um, to, to kind of like, go into that meeting. And also, I had been spending so much time dreaming about Twitter and trying to mold Twitter in a direction that I f- I thought was compelling and working with a team of people to, to do that, and to meet someone who, who, who also had a similar sort of ambition, um, but, you know, obviously in, in different ways. He had his own dream for, for Twitter but, it, it was sort of really bizarre and unique, um, and surreal seeing that glimmer in someone else who was like, "Yeah, I also just bought this thing, so I can actually do whatever the hell I want. And by the way, here's some, like, crazy dreams I have for doing it." It was just a really... As someone who had, had their own dreams for the product, like, witnessing that was, um, it was really unusual and, and, and cool. Um, so I think that's like, probably what Scott m- meant when he said to you, "You should ask Keyvan about that." Was that, I think we both kind of recognized that like, something's about to happen here. Obviously, like, we have this very public spectacle of someone essentially having a takeover of a public company, but all of that stuff aside, also, um, you could tell Elon was scheming and cooking up like, you know, "What am I gonna do with this?" Um, and that was... It was cool to, to, to see that.

   7. LRLenny Rachitsky9:12

      That's, uh, an amazing story. I love the Walter Isaacson component of it. Do you feel like you made a dent on his approach and way of thinking? Obviously he made a lot of big radical changes, uh, do you feel like you made a dent in his view of where Twitter should go?

   8. KBKayvon Beykpour9:25

      Uh, I don't, I don't know. I certainly don't think I've m- made any impression on how he should run the company. I think, I think that's, you know, Elon's gonna Elon in his way and I think, you know, he certainly has had some radical moves in terms of, you know, how, how he's running the company. There's decisions he made, like, how many people he let go, how the company is structured, the culture and all that stuff is like, you know... We didn't even talk about that stuff. I think, you know, what we spent time talking about is, um, I, I, you know, shared my perspective with him of like, people I thought who were exceptional who are at the company and, and if I was in his shoes, like, who I would spend time with and embrace and, you know, most of the people that I mentioned are still there which is, which is awesome, and they s- they seem to be A, empowered which is great, and B, like, having fun, which is, which is awesome. So, you know, hopefully that stuff is useful. Um, we spent a bunch of time t- brainstorming products, um, and, and, and, you know, I had sort of my set of projects that I was very passionate about because we'd sort of given, given birth to them and, and you know, I think a bunch of those projects it seems like Twitter's still investing in and putting a lot of energy behind like, Community Notes, which at the time was called Birdwatch, um, but, uh, you know, I always felt really bullish on that being the future of how... Essentially how, like, content is moderated on Twitter, just 'cause it was very clear that the way we were handling content moderation, um, among many other flaws, just wasn't scalable. Um, and Spaces and Communities and, um, the Creator program of help- you know, helping people make money on the platform, like, it's, it's... Those are projects that we started, um, you know, over the year-ish prior to Elon taking over, and my hope, you know, having left the company and having had a new leader come in, was that those things would be given more oxygen and it's been, it's been awesome to see that those have been continuing to, to sort of grow and be molded in different directions. So, um, yeah. I think it, in, in that sense, I hope that, you know, our brainstorm was useful but I mean, for all I know, he doesn't even remember the conversation, so.

   9. LRLenny Rachitsky11:30

      (laughs) I freaking love Community Notes. It's such an amazing product.

3. 11:34 – 21:09

   The story of being let go from Twitter after Elon’s acquisition

   1. LRLenny Rachitsky11:34

      Let me ask two more dramatic-oriented Elon Twitter questions just to get these out of the way. The first is you said you were fired. I don't know if you've shared that publicly. People always wondered, I think, what happened. I know you tweeted during your pat leave, "I'm leaving Twitter." And no one really knows the story as far as I know. What actually went on there?

   2. KBKayvon Beykpour11:53

      It was weird, to say the least. Um, it honestly took me some time to kind of come to, come to peace with it 'cause it was, it was frustrating and, and surprising. Um, I guess the, the story of what ha- wh- what happens starts with, um...... Jack resigning. Jack resigned in, it was like November of 2021, at the end of the year. The board, you know, uh, chose Parag to be the CEO of the company. And, um, you know, I've- I've had a long relationship with Parag. I, I respect him. Um, but I, I had mixed feelings about, about that. But to his credit, you know, Parag very quickly addressed one of the biggest things that I was really frustrated about, like in the last three years, essentially, of my time at Twitter. One of my biggest, uh, points of feedback and, and points of consternation was the structure of the company, and that we had a functional organizational model, meaning we had, like I had a consumer product, me, um, we had a head of revenue product, Bruce, we had a head of engineering, we had a head of design, we had a, you know, research. It was like a functionally run organization. And the combination of that model and the type of leader that Jack was, um, w- wasn't working in my view. Like, uh, I think, you know, if you're gonna have a functional organization, you need to have a GM or a CEO who's like extremely leaned in to tie, break, and, uh, resolve conflict and make sure the team is moving quickly. And, uh, Jack, for all of his amazing qualities, like just wasn't operating that way. And so you had a group of highly opinionated people that often disagreed and would create either the need for consensus or deadlock. And that just was driving a lot of people crazy, including, including me, and I think it really held us back from, from, from living up to Twitter's potential. Anyway, all of that was super frustrating for me. And the combination of that and, you know, a, a new, a dramatic change in leadership with Jack leaving and Parag coming in, you know, I wasn't feeling too stoked. Parag, to his credit, um, when he became CEO, quickly changed that and, you know, shifted the company to be a GM structure. So, um, and he, you know, promoted me to being the GM of consumer. So for the prior three-ish years having been responsible for growing Twitter's consumer product, I was only responsible for the product management team. Um, I didn't have engineering or design. Um, and that honestly was like difficult. It's like trying to ... It's very difficult to change culture with like one hand tied behind your back. Um, still no re- regrets, had a lot of fun, you know, I think we had some impact, but it was, it was frustrating. So Parag changed that, you know, and, and, and the irony of this, by the way, is he was one of the biggest proponents of the functional structure. W- you know, when he became the CEO, he, um, changed the structure, promoted me to the, be the GM of consumer, and that was, you know, I- I was at that point like one month, this is one month before I went on pat leave 'cause my daughter was, um, was due. And so I sort of went into my paternity leave being like, "All right, m- I'm, you know, gonna give this a shot." You know, like, "We'll see, we'll see how this goes." Like, you know, is Parag addressed to the, the biggest frustration, um, that I had with the company and how it was being run? And so I kind of like, you know, had some trepidations but went into my pat leave feeling optimistic. Uh, mind you, this was all before Elon was even part of the picture. He had not become a board member. He had not ... There was no like news about him having beef with the executive team or, or for that matter, like trying to buy the company. So I went on pat leave, you know, maybe a week and a half before our daughter's due date. Um, three weeks goes by. Um, in that three weeks, uh, Elon joins the board, leaves the board, makes an offer, has a short, dramatic feud about whether that, that goes through, and, and also during that time my daughter was born. Um, some drama at the hospital for us, but a week, a week afterwards we come home, mom's healthy, daughter's healthy. Um, the day after we get home from the hospital, Parag called me and said that he was letting me go and that he was taking the team in a different direction. And that night Twitter signed a term sheet with Elon to sell the company. So, you know, a lot happened in a very short period of time. Um, and the r- the, you know, the reason that Parag gave is exactly what I shared publicly, which was that he wanted to take the team in a different direction. Um, uh, you know, the only, only other thing he said is that, um, given that new direction, he thinks that the things that I'm good at Twitter doesn't need anymore, and the things that Twitter needs are not particularly in my skill set or in my interest. He wasn't particularly expressive about what those, um, what that direction was, but that was the reason he gave me. And that was a huge bummer for me for a bunch of reasons. One, um, I love the company, I love the product, and, and also it just like sucks to leave not on your own terms. And two, I was just confused. The timing was very frustrating and confusing for me, not least of which because I had just come home from the hospital while on paternity leave. Um, but also because, you know, at that time especially, the fact that Elon was buying the company was, um ... Well, I was conflicted honestly. I was very excited because Elon is someone that I looked up to immensely, um, and, you know, just look at the things that he's achieved in the world, um, and you can't help but be inspired by that. And two, Twitter for all of, I think, the impact and progress that we had made, had a lot of challenges associated with its, its governance and the fact that it was constantly vulnerable as a public company. So there, there was just always this drama associated with Twitter as a public company, even a private company before that, that made it extremely hard as a builder to get shit done and like have a, have the product live up to its potential. And, um, one of the benefits of this particular takeover was that Elon offered a path towards solving all that. It's like, oh, cool, now you've got one owner who happens to be, by the way, extremely opinionated about the product, and a voracious consumer and creator of the product. And there was something, I think there's an incredible opportunity in that, that now you have this organization and this product and this incredible ecosystem that can be devoid of all the political bullshit associated with being a public company, and, and now it has this like conduit to just living up to its potential. So it was a bummer to, to have, to be removed from that, I suppose, um, without, without having any agency myself. So that was a very long-winded answer to your question, but that's kind of what happened towards the end there. And, and like I said at the beginning.It took me a while to come to terms with it and to be at peace with it. And I, I did eventually... I mean, listen, there's like a, a huge silver lining of I spent the first, you know, year of my daughter's life with her and my family. And my wife, Sarah, had left Twitter, uh, like eight months prior to, to me leaving. And so, you know, when's the next time we could all be together and have time and space to just enjoy each other and our new, our new family? And, and frankly to avoid a lot of the drama that ended up ensuing that not a lot of people could have predicted. Um, the deal was on, the deal was off. You know, it's just a whole lot of drama that I, I got to miss, um, which, which, you know, is the silver lining. And then it was con- confusing there for a bit, because when, you know, when Elon did end up buying the company, um, you know, in that, in that conversation, eh, that I had with him, it was definitely... I was conflicted about like, "Do I wanna maybe spend some time working on this still?" And you know, Elon was very cool about... He actually used this phrase at the end of our conversation, which I still find hilarious. He was just like, "Do you wanna just, like, come? You seem like you care about the product and you don't have dumb ideas. And like, do you wanna come, like, hang out?" And I was like, "What would my job be?" And he was like, "Dunno. You know, just like hang out and you can swipe left or swipe right, or I can..." You know, he, he used the swipe right, swipe left Tinder metaphor, and I thought that was kind of hilarious coming from him.

   3. LRLenny Rachitsky19:38

      For like ideas that come up, just like here's...

   4. KBKayvon Beykpour19:40

      No, like-

   5. LRLenny Rachitsky19:40

      ... I like this idea, don't like this idea?

   6. KBKayvon Beykpour19:40

      ... "Swipe right on whether you want to be here, or swipe..." Like, we did, he's like, "You don't have to, we don't have to make this a thing."

   7. LRLenny Rachitsky19:46

      Love that.

   8. KBKayvon Beykpour19:46

      Just like, you know-

   9. LRLenny Rachitsky19:47

      Yeah, yeah, swipe...

   10. KBKayvon Beykpour19:47

       ... "Do you wanna hang out and, and, you know, work on, work on the product with us?"

   11. LRLenny Rachitsky19:51

       So funny.

   12. KBKayvon Beykpour19:51

       And then so I, you know, I, I sort of ended up deciding that actually I'm just ready to, I'm ready to move on. You know, I've spent enough time at this company, at this product trying to, um, shape it into something that I was passionate about. I think it's someone else's turn. And, and, you know, especially Elon. Like, you buy it. It's your turn, you know? It's, you can do whatever you want with it. So, um, that was conflicting for a bit. But I would say towards the end of the year, it was pret- pretty clear in my mind that I was ready to move on and, and start thinking about other problems.

   13. LRLenny Rachitsky20:23

       What I think about is there's always this, uh, tension being a PM at a company with a very strong-minded, product-oriented founder. And I feel like you would've been in the epitome worst possible situation there, where you're a product leader between Elon and the rest of the org. So I think it probably would not have worked out.

   14. KBKayvon Beykpour20:40

       I'm not sure I would've been able to articulate it as sin- succinctly as you did just now. But I think that is the feeling that I had. Um, that like it just, it's not, it's not my place anymore, you know. Like I, I don't have the canvas to try and exert my, my dreams on this place. I think it's, you know, Elon, Elon took that mantle and, and, you know, I'm excited to see what he's, what he's, what he's gonna do with it, is, is kind of the feeling I had.

   15. LRLenny Rachitsky21:07

       Yeah.

4. 21:09 – 29:44

   Changing the product culture at Twitter

   1. LRLenny Rachitsky21:09

      So you've touched on this. You, you're kind of known for, at Twitter, as someone that turned the culture of the product team, and Twitter in general, from a very stagnant, nothing is changing product, to shipping all the time, all kinds of stuff. So here's a list of stuff that I've gathered you, your team shipped while you were there. Twitter Blue, which is I think called Premium now, Spaces, Super Follows, Communities, newsletters, Topics, Fleets, being able to see Instagram photos inline, testing Reactions, edge to edge photos, tons of UX improvements, and obviously live video. What did you learn about how to change a product culture from a company that's very risk-averse and essentially just not shipping a lot to taking big bold bets and becoming a lot more open to new stuff?

   2. KBKayvon Beykpour21:53

      Trying to drive culture change is both one of the most challenging things and rewarding things. Um, like I, you know, for the first year of my role, the, there is a, there is a chapter of my time at Twitter, uh, maybe just a backup, that was just leading Periscope. And, and in that first chapter of maybe two years, I was not really involved with Twitter stuff all that much. Um, that started to change when we really tried integrating Periscope with, with Twitter. But sort of chapter two of my time at Twitter was when I became the head of product. And that first year of being the head of product was like one of the most difficult of my career. Not because the work was difficult, but because it was so politically and bureaucratically exhausting to try and change culture in a way that just there wasn't alignment around. And it comes back to the point I was making earli- earlier around like the organization of the company was, was functional. And so you know, it's one thing for me to have some ideas and, and a plan and a strategy or that, that I felt compelling. But you know, when, when you have to essentially drive consensus amongst your peers across the other functions, it's, you know, it's, that's a different game. That's not execution. That's like politics and consensus building. And I, I both can't stand that stuff, but I'm just, I think like, um, this is gonna sound like I'm tooting my own horn, but like good enough at it and pa- and I have enough patience at it that I kind of like, you know, I invested the time and en- and the energy. I think a less patient person wouldn't have bothered and would've throwed their hands up. Um, but I, I think honestly a lot of it just comes down to like I ha- I had practiced. Like at my first company that I started in college with one of my best friends, Joe, who I ended up co-founding, uh, Periscope with, we got acquired, um, by a big public EdTech company called Blackboard. And like, we were 19 at the time and we got thrown into like a public... You know, I was a senior executive at a public company that was like not your quintessential tech company. It was even more difficult to kind of get things done. And so through the four years I spent there, I learned a lot about how to, like, navigate that type of environment. And so it was, it all kind of came coming back when, when I was given the product role at Twitter. You know, that first year of changing culture was like, you know, walking through mud. Um, and it was really difficult. But I think when we started building that alignment and sort of like building excitement that like, oh, actually maybe this, maybe we should be taking some bigger swings. Um, and we, when we started seeing through the execution against some of those plans, I think it ended up...... um, it- it got easier and easier, right? 'Cause it- it sort of- it becomes addicting. I think people end up feeling like, "Oh, wow, maybe these sacred cows we had didn't need to be so sacred." And I, so I think, you know, after that first year, it became a lot more fun. Uh, it was still difficult, but it ca- it became, it felt like we were all swimming in the same direction a lot more. But I think my- my takeaway there is, like, you just- you can't change culture without having alignment from the- from the top. Um, you can't change culture. It's difficult to change culture when you have, like, a pocket of the company trying to, you know, advocate for change, which I think we- we got there in the end. We didn't move... As much urgency as I think we drove, we were not fast enough, we were not bold enough. Like, I'm like, I was consistently dissatisfied with what I was achieving and what our team was achieving, but I think, you know, we- we did make a change, you know? Like, Twitter was an organization that had a lot of sacred cows and became very calcified in its ways. L- literally, when- when, you know, in the first two years I was at the company, the stated product strategy for Twitter was refine the core. It was like, you know, "Don't- we're not ch- you know, we're not making any big bets here, team. Like, our goal is to, like, keep turning the knobs that are working." And listen, like, as much as I was kind of throwing stones from the side- sidelines through that period when I was in Periscope Land in our separate office a few blocks away from the mothership, that focus actually did help the company for some period. Like, the reason why Twitter went from stagnant to declining DAU growth to growing DAU again bec- is because they refined the core, right? This is when they went from the reverse chronological timeline to the rank timeline, and that, you know... And the- and the year after that was a lot of knob turning. It was like, "How do we make these recommendations better? How do we make our push notifications more relevant?" Now, that is not an inspiring product strategy. That does not result in the product feeling materially different or adding new capabilities, but it did, uh, return the company to user growth, um. And I think that the fact that it did actually calcified the organization's, um, sort of reticence to take any risky bets even more. So it was a very interesting, uh, sort of predicament, um, 'cause when I- when I, you know, got into the role, the goal wasn't to change that prog-... Like, we wanted to continue reaping the benefits of refining the things about the product that were working really well. What we wanted to change was the lack of ambition, the lack of creativity, the lack of customers feeling that the product was- had changed at all. 'Cause you would hear people... And one of the beautiful things about working on Twitter as a product is that you have literally, like, customers being injected into your veins. Every single day, whatever you change about a product or whatever you don't change, they're telling you what they love and what they hate, um, and it is both exhausting and exhilarating. It is like the- one of the most ridiculous luxuries of product development is working on a product that that many people use and therefore you get that much feedback around. And it was a very common thing for us to hear people say, "What are you- what are you all doing over there? Like, the product hasn't changed in, like, eight years." And that was horrible to hear. That was like... And I felt it too as a- as a critic on the sidelines, as a- as a user who wasn't an employee who eventually became an employee. Like, I had the same feedback. And so that was my, uh, I- that was my mission. I was like, "I am s- you know, somehow ridiculously in this fortunate position that I've been entrusted some responsibility here. Like, I might, you know, flame out completely, but hell if I don't try." And so that was both fun and exhausting, like I said, but, um, it was really... It was as simple as starting with, like, we are voracious users of the product ourselves. A- and if we aren't, by the way, that's its own problem. Like, I don't... I- I- I think that in order to build, um, in order to build something wonderful, like, you have to be a customer of the product. And sure, I'm sure you could- you could point to businesses and products where that's not the case, and I'm sure there's a flaw in that philosophy somewhere. But like, I've always believed that, you know, one of the best ways to build products is to be a customer yourself, and- and to find your own pain points and to, you know, build the product that you want to use. And so, that's, like, actually not that hard to do if you're a- a user of- of Twitter and you, like, you know, can think critically. W- it was ripe with opportunity, and so it was actually really fun and amazing to, um, be able to kind of craft a- a- a plan that started to take a swing at some of these things. And the other thing I'll add to this is that there were so many sacred cows at Twitter that the sacred cows are like their own roadmap. It's like a built-in free roadmap of like, "All right, what are all the things that you think we're not allowed to change? Let's start there." You know? Everything from like- ... you know, moving from reverse chrono to a ranked feed, that was a sacred cow. Um, text and 140 characters, that's a sacred cow. Not letting anyone control any tweets that they see on the platform. Like- like, the notion of Lenny owning his reply- reply space was anathema at Twitter, you know? It's just like, if a tweet gets... You know, if you get an at mention and you don't like it or it's abusive or it's por-... Like, we're not touching that. We can't annotate tweets with community. And it's all that stuff will... Those are all sacred cows. And you know, when- the- the process of starting to address those one by one reveals a lot of the cultural hesitations that- that existed. Um, so

5. 29:44 – 32:02

   Building the “hide replies” feature

   1. KBKayvon Beykpour29:44

      yeah, I- I'll- I'll never forget one of the first features, and this is such a tiny feature that we worked on, you know, after I'd, um, after I started in the role. Uh, we- we were building this feature called Hide Replies. Still exists in the product today. If- if someone replies to one of your tweets, like in- in the tweet details area, like as a reply to one of your conversations, before, you had no ability... The only way you could kind of like address unwanted content was reporting it. So that was like Twitter acting as policeman and policewoman. Completely unscalable and- um, and- and challenging, especially in the context of like someone replying to one of your tweets. And so we wanted to add a feature that let you hide a reply to one of your tweets. And it's not impacting what people say. It's not impacting free speech in the sense of like, you can still broadcast whatever you want, but if you're gonna like come into my res- reply space...... and say some shit that I don't want to see in my feed. Like, I should be able to hide that. You know, tweet what you want, but don't at mention. Don't, you can't scream in my face, basically. Um, and I remember w- uh, we had a PM on our team who was, who was leading this feature who, you know, a few weeks into this project, uh, mentioned to me that she had had a conversation with someone on the engineering team that told them, "Don't work on this feature. This is bad for your career. This is not gonna launch. And you don't want to work on this... You don't want to be seen as having worked on this." 'Cause it was so, uh, kind of anathema to, "We can't let people hide replies to their tweets." Um, and I just remember hearing that and my, my blood was boiling. Like, that is like the, the most, um... It was such an interesting representation of the culture. Like, not just like hesitation to try new things. Which by the way, that, that product might have failed, and that's fine. But we sh- it doesn't mean you shouldn't try. But like, to go so far as to dissuade someone else who was excited about experimenting with a hypothesis to see if it could help customers on the platform, and telling them, like, "Don't work on this. It's bad for your careers." Like, it, as a microcosm for some of the cultural challenge we had around trying big, bold bets, which by the way, this wasn't even a big, bold bet. It was like such a innocuous thing to try. But there was a lot of that, and it made it very difficult without, like, sh- some sheer force of will, and also, like, just a lot of, um, just a lot of effort. Um, so there was a lot of that.

   2. LRLenny Rachitsky32:01

      Hey, Vaughn, you're

6. 32:02 – 34:28

   Sacred crows, taking bold bets, and reigniting growth

   1. LRLenny Rachitsky32:02

      s- there's like fractals of stories that I could infinitely follow. There's so much interesting stuff here. Um, one that's just stands out is this idea of the sacred cows become like your future roadmap. It's like flipping it from, here's the thing that we're all afraid of. No, this is not what we should be doing. I think that's really interesting and could be a lesson to people. The other is just, I, I love this point you made about like the growth was most accelerant when you're just focusing on the core. Like, that's what actually, like optimiz- like people bash on optimizing the existing experience and then just micro-optimizing and improving, versus trying to take all these big bold bets and experiments. Obviously, that's also valuable. But I think it's really interesting that that's what reignited growth and was responsible for growth for a while.

   2. KBKayvon Beykpour32:44

      Yeah. And, and continued to be, by the way. Like we-

   3. LRLenny Rachitsky32:46

      Mm-hmm.

   4. KBKayvon Beykpour32:46

      Like one of the, um, lines that we had to, um, maneuver was creating a portfolio of bets where some of them were not speculative at all. We knew that if we continued to invest in ML, um, and, um, getting better with recommendations in the main feed and through notifications, and you know, things like improving the onboarding flow, like it was just, it was rife with opportunity. You'd have like really dumb things happen. Like we, we would learn, uh, uh, there was this wonderful like just bizarre and incredible meeting we had where we finally had gone more rigorous amou- around instrumenting our onboarding flow, and we found that like in a couple countries, like we, like one of the... We had a bug with our like SMS verification, um, flow, where, you know, we wanted to verify users, um, who were signing up with phone numbers, and our telco integration to send SMS verification codes just wasn't working. And so like a huge percentage of people signing up in like the UAE and other countries just like couldn't, couldn't use Twitter. And so like of course at the scale that we were operating in with hundreds of millions of users, you need to be able to refine the basic building blocks of the product, and that's gonna lead to, to reliable growth. But we wanted to balance that with a portfolio of other bets and, and product improvements that would materially add new capabilities, and, and that, that was a balance that I don't think existed. Um, and, and by the way, I don't think we nailed it either, um, under, under my tenure. But it, it is, it was the driving force of what I wanted to achieve was to create a better balance that would result in evolving the product and introducing new capabilities. And, um, and, um, so that's, that's, you know, that's what we tried to do.

7. 34:28 – 42:40

   Aquihires and their impact

   1. KBKayvon Beykpour34:28

   2. LRLenny Rachitsky34:28

      In terms of what actually helped turn things around, things that I kind of gathered from what you shared so far. One is just building a little momentum, having some quick wins of new products that people start to get excited about, so creating more excitement down the, "Oh, wow. We can actually try new things." There's also felt like there's like a sense of trust that you built with Jack and the execs of just like, "Okay, we can actually trust this team." Also feels like because growth started up, there's probably a sense of like, "Okay, we can try some new crazy ideas." It feels like another part of your strategy was acqui-hires, and bringing in these like entrepreneurial folks to take the lead on some of these big ideas. Can you talk about that? Was that something you actively thought about, and was that a big part of the, the impact there?

   3. KBKayvon Beykpour35:07

      Totally, yeah. So I, I think the, there are a couple things. One, the only thing I'd add to what, to what you said in terms of the ingredients, i- it was also just like storytelling, uh, and just like repetitive storytelling around like this is the vision. This is w- these are the bets we're making and here's why. And like you can't just tell that story once. And I'm talking internally about it. There's a whole other component of this, which is like externally, how do you tell, uh, especially for a product like Twitter where y- you know, you have, it's a consumer product that hundreds of millions of people are using. And you have many constituents. You have users, you have advertisers. And so it was very important for us to tell the story of like, "Here's what we're doing and why. Here's why you should believe in us. And by the way, give us all your constructive criticism too, because like we're listening, and we're gonna build that int-" You know, so that storytelling was really, really important. And you know, there's, this is like oversimplifying the world, but there's two types of internal team members. There are people who hear that story, who have been a part of the organization who's been slow, or maybe they've been outside the company as a user of Twitter, and they're like, "Ah, I'd never really want to work there, because like doesn't seem like a particularly ambitious product company." And one of two things happens when you hear that story. Either you're inspired and you're like, "Yeah, we can finally like take a swing at making this product better. Maybe I wasn't interested in working at this company before, but I mean, this is an iconic product, and to have an opportunity to, to reshape it is really exciting." Or, you know, again, oversimplified world, but there, there are people who are very pessimistic and maybe aren't excited by that vision of like, "Why? Let's just stick to what we know works. You know, we're not gonna take any big swings. That's a waste of time." And I think one of the really important things about...... driving cultural change at the leadership level is you gotta identify whether someone's on the wagon or off the wagon. And, and either quickly convince them to get o- o- on the wagon, or if they're not on the wagon, like, they shouldn't be there. And that's something that we were terrible at. And we didn't have the, we didn't have the organizational structure to be able to enact that.

   4. LRLenny Rachitsky36:59

      The wagon is they're excited and bought into this vision, and want it, want this to happen?

   5. KBKayvon Beykpour37:03

      Correct, and contribute towards it. And, and off the wagon is like you're not at the company. And we didn't have an organizational structure that could allow for that, nor frankly the fortitude. I, I, I include myself in this. Like, I feel like I've learned a lot about, you know, how to make that determination. So I think we were, we were terrible at that. And, and Elon is like the whatever, like if that's a spectrum, Elon is like the opposite spectrum of that. Like, his tolerance for people who are not aligned and his tolerance for low performers is famously extremely low. And I think it's, you know, it's, it's, it's one of the things that when I, when I said AB test, it's like very interesting to see the extreme to which he has operated. And, um, and I think like I've learned a ton around like we just, you know, we didn't have the organizational structure nor the fortitude to be swifter, and that made cultural change way slower. We still were able to change the culture, not as much as we should have, and not as, certainly not as quickly as we should have, or efficiently as we should have. 'Cause by the way, a lot of high performers who have, who are aligned with that desire to, to change and build, they don't have the, they don't, there's not like a equal distribution of patience, right? If you're extremely talented and you're dealing with like organizational bullshit, you're gonna go find someplace else that lets you do your craft and, and have impact. So it's, it's very difficult. But anyway, going to your, your question, one of the things we found that was re- a really effective way of accel- accelerating cultural change and also helping drive some of these product initiatives that were particularly speculative was doing small acquires. And, and, and really the benefit of that was, A, you bring in a, a founder type who is an entrepreneur, who's drives urgency, who has ambition, who's ideally savvy enough to also work on the context of a large organization, which, which, you know, sometimes is a totally different skill set. I mean, it is a totally different skill set. And a lot of our most ambitious bets that were the riskiest and most, um, kind of misaligned with how the, the product worked or, you know, it wasn't like an easy, there was no easy staff for them to build on top of, a lot of those bets were driven by founders who we basically acquired and said, "Here, you, you know, you're gonna run this." Um, and they believed in it. They were able to rally team around them, and it's all the attributes of a startup but with the, the canvas of a product that hundreds of millions of people use, and, and, and more resources. So, you know, spaces, communities, community notes, called Birdwatch back then, um, these were all projects driven by, uh, uh, fleets as well. These were all projects that were run by, um, small teams led by entrepreneurs who we acquired, right? Like Keith Coleman, um, ru- runs community notes, so that's Birdwatch, you know. We, he actually was my predecessor. We, we acquired his company so that he could be the head of product, and then when he moved on from that role, he was extremely passionate about this idea of sort of crowd-sourced moderation and, and, and letting people annotate, you know, uh, misleading content on the platform without Twitter acting as a, a policeman. And that was a very speculative bet that, by the way, a lot of people thought, most people thought was a terrible idea. We gave Keith a little silo to go build, you know, build this vision, and then, you know, it was, it was our job to make sure that bet didn't get suffocated in the context of the big organization that, that would otherwise not, had not had patience for it. Um, you know, all of the community eff- or the, um, creator efforts started with super follows and tipping, and all these things were led by Esther Crawford, whose company we acquired, and who had her own couple viral moments, um, with the, with the Elon acquisition. But she's a phenomenal leader who, again, is a perfect example of balancing that sort of entrepreneurial startup muscle with the savviness to be able to get things done at a large organization. Fleets was run by Mo Aladam, who's an entrepreneur. Um, and, you know, communities was, um, you know, has gone through a few iterations obviously and still is in the product, but John Barnett and a team of people who we acquired from, from Chroma Labs. So I think that that story of acquiring hungry, ambitious f- uh, founders and giving them responsibility and latitude is a success story of Twitter's his- I mean, I, I'm a beneficiary of such a bet as well, right? Like, my company Periscope was acquired, and I was given the responsibility to eventually lead the product team. So I, I feel like my, that we ended up realizing both through the company's history and, and, and through learning is that this is actually a very effective way to drive cultural change, um, and to, and to, to deliver impact. 'Cause you need a special type of person to be able to both operate within the existing structure and change the structure. Like, to know when to use the system and to know when to fuck the system. Um-

   6. LRLenny Rachitsky41:52

      (laughs)

   7. KBKayvon Beykpour41:52

      ... and that, um, I feel like I'm, my whole life has benefited from other people taking bets on me like that, um, to the point that I, even I was like, "Really?" (laughs)

   8. LRLenny Rachitsky42:02

      Hmm.

   9. KBKayvon Beykpour42:02

      You sure you wanna ... You know. Um, and so I've, I think I've enjoyed trying to pass the buck, um, and, and, and do that for other people, and I've never regretted it. It's y- you always, like taking a bet on people and, and especially throwing them in the deep end, which on paper they may be unqualified for, I think is like one of the, one of the best ways of driving change. And, and by the way, supporting growth. Like, you learn so much more when you're thrown in the deep end than in other contexts. And so I think it's a, I think it's a fantastic ... It's an expensive strategy if you're gonna go buy a bunch of companies, but it's a great strategy for the situation that we found ourselves in, uh, at Twitter.

8. 42:40 – 47:00

   Tips for successful acquisitions and staffing

   1. LRLenny Rachitsky42:40

      It feels like every, every big bet was like one of these companies from w- from the list that you just shared, and I'm glad you shared Periscope, obviously that's a great example too. I guess, maybe just a follow-up question here, is there anything you learned about how to do this? Well, I know you talked about maybe creating a little silo for the team, 'cause so many companies acquire and acqui-hire, and they just go nowhere. So, I guess just like a two-part question, just like, what are some tips for how to do this well at a company? And then two, you also, we were talking offline about this previously, and I think it's a really interesting point that a lot of companies staff based on who's available versus who is right for the role, and let's wait until that person is there for us to bet on this. Just talk about lessons there.

   2. KBKayvon Beykpour43:19

      Yeah. That, that last one is, um, is like a huge pet peeve of mine that I feel like we learned the hard way. Um, and, and it's particularly, it's a, it's a common pattern, I think, in highly functional organizations where you have different, different people making decisions on how to staff projects. And there's nothing inherently wrong with that, but I think in the situations we, we... In the situation we found ourselves in, where we had this sort of like cultural, um, evolution that we were going through, where some people just didn't, like, agree with th- the things that we were prioritizing. Um, they're sort of like begrudgingly going along with it, but you would end up in a situation where the combination of that sort of like cultural shift in strategy, and the fact that the way teams were being staffed was not... There's no s- ultimately no single decision maker other than the CEO, and like Jack is not gonna get in the weeds and debate a staffing decision on the team, resulted in a situation where oftentimes we'd have projects, like the one I mentioned about hiding replies, where there wasn't even agreement on the team about whether this was a good idea and whether this is worth trying or how to do it. And imagine, it's hard enough to build something from nothing. It's even harder if, like, the team doesn't believe in it. Like, like this, this is like to the point of just being toxic. Like, you, the startup would never succeed if all the people who are working on that startup aren't, like to the point of being perhaps irrational, obsessed with that idea. A- and still willing to see truth, like you need to be able to see whether the thing is working or not, but if you don't believe it in the first place, you know, I'm not betting on that succeeding. And so this was common. This is... And, and sometimes not exis- as extreme as the examples that I mentioned, but, um, you know, I think one of the lessons I learned, uh, and it's not even, it's, it's quite intuitive actually, is like you need to staff projects with a team of people that are well-equi- equipped from a skillset standpoint, but more importantly have, have an obsession with the idea of they want to pursue. It's gonna make them work harder, it's gonna make them be more creative, it's gonna make them have the, th- th- um, the sufficient level of ambition and desire to will this thing into existence. 'Cause every project, you know, whether big or small, like there is an element of need to like will this thing into existence 'cause it, it's hard, right? Like th- the only way you're gonna get through that pain is by having that desire. And I think a very easy cheat code for an organization to employ is to say if you're gonna work on something, especially if it's speculative or risky, staff it with a set of people who believe in it and want it, like really want to learn whether this solves a customer problem or not. 'Cause if, if you don't have that ingredient, it's gonna drag everyone down and it's just not gonna be as successful.

   3. LRLenny Rachitsky45:56

      (instrumental music) This episode is brought to you by Heap, the product analytics solution that shows you everything users do on your digital product, website, mobile product, or other digital services. We all know a great digital experience when we see it. It's intuitive, it anticipates your needs, and it makes it easy for you to do your job. If you're trying to build that kind of experience for your users, you need up-to-date, reliable information about what your users do in your product and why they do it. Want to know how your users behave across platforms, what keeps them coming back, what they're doing that you're not even aware of? Well, I have some great news for you. Heap captures all of this user activity for you automatically and then gives you definitive answers to all your questions about user behavior in seconds, not weeks. With Heap, it's easy to prioritize the product investments that improve conversion, engagement, and retention. Visit heap.io/lenny to get started with a demo. That's H-E-A-P .io/lenny.

9. 47:00 – 53:20

   The limitations of frameworks like JTBD

   1. LRLenny Rachitsky47:00

      This kind of touches on something I wanted to touch on, which is jobs to be done. This is like a may- one of the most recurring controversies of the podcast is like, is jobs to be done amazing or is it really bad and not something you should do? We've had many guests share their opinions. I feel like at Twitter it, like jobs to be done was implemented so strictly that it burned a lot of people out on. It's like, "Oh my God, this is not anything s- anyone should ever do." I'm curious just your lessons and experience of just with frameworks in general, with jobs to be done specifically, maybe even OKRs.

   2. KBKayvon Beykpour47:31

      I've only seen one interview on your show that covers this, and it's, you know, Sriram was particularly spicy when he talked about jobs to be done, which is unsurprising 'cause I spent a lot of time talking to Sriram about jobs to be done. I mean, I'll, I- I guess I'll just start by saying like I, I was not a fan of how we leveraged jobs to be done at Twitter. I thought it was, um, exhausting and not particularly helpful. Um, and so I, it's a particularly s- particularly sore subject for me because, you know, I, I was sort of charged with defending it and rolling it out, and it's hard to, hard to do that when you don't really believe in something. But to me the critique is less about jobs to be done, though there are many critiques of it, and more about like every framework at its limit if followed to such a religious extent is just unhelpful. Uh, you need to have nuance in how you leverage these frameworks, otherwise people, um, i- i- y- you lose the forest through the trees and you end up following a process for the sake of following a process. And that's what happened with jobs to be done. So I, I think that's my real critique of it. It's not the... I mean, listen, the premise of jobs to be done and my, my, my most ch- charitable take on jobs to be done, which is actually useful, is that it forces you to look at things through the lens of customers and understanding what their needs are and understanding what their true alternatives are outside of the narrow lens of your product. And I think that's just like healthy product thinking, right? Like y- you don't need a framework called jobs to be done, you don't have to think about milkshakes in order to be able to do that. It's just like-... you could employ common sense, or you could leverage something like Jobs to Be Done to, to be able to, like, force your mind to think of things through, through that lens. So I think that's, like, my charitable view on what Jobs to Be Done can help you do. But as a framework into and of itself as a sole governing principle of what to build, it's just not useful. By the way, in the same way as, um, and I think t- Twitter had this, this problem as well prior to our detour around Jobs to Be Done, if you are, if the only way the organization is trained to think about what to build and what not to build is OKRs, it's equally unhelpful at the limit. Because, you know, sure, you can, um, have a good sense of, like, what you should build to drive metrics. But by the way, you might be focusing on the right, wrong metrics, or you, that might not help you have the right balance of, of things to build. That might not help you see when the things that you're building are actually hostile to customers. So, you know,, uh, just as an interesting example, the thing that, that I remember about your interview with Sriram, if I'm not mistaken, he, um, I think he mentioned, um, and I love Sriram, I'd be happy to debate him about this on his podcast, um, but he mentioned, um, eh, one of the examples I think was the Amazon, i- i- uh, Amazon doesn't send you, um, when, when you get order confirmation from Amazon, they intentionally bury the order details. You have to, like, click the link and authenticate to go see what you ordered. I don't give two shits what metrics that drive for Amazon. That is one of the most customer hostile things I experience in my daily life. I order a lot of things from Amazon. I hate the fact that I can't, you know, search my email to see, to see what I ordered. And so I think the problem with these frameworks is that you lose nuance and ultimately, and this is where I agree with Sriram, he actually mentioned this on your podcast as well, you need to be able to make trade-off decisions that balance what's right for the organization and what's right for the customer. And sometimes, based on how you devise your frameworks, your metrics aren't actually aligned with the customer's benefit, like the Amazon example. And we had many famous examples in Twitter's history which were the same. Like, we, uh, one of our key metrics that we always optimized around growing was DAU. And we had a, you know, a- obviously the rank timeline did wonders for, for growing DAU, and it was a great experience for many customers. But we often had features that would not lead to a good customer experience, and the team would just be blind towards leaving sort of like hostile customer experiences in place because it was good for metrics and aggregate. And the famous example of this is we had this toggle, which we called Swish very affectionately, but it was like a sparkle icon. Before you could switch between the rank timeline and the following, the reverse chronological timeline, which is still in the product right now, it was just a toggle, right? You would, you would press a button and it would turn your feed reverse chron, which very few people used as a percentage of, of our users. But you had power users who really cared about having a reverse chronological timeline, and we had the, you know, we took so many baby steps on the evolution of this product. But the very first baby step was, you pressed the toggle, it turns you to reverse chron, and then we would pull the rug out from underneath you and, and make the experience go back to the ring timeline after, like, I don't know, 24 hours or something like that.

   3. LRLenny Rachitsky52:03

      Oh, wow.

   4. KBKayvon Beykpour52:03

      And the reason th- the team felt strongly about keeping it this way is 'cause it was better for metrics. Why? Like, even if Lenny wants a reverse chron feed, we knew that he would spend more time in the app if we put him in the ring timeline. And, and this was like, you know, the, the number of debates that we had about this because, you know, the team understandably was like, "This is good for metrics." But at, at the same time you'd have customers being like, "I fucking hate this exper- wha- I'm telling you I want reverse chron. Stop randomly changing it for me." Um, i- you know, Instagram has, I think, gone through their own struggles with this as well. They, they sort of tiptoed their way towards ultimately giving people control and, like, the, the, the difficulty in making product decisions comes down ultimately to making these trade-off decisions, and y- you have to look at things through the lens of the customer. You have to balance that with, you know, what's driving the right business outcomes. And sometimes those things are aligned and sometimes they're not. And the answer isn't any one framework. Sometimes it's just good old-fashioned, like, judgment and product taste. Um, and so that's, that's where, like, I, my take is different. I don't think the issue is Jobs to Be Done, although I'm not, I'm not the biggest fan of Jobs to Be Done. The issue is just, um, having the right nuance and ultimately the right leadership to be able to weigh these things and see when your frameworks are not actually helping you make the right decisions.

10. 53:20 – 57:44

    Signs you’ve gone too far with a framework

    1. KBKayvon Beykpour53:20

    2. LRLenny Rachitsky53:20

       I think that's really important advice, and I, I love hearing the details of how you actually think about this stuff. Actually finding this balance is very hard in practice. I'm curious if there's something you could recommend or have learned about how to know when you've gone too far with a framework. Like, signs like you have implemented this too religiously and maybe you should be thinking a little more broadly.

    3. KBKayvon Beykpour53:41

       There's two, I think, simple and obvious ones. Um, one is, if the result of your framework is that subjectively bad decisions are being made, then something's gotta change, right? Like, assuming you have, the person who's making this assessment has good product taste, which is in and of itself subjective obviously, but m- my personal, um, view on this would be, like, in the role that I had, if I, if I saw that our organization was being incentivized to make decisions that, to some non-trivial degree at the time, were just bad decisions that I don't like as a user, I can't stand by as a, as a user or builder, then something's gotta ch- either, either bad judgment was made in following the process or the process was wrong. Or if that framework did not, didn't e- didn't even lead to the right debates, then that's, that's how you know, you know, like, either you have an incentive problem, um, and, like, the team did what they were incentivized to do (laughs) or, you know, that there is bad judgment and that's, that's a different problem obviously. But I think, you know, in the situation we found ourselves in where the team was, um, again understandably hyper-focused on driving DAU 'cause that was the strategy for so long and it left so little room for even, you know, taking ambitious bets that in the short term wouldn't drive DAU. Like some of our bets that I, I still to this day believe in-... hurt DAU in the short term. But, you know, you had to squint and believe over time would improve some metric, DAU or otherwise, right? Like a product like Spaces. In order for Spaces to be actually used, you needed to make sure that when Lenny starts a Space, that people would join, right? And how do you get people to join Lenny's Space when they're used to having an asynchronous, you know, feed of tweets? Well, you can send push notifications. You can occupy really prime real estate at the top of the app that lets people know, "Hey, Lenny's live right now." He's e- he's in the space. "There's people here. Come join." Guess what happens when you put a bar at the top of the app that tells people when they're live? You move tweets down. You move ads down. DAU goes down. Revenue goes down. And so if you have an organization that's like just hyperfocused on like the thing that matters is driving DAU quarter over quarter, then that doesn't leave enough room for nuance to accommodate new speculative bets that might hurt one metric, but over time have other consequences that are positive and, and beneficial, like enabling an entire new vector of content creation and conversation on the platform. I guess the other answer to your question in terms of like how do you know when the framework's not serving you right, when you start imagining and planning for a bunch of bets that the organization then sees ... it's like disincentivized to make successful, then something's gotta change. Either your strategy is just not the right strategy 'cause it doesn't abide by the frameworks, or the framework needs to accommodate the fact that actually we're gonna try some things that in the short term either might not, you know, show up as blips on our DAU radar or are gonna help some other metric that's important. And so that took us some time to, to get right. And we tried a variety of schemes to, to make that work, right? Like Community Notes, like Keith, the project that I was mentioning Keith, you know, started, we intentionally structured that like a startup. Like w- it was literally like... We made a seed bet on Keith and his team. And we were like, "You don't, like ... Don't worry about the OKRs. We're not gonna judge you on the basis of your OKRs." And there's some pros and cons to that. Like a lot of our projects worked that way. Fleets started that way. Community Notes started that way. And some other projects started more kind of like part of the core organization, um, because they were so intertwined with how we were, like the, the nature of the product, that it just made sense to i- i- ... Separating it was gonna be, um, it was gonna do more harm than good. Um, so you just have to figure out based on what, like h- based on how execution is going whether you're, you've got the right framework. And you've got to be willing to, to make adjustments when it's not working.

    4. LRLenny Rachitsky57:29

       That's actually really helpful. Kind of the two that I'm taking away here is how often are you feeling bad about the features you're shipping, like they're bad for users and you think they're bad for you as a user potentially? And the other is, is it keeping you from taking big bold bets that don't necessarily drive the metrics you're focusing on?

11. 57:44 – 1:00:41

    Lessons from building Periscope

    1. LRLenny Rachitsky57:44

       Okay. So before I'll let you go, I want to spend some time on Periscope. I don't know if everybody knows the history and story of Periscope. Basically, it was the biggest live video streaming platform in the world, and I imagine inspired basically every other social network to build a live streaming platform, Instagram Live, Facebook, TikTok, obviously Twitter. Uh, so I want to spend a little time here and see what you learned, and, and also just broadly consumer products. But, um, first of all, I hear there's like a story with Kobe Bryant and Periscope about him using it in some form. Could you share that story?

    2. KBKayvon Beykpour58:18

       Before Periscope launched publicly, which was in, it was in March of 2015, I want to say, we had a small beta that grew to maybe 500 people in total before we actually, uh, released the app publicly. And in that time, while we were still in beta, I was personally onboarding. I was trying to personally onboard every single user. Um, and I had sort of like a, a shtick that I, that I did, which was we'd get them in the app and I would start a broadcast. Um, we had a feature called Private Broadcasting that we basically built for, for this use case, which is someone joins, I'm gonna go live, and they're gonna join and I'm gonna show them how the app works. And, and, you know, we were spending a lot of time in our office. So, you know, Kobe, you know, one of, um ... Chris Sacca actually invited Kobe to the beta. Um, and so, you know, Chris connected us, and, um, I did a private broadcast, and Kobe joins. And I was like, it was like 10:00 PM in the office. And my, my routine was like let me just walk around the office kind of like talking through the mechanics of Periscope through the lens of this demo of like, "Hey, let me show you around the office. And like here's how the chat works. And you can tap the screen to send hearts. And if you want me to do something like go to, go to the room over there, just type it in the chat and it'll come up." And it sort of showed the, um ... One of the things that was unique about Periscope is, you know, it, it was a multi, uh, it was a one-to-many broadcast but still low latency enough that it felt like a FaceTime. So you could have the, the bidirectional communication between multiple viewers and a broadcaster. And so, um, I was teaching him how to use the chat and showing him the office. And he was, he was playing with it. And at one point through the end of the demo, he, he posts a comment that was like, um, "Why the fuck would anyone want to watch someone else stream live?"

    3. LRLenny Rachitsky59:56

       (laughs)

    4. KBKayvon Beykpour59:56

       And I remember like my heart sank. And I started like fumbling through like, "Well, we think it's, you know, cool," and like before I could even get the words out, he posted like, "I'm just fucking with you, bro. This is incredible."

    5. LRLenny Rachitsky1:00:08

       (laughs)

    6. KBKayvon Beykpour1:00:08

       And I just remember it was just such a surreal moment that has, I'll never forget. It was, I mean, Kobe's a legend obviously, um, but like to have that, to have him essentially troll me w- while also like putting a point on like what was cool about the experience, and that was like bidirectional, and like something he ch- commented on could cause the broadcaster to kind of like change their behavior or change their, change the experience. It was like a really ironic full circle of, um, showing off how the product works. But yeah, it's one of my favorite early Periscope stories.

12. 1:00:41 – 1:07:24

    Reasons why Periscope failed

    1. KBKayvon Beykpour1:00:41

    2. LRLenny Rachitsky1:00:41

       Oh, man. If I, if I kept doing that, uh, manual onboarding, I'd be like, "Sh- Who's coming?" Just always worried that someone else fancy is gonna join. You never know what Chris Sacca's getting, who he's talking to. That's amazing. Okay. So it's been about 10 years, I believe, since you sold Periscope, and about six years-ish since you've stopped running it as CEO, something like that. Now, as I shared, every single platform basically is doing this. They've added live video streaming.I'm curious if there's anything you learned about just competing with these major platforms.

    3. KBKayvon Beykpour1:01:14

       Well, the, there's, there's a few reasons why, why Periscope failed ultimately, and why we shut down the app. I mean, obviously, P- Periscope, the technology and the mechanics still lives on because you can go live on Twitter. Um, you can watch live broadcasts on Twitter. You can do audio conversations on Twitter, and all of that is the Periscope stack, still there, which is like... It's awesome that the legacy lives on in a different form factor. The reason that the Periscope app failed, i- it really comes down to a few things. One, we did not address the core problem that retention wasn't good. Our poor retention was mapped by just an incredible surge in top line user growth. Um, you know, and, and, and it... And for Twitter it was very, um... Or for Periscope, it was, it was interesting because like every month or two, we would kind of blow up in a new market that would just bring along an incredible surge in usage. Um, you know, we, we blew up in the U.S., we then blew up in France, we then blew up in Turkey, we then blew up in the Middle East, and you had these, like, incredible surges. But underneath that surge, like, the core product had retention issues. And it was... We ultimately just did not... We didn't spend enough time prioritizing addressing those. And in fact, we shipped product changes that made those retention issues worse.

    4. LRLenny Rachitsky1:02:24

       Mm-hmm.

    5. KBKayvon Beykpour1:02:24

       Um, compounding that was the fact that one of the theses behind our acquisition, uh, the Twitter acquisition, was that we would leverage the scale and the community and the product mechanics of Twitter to, um, make the p- the product grow faster and also become more durable. Um, and I think that sort of connects to one of the reasons why I feel like one of the, one of my learnings and one of the things that we knew but just failed to execute on, was that I still am very skeptical that there can be a, um, a consumer product that is just focused on live video. Like, a generalized, uh, synchronous live video application for short form video I don't think can be durable on its own. I think you have to surround that product with enough features and capabilities to allow a community and an ecosystem of users to be able to stay in touch with one another asynchronously and synchronously.

    6. LRLenny Rachitsky1:03:10

       Mm-hmm.

    7. KBKayvon Beykpour1:03:10

       This is why a lot of the other products that you mentioned that incorporate live capabilities and, and we're shameless about copying, you know, our, what was working about Periscope, they're, they're surrounded by a scaffolding that lets people also stay in touch with the other, with, with each other asynchronously. Like Instagram, it's an asynchronous product that has synchronous features like Live. Um, same with TikTok obviously. Um, and I think we were in this position where it was a live only product, right? Like, you know, you are con- you're connecting with your audience and having a great time when you're broadcasting live, but you're not using the product, um, to keep in touch with that community when you're not broadcasting. And how often throughout the day would you, would you broadcast live from your phone? And mind you, this is different for products that are li- like, live, 'cause there are live consumer products that focus on a specific vertical, like Whatnot for selling or Twitch for, for gaming, that have very different properties that make it more durable as a standalone live product. But, but, but Periscope was really in this sort of consumer, um, generalized live streaming from a phone land. And I think it was just not durable to have the product be live only, and the time it took us to integrate with Twitter was way too long. And there was reasons for that, that, um, you know, come down to just how distracted Twitter was with its own roadmap and, uh, refining the core, and they just had other, other fish to fry basically. And that, all of that leads us to competition, because at some point Facebook woke up and decided, "This is cool and we need to go build this." And, you know, I, I obviously wasn't, wasn't there on the inside, but, like, legend has it, Mark says, "Hey, like, three- you 300 people stop what you're doing. Go, go basically, you know, make Live exist in our product as a first-

    8. LRLenny Rachitsky1:04:55

       Oh, wow.

    9. KBKayvon Beykpour1:04:55

       ... experience."

    10. LRLenny Rachitsky1:04:55

        I didn't know that.

    11. KBKayvon Beykpour1:04:56

        Um, and you know, if you have that level of organizational effort put on building something that by the way isn't even... Like, there's, there's no... You don't have to spend any time wondering what the product looks like. Just go, like, copy these features basically and make them, make them work. Um, and they did a lot of other savvy things too from a partnership standpoint. Like, you know, we had a lot of prominent streamers that ranged from kind of like influencer or just creative people that became known on Periscope all the way up to celebrities like Kevin Hart and, um, others who were like prolific Periscopers that Facebook just went and bought out, and they were just like, "Cool. We're gonna pay you, like, a bunch of money to stream exclusively with us." So, they kind of hit us from all sides. They had the entire company put their effort towards building Live in a way that was cohesive in the core product, first with Facebook and then, and then Instagram, and then also attacked it from the, from the creator side as well. Um, and, um, we were too slow, and it was, it was, you know, very painful to think about 'cause, um, it, it was... You know, like many other, uh, insights that Twitter had early, Twitter had the right insight but botched the follow through. I- I'm not pointing fingers. Like, I, I blame myself for that just as much, um, but it's, it's... There's a pattern where Twitter is really great at spotting meaningful consumer behavior changes, right? Like, they spotted Vine and acquired Vine, botched it. Spotted Periscope, botched it. Um, so, and spotted Instagram by the way before, you know, Facebook tried to buy Instagram, um, Twitter was trying to buy Instagram, and there, there are other reasons why that didn't fall through, but it's just, it's interesting to me. It's one of the interesting s- um, aspects of Twitter's history. They were phenomenal at spotting meaningful changes in consumer behavior, consumer social behavior and, and actually, like, putting their money where their mouth is in terms of, um, trying to follow through on bringing those, those bets in-house but then botching the execution. And so that was, that was one of the things that was really motivating for me when I, uh, I was in my role, um, leading product at Twitter, was like, I didn't wanna... I didn't wanna make that mistake. Um, and, um, you know, we didn't end up buying anything as... We didn't end up buying anything like Vine or Periscope and, and keeping the product in-house. We obviously bought lots of small acquihires but, you know, we did, we did obviously, um-... uh, have a bit of a, a, a story with Clubhouse that ended up with us building, um, you know, building Spaces and competing with them. But anyway, that was a long, rambly story. Hopefully that answered

13. 1:07:24 – 1:12:05

    The challenges of implementing video at Twitter

    1. KBKayvon Beykpour1:07:24

       your question.

    2. LRLenny Rachitsky1:07:25

       Yeah. There's, there's so many... Again, it's all these fractals of th- threads I want to follow and ask about. Real quick on the Vine and Periscope point, I was gonna ask this. Uh, Twitter, as you said, had the opportunity to win in video in so many ways. Vine was amazing, killing, everyone loved it, and then it's like, yeah, then it fades away. I guess you already shared a lot of challenges Twitter had with executing shipping, uh, sacred cows, things like that. Is there anything specific with video? Was it just like, "Oh, this is not actually a huge priority and we're just gonna kind of ride it for now," and that's why it didn't work out? Or is there anything more to that?

    3. KBKayvon Beykpour1:07:53

       No, it's e- it's even more, like, pathetic than that. Um, 'cause I think Twitter did believe in video, but it did... made this classical mistake that we, we also unfortunately recreated with, with Periscope, which was they had the insight around short form video, they bought Vine, they then competed with Vine internally. So Vine was a separate organization within, within Twitter, separate office obviously in New York. Um, and then Twitter, rather than, like, integrate it holistically into the product and, and sort of pour gasoline on it, um, they, they built a native Twitter video feature that was like a different stack, different team. And it, it became what you think of as Twitter video now. It's like, you know, the simple act of uploading video and of all the, uh, professional, um, video tools, um, called Media Studio that let publishers like ESPN put content. Like, all of that was basically built as a separate team, separate organization, separate product, that was fundamentally competing with Vine. You had, like, two visions for short form video, um, that were manifesting. And that's, like, that's, like, the quickest way for, for like, uh, for things to get messy and for, uh, o- of course, like the, the, the separate startup team is gonna get, um, um... You're not gonna be able to make good on the vision of buying the company and integrating it in all the right ways if you compete with it internally.

    4. LRLenny Rachitsky1:09:11

       Mm-hmm.

    5. KBKayvon Beykpour1:09:11

       Um, and we had a s- you know, a similar thing happen with, with Periscope. Um, you know, we were very focused on, on Periscope with, you know, separate organization, separate structure, separate app. Periscope at the time primarily v- was focused on UGC live video. Um, so user-generated content being streamed from a phone. Twitter then, you know, decided to get in the premium live video business, um, very famously with the, you know, acquiring rights to Thursday Night Football, the NFL. Guess what happened? We competed internally. Um, rather than, um, have a cohesive technical and product vision for, for how to embrace live video across the spectrum of UGC and premium live video, Twitter, you know, put a separate team in charge of premium with a separate product, separate technology stack. Um, and, um, so you had, like, two ways to manifest live video on the product. There's like UGC live video, which was kind of like awkwardly not even really implemented well with Twitter at the time, and then premium live video, which had totally different UX, total different team, totally different architecture. And by the way, like, the company put in, in a tremendous amount of energy and investment in talking about Twitter being a place to watch the NFL, and meanwhile you had this burgeoning UGC ecosystem. So this was like, we're making the same mistake all over again. Now luckily, with... in the Periscope case, with a lot of persistence and impatience and table pounding, we eventually fixed that mistake. But we wasted a tremendous amount of time. Um, right? It was just a lot of head-butting and politics, and, um, and eventually it took us a lot of time to technically reintegrate things together. And, and now it's clean and awesome, right? Like, ESPN can go live with behind the scenes content at Wimbledon and it's, like, the same technology stack and the same user experience that powers Lenny going live from his, from his iPhone. Um, but I think that was, like, one of the reasons why... It was an, it was an example of failed execution that ended up wasting time, resources, and just leading to a subpar product experience that other companies I think have, have avoided making such mistakes. Facebook being a prime example. As, as frustrated I am with them as a competitor for having, you know, um, really taken over the use case for live video, got to hand it to them. Brilliant execution. Um, I have a lot of respect for them. Um, so, um, we made sure to not make that mistake moving forward.

    6. LRLenny Rachitsky1:11:33

       Mm-hmm. I imagine there was reason for that partly... I imagine this calcification of just like, "We can't get shit done. We just need to start a new team and do this thing." Like, I imagine it always comes from like, "Oh, this makes sense," and then you realize, "Okay, this was a terrible idea," down the road.

    7. KBKayvon Beykpour1:11:47

       Yeah. It's, it's that and also leadership. When you, when you don't have unified leadership around these things-

    8. LRLenny Rachitsky1:11:51

       Mm-hmm. Mm-hmm.

    9. KBKayvon Beykpour1:11:51

       ... um, you end up-

    10. LRLenny Rachitsky1:11:53

        Yeah.

    11. KBKayvon Beykpour1:11:53

        ... making decisions that are in conflict with one another. There, there just needs to be a highly opinionated person at the top that avoids that kind of, um, that kind of, um, messiness from a product and engineering standpoint.

14. 1:12:05 – 1:17:58

    Copying ideas in good taste

    1. KBKayvon Beykpour1:12:05

    2. LRLenny Rachitsky1:12:05

       You mentioned Clubhouse. So I think what's interesting is, one, many people copied Periscope as a product. Uh, I don't know if you'll describe it as copying, but it feels like Spaces, very inspired by Clubhouse. Do you have a, just a, uh, current philosophy on when it makes sense to be super inspired by another product and build it into your existing product versus like, "No, you should not do this?"

Episode duration: 1:35:26