Behind Duolingo's Multi-Billion Product Strategy, Scaling to 100M users - Cem Kansu

0:00Google never had this problem. During my time, it didn't. And this was a new environment for me to walk into. In a way, I was excited about having to figure this out, figure out how to monetize. I was tasked with that. But also, if I don't, we might not make it.

0:18What would happen if we just told the truth? Welcome to TruthWorks, where we dig into the nitty gritty of leadership and work. And what needs to change? I'm Jessica Neal. And I'm Cati McCord. Our journey together started in HR, but trust us, it's evolved into something wild, honest, and well, a bit rebellious. So throw out the handbook. We're here to redefine rules to work for us, not against us. Let's dive into another episode.

0:56Welcome to another episode of TruthWorks. I'm Jessica Neal, your host. And today I have a very special guest. So I'm just excited to talk to you about all the things. But we have Chim Kansu. Or no, you actually told me to say Kansu. And now I said... Kansu is better. Kansu is better. But I said the first name right, which is good. But Chim is someone who knows how to make learning addictive in the best way possible. He does this in the best way possible. So he is a longtime product leader and he's currently at Duolingo and he has helped scale this app into like a cultural phenomenon, right? Like everybody does Duolingo, including myself and my six-year-old son who is learning Korean.

1:46So my son's half Korean and his dad doesn't speak to him in Korean. We're supplementing with Duolingo. We're happy to help either way. Well, thank you. Jim, welcome to the show. I want you to tell people a little bit about your background because before Duolingo, you had a whole life. You're not just defined by Duolingo, but you were at Google. You were at Jawbone. You've been a product leader. Tell us a little bit about your story and how you got to where you are. Definitely. And I'm very happy to be here. When I've listened to a few of your episodes, it always felt like an incredibly honest conversation.

2:27That isn't just about product. And this is a new territory for me. So I'm excited to test myself out talking about all these topics we're about to enter. But my story is I actually use almost the same words when I talk about my Duolingo time, which is I do feel like I grew up at Duolingo. So I now am the head of product here at Duolingo. I lead the product management team and I'm involved with all of our product decisions. I have been here nine years. I just passed my nine-year mark, actually. Thank you. And I started in 2016 when Duolingo was about 55 employees and I was the third PM in the company.

3:07And I have worked on all of our revenue products for about five to six years. I built our subscription products. That really took a life on its own, defined our business model. And then for the past roughly four years, I've been leading the product team. And I've seen Duolingo through various phases. Duolingo was 50-ish people, and now it's about 1 ,000 employees. So things have certainly changed and grown. And then before Duolingo, right before I was in business school, during business school, I had a short stint at Jawbone. I don't know if people remember anymore. Actually, I know they probably don't, but I remember Jalman because what was the name of the CEO?

3:49Oh, my God. Now that you asked, I'm blanking. Because I got introduced to him and we really liked each other. I was still at Netflix. Let's see, because it's driving me crazy. I need to look this up. I'm looking it up right now. Okay. Hussain Rahman? Hussain Rahman. There we go. Yes. Yeah. Yeah. So this is so long ago. But anyways, he and I, I think it was Andreessen Horowitz. Okay, yeah. Yeah. And they were like, you got to go meet him. And I went up to San Francisco and we hung out and we literally like we're supposed to spend like 45 minutes together. I think we spent like five hours together.

4:28Wow. And I really liked him. He would have really long meetings. I do remember that. But anyways, he gave me one of the jam boxes because that's what Jawbone. Such a beautiful device. Yes. And I still have it today. It still works. I still have it too. Yeah. It was really good technology. So for those of you that ever had a jam box, Jawbone was the company that made it. But I think that company, it was just a hard, like, it's a hard business, consumer electronics. Hardware is hard. Consumer hardware is really hard. and Jawbone, to their credit, was an amazing design company. They would build, they were in a way very similar to Apple.

5:09They would kind of come into a category, build a very well-designed version of that hardware. They did it with Bluetooth headphones and then fitness wearables and also the Jambox, the Bluetooth speakers. The challenge with hardware though, there's a lot of race to the bottom with these devices. So anyway, I think eventually that was the difficult turn of events for Jawbone where they would try to sell the well-designed version for$200 where someone else was selling it for$20. That's doing the same thing. Right, right. Yeah, and it is no longer. It is no longer, unfortunately. Unfortunately, yes.

5:43And I was, for the record, I was only there for an internship, but it did teach me a lot about design. And that's kind of eventually what led me to come to Duolingo, which is another company that cared a lot about design. Yeah. So it had a lot of formational impact on what I chose to do. But anyway, so that was during my time in business school. And then before then, I was at Google for about four years. And I was not actually a product person there. I was doing business development for Google's ads business. So my career switched from business development to what I really wanted to do, which I realized over my time at Jawbone, products.

6:19And then if you go way back, well, I'm originally from Turkey. That's where I grew up. And hence my name spelled with a C, but pronounced like a G. That's the Turkish C. Well, you're a gem. Yes. Yes. And no one can deny it. Okay. Especially Duolingo. So I want to talk a little bit about like you come into Duolingo. It's what? 55 employees. And I'm guessing the product is pretty in its infancy, if you will. Right? Like not very mature, not the Duolingo it is today. So when you walk in and you're part of this team, like what did you do? What was obvious to you? What wasn't? I think if I look back at that time, it's filled with excitement and anxiety and at its maximum.

7:09How many, well, I guess you didn't have subscribers then, but how many like, do you call them learners? What do you call them? That actually that term learners solidified over time. At the time, we would just say daily active users or daily active learners. We had 3 million daily actives when I started. Right now, we have about 50. So we've grown a lot. And the product, you are right, was way earlier in its evolution. I was really excited because there was just so much to do. Clearly, the company's mission was quite well-defined. And that's part of the reason, actually, the majority of the reason I wanted to help this mission, which is provide access to education.

7:46That's what Duolingo stands for. It broke through, if you go way back in time on what language learning products look like, like the Rosetta Stones of the world, these were really expensive CDs, basically, is what they were. And it was a kind of inaccessible way to learn a language, especially if you think about learning English in certain parts of the world or like in countries where English is not the dominant language. it opens massive economic opportunities. And Duolingo's mission was built around trying to help people have access to education that's easily accessible through mobile or web, and it's free to use.

8:22So that was really exciting to me, and that fueled my excitement coming in. The anxiety part was also quite real, because when I started, we were very much a startup, and we were making zero revenue. So if you're an investor looking at this business, you would quite be concerned in the sense that will they ever figure this out and um the way a startup operates and communicates is we would say things like okay we have this much money in the bank and this will last us 14 months so runway is 14 months and that i had never worked in an environment where it was like okay we don't figure it yeah exactly google never had this problem or at least during my time it didn't and this was a new environment for me to walk into uh because i I also felt in a way I was excited about having to figure this out, like figure out how to monetize.

9:11I was tasked with that. But also if I don't, we might not make it. And that was a mix of emotions. Yeah. Yeah. So then you have to have all this pressure. You have to figure out how to monetize. Do you just start testing? Is that what you started to do? At its core, yes. We started testing and in hindsight, everything looks super clear on it should have been freebium. Of course, we want to provide education for free and monetize the bells and whistles. But at the time, it really wasn't. Actually, the common advice for any education product, honestly, it still is, is you have to put everything behind a paywall.

9:48So meaning if you're monetizing educational content, there's really not a lot of creative business models. Lock it behind the paywall, you charge a subscription. Duolingo wanted to be on the other side of history where we will provide educational content for free, but then we'll figure something out on how to make money was kind of the current state of things. So we started testing. We started testing ads. We started testing different in-app purchases, and we started testing a subscription model, which honestly, I don't even know why people bought our first version, but people did, and that gave us a lot of fuel because the subscription, we had very minimal ad load.

10:23It removed ads. it gave you offline access, which most people didn't use. And then there was a third feature that said you get to support Duolingo. And a lot of people bought it. They love Duolingo and they wanted to contribute. And they gave us a lot of fuel that we thought, okay, even with a bare minimum set of features, this is a good business model that's working. Let's test more. And as we experimented, we started seeing things that are working. I will call out the, of course, there was a lot of uncertainty and we figured out as we tested and got data. The biggest challenge I felt like we had at the time was convincing internally that this was the right thing to do because a lot of people until that moment in time that joined the company had this idea that Duolingo is free forever.

11:08And anything that goes against that felt like we were breaking an agreement we made, even though we knew we had to survive as a business. I know. I was at Coursera. So I left Netflix for a brief amount of time and then came back and I was at Coursera and we had the same challenge and because we gave it away for free and everyone had joined for the mission. And then it was like, but we have to make money, right? And it was really hard to navigate. It is really hard. How did you guys do it? I would say I think the approach that ended up working is one, we, of course, our CEO, Luis, he was quite aligned in testing and figuring out what works.

11:52So we had to at least have alignment to test. But as we tested, we were able to alleviate some of the worries on this. Anything we do with Monzation ruins our product was the general, like, let's not touch this vibe of the company. But the data showed that users didn't run away, actually. Retention was fine. Our user love was fine. And we were able to quantify what that means rather than this dark cloud that was pointed at as monetizing Duolingo. That's one. Two, I think it became clear also as time went on that if you want to actually maximize your impact, as a company, it does actually help a lot to have your own revenue.

12:31Otherwise, you could become a nonprofit, but it was quite clear eventually that if we became a nonprofit and decided not to monetize, the impact we're trying to have on providing access to education would actually be quite handicapped. So I think that was, one, the kind of the emotional appeal of monetization actually helps us. And two, backing that up with the data we're seeing, which is like our users are not running away. They're actually happy to pay for a certain set of features. Yeah, people like to pay for things that are useful for them. Yes, turns out. Turns out they do. Having worked for a subscription business for most of my career, subscriptions really work.

13:11They really do. They really do. I mean, I think you said it, and it's not rocket science. If people enjoy using a service, paying a continuous subscription is a quite logical way to get that service. Right. Yeah, absolutely. Okay. So you figure out this subscription thing. You kind of unlock monetization for the company. You were at$3 million. Now you're at$50 million. I'm guessing your job has changed quite a bit because how many people were you leading when you were at$3 million? Zero. I was leading myself, if I could even say that. I was managing myself, yes. And unsuccessfully, probably. But yes.

13:56Yeah. So my job evolved quite a bit. It was, again, this is, I love that you say this because I say this light. I really feel like that's why I grew up at Duolingo. It started with, I was responsible for a portfolio of product features and certainly no other humans. And then I moved on to what we call a team lead role. That was kind of my first step into leading at least a group that wasn't direct management, but it was more deciding the kind of the direction of a team. And then that eventually turned into managing some other product managers. And then that turned into what we call like an area, which is a group of teams leading a group of teams direction and managing the product managers on those teams.

14:33And then that kind of grew as the company grew that the scope of that grew. And then now, as I guess, for the past four years, I'm managing an org, which is the product management org. So that's certainly all of these stages looked very different. And I would say there was a lot of trial and error on things I messed up and I continue to. So that's been the journey, honestly. Well, you know, I think great leaders are self-aware. At least we have that. Yeah, that's self-aware. So how many people work for you now? We have 55 product managers. Okay. That's a lot. That's a lot. So you went from yourself to 55.

15:12So that's pretty good. So culturally, I think this is true, but you can correct me if I'm wrong, but you all are known for like ruthless focus, like focus. We had a similar DNA at Netflix, but tell me a little bit about that and maybe the pros and cons and how you stay focused? I think we do care a lot about, I think the exact operating principle we have written and communicated and honestly at standby is prioritized ruthlessly. So it's certainly ingrained in how we make decisions. And it's not one of those operating principles that sits on a wall and looks cool. It really, we say this a lot. But you don't just say it, you do it, actually.

15:57We say it, exactly, and we do it. I would say maybe two things. One of them is if you have such a grand goal to solve. For us, it is one, teaching you a language fluently. And now we've actually expanded our problem set to teach you a language, math, music, chess. So these are very hard subjects to teach if you're starting at zero or anywhere and to get you to be fully fluent in any of these topics. We just have a very hard problem to solve. It is easy to get lost in the details. So it matters so much to pick the one thing that matters the most in making progress. So what turned out for us as a company, at least in the early days, now it's branched out as we grew.

16:35But the one thing that mattered the most for succeeding in teaching a language was making it fun. So when we came to that insight, majority of our efforts, and still today, honestly, is focused on making teaching fun. with self-taught learning, as opposed to, for example, if you need to sit in a classroom for an hour, that's a very different environment than self-taught learning or self-chosen path of learning. You're competing with the most fun things on your phone. That could be Netflix, that could be Instagram, that could be any game you can play. So unless it's really fun, we don't have a big business.

17:11So we spent a lot of time basically investing in and prioritizing how to make things fun. And that's why Do Ango is the way it is today. And it's funny, we get criticized for it sometimes being like, oh, this is a game. Well, it's meant to be as fun as a game, but it's actually going to teach you the thing you came here for. And that was an example of, for example, ruthless prioritization. And the kind of the fundamental insight behind it is you can learn the language in many different ways. You can learn a language by reading a textbook, but nobody does that because that's not fun at all. It's not fun.

17:44nobody wants to learn that way actually i did rosetta stone back in the day and i hated it i tried to learn korean and it was so hard and it was like it's so hard i don't know what the method of rosetta stone is called but you know it would show me like four pictures and it would say the word in korean and then i would i don't know what they're saying but it was like there's a pic you know because i don't know korean and there's a picture of a horse a picture of a frog, a picture of a duck and a picture of, you know, a snake or something. And then you just like, and then you're wrong. And then you finally figure out that that, that word actually is horse.

18:23You know, like, and I was like, this is not, I can't, this isn't working. I told my husband, I said, I don't think I'm going to learn Korean. You gave up eventually. We had just gotten married and I was trying and I was like, I can't, I can't do it. Well, we talked about business models. I think this is a challenge of that business model because you already had bought the cds so there's not a lot of incentive for continuous improvement on how they teach korean in a way so yeah because it was like well back in the day when we had cds you know that dates me well thank you um okay so so we were actually pretty ruthless at netflix too and i think you know we our goal was to entertain the world right and so in order to entertain the world, you have to have the best content.

19:12And the content isn't necessarily what you think is entertaining. It's what the world thinks is entertaining. And then you have to be able to find things easily to entertain yourself. The uploading of the video can't take forever. It's got to be seen, all those things. And so I like that. But I think that a lot of companies don't have that focus, right? They don't. And I think it obviously is great for us. Obviously, it's helping you. But why do teams, organizations have such a hard time focusing? It's a great question. And I don't know if I have the perfect answer, but I would say maybe two things are hard.

19:59One is it's always easy in hindsight for me to even tell the story and say, oh, yeah, we had this one insight, we built product around it, and it worked really well. But I think getting to that was a quite messy process, meaning we were testing and investing. It's also not like we did one thing, but it was one insight that ended up changing the trajectory of Duolingo's growth. And I think it was, it's not simple to get that insight, and you end up chasing a lot of dead ends until you get to that insight. And I think that I feel like the formula that ends up working is really fast iteration cycles on your hypotheses until you land on the one or two or a few that is really going to drive the business.

20:38But I think getting there is always going to look messy. It's going to take some time. But I think the way to get to the end state, which is clarity on what really is going to make a difference in the business, happens through testing various hypotheses. Whether that's honestly a product challenge, an org challenge, a business challenge, whatever it might be. It really requires trying. Action really gets you the insight rather than whiteboarding and writing strategy docs. I feel like people make the mistake of doing the work that is not the actual action, but the discussion of the action. And that really doesn't move your thinking forward.

21:12That's kind of what I believe, at least, is the solution to a confusion. I think you're right. As I'm thinking about it, I think there's a lot of discussing and analyzing and thinking. And sometimes there's not a lot of doing. And I think a lot of humans, but leaders and organizations are afraid of risk and failure. We were very, I mean, we were a huge A-B testing culture at Netflix. But we also were like, let's just try it and if it doesn't work we will know it doesn't work right like there's no if you don't try you don't know and when you're in these businesses like like i i was and you are nobody knows the answer you're making it up as you go along you don't know you gotta try i think that's why the building a culture of experimentation is more important than the specific experiments you decide to run because it compounds, right?

22:17It's like, I think we can debate, for example, the best way to run a subscription experiment, but it's way more productive. We build a team that can do that at large numbers that can run good hypotheses. And in a span of six months, rather than us debating what's best, we will cover way more ground by letting the system kind of experiment. Yeah. So how, as a leader of a system of experimentation, how do you get your team to to be on board with that how do how do you foster it i love this question and i i feel like we love talking about this a lot internally at least one of them is make it easy like if remove any friction to for anyone in the company as much as possible at least let's take product experimentation which i i also now my new favorite thing is org experiments which is which is something i enjoy now making org changes not like concrete and stone but more like an experiment.

23:09But anyway, let me get to that. I want to talk about this. I like that. This is right up my alley. I assumed. But I think one is reduce friction. For us, we built our entire A-B testing tools ourselves for our use cases. And we made it quite easy for any team to go run an experiment. So that at least makes it easy to get started. Two is encouraging teams to be experimental. As long as we align on kind of what designs we're showing the user or what change we're going to make as long as we're aligned you can't change the logo exactly let you don't change the logo or we have like a bunch of design principles like duo has to look a certain way he needs to be unhinged whatever it is there's many of these as long as we're aligned on on the design guidelines and what we're gonna do teams should be running many experiments and honestly it's a bit of a social currency culturally to run a team that runs 100 experiments is kind of cooler than the team that runs 70 is also the culture in a way because it's a good thing.

24:11And I think that's maybe kind of the culture thing, which is encouraging teams to be experimental. Yeah. Is there something that you also do around failure? Because a lot of the experiments fail, right? Half of them fail on average. Yeah, like you do a hundred of them and, you know, half of them aren't great. And some of them are just okay. Like it's like, you know, a little bit, but not enough. I think it is completely expected that half of experiments will fail. So, and you mentioned failures. I feel like we don't blink an eye when an experiment fails. I think that the, at least for product experiments, the lens we have collectively created is if the experiment failed, the thing, the only thing to do is look at why.

24:58and then if there's a new experiment we can run that would fix the why on why it failed. If not, throw it out. There's nothing to do here. Let's move on to the next hypothesis. I feel like that's rather than just, I mean, celebrating failure is certainly important because trial is important, but I feel like our lens is rather than celebration or not, it's what can we learn from this? And once we've extracted that, let's move on to the next. So to sound like maybe like a completely always on the grind mentality, I feel like that's how we view it. It's like, okay, what did we learn? We ran this thing.

25:28The insight was maybe the color should be red, not green. Okay, let's try three different shades of red. That's now the new experiment to run. Right, right. Yeah, I love that. Okay, so let's talk about these org experiments. I am curious, tell me. Well, I've realized, I think, making org changes is always painful, especially for someone that is responsible for the health of an organization. Doing reorgs, nobody loves to do. However, this is kind of a mental trick in a way, which is if you call something a pilot that we're going to try for a while, or if you call it an org experiment, everyone is, number one, way more open to it.

26:05Two, it gives you the chance that if it doesn't work out, because org experiments also half the time don't work out, it's a very easy reversion because you never said, I am putting, you know, A person here to lead B. You never committed to it. So it makes it way easier to try something. And that's why I now I propose a lot of org changes as experiments. Not everything, of course, but a lot of things. And it honestly gets you to try more things, to learn more from those experiments. Rather than I think when I first honestly stepped into my job, every org change felt like, okay, I got to convince 18 people.

26:40They all have to say yes. And then we are going to commit. We're going to do it next quarter. It became a whole thing where most things can be just tried. And if it works, we make it permanent. If it doesn't, we revert. Yeah. No, I love it. And it's actually how I approach things at Netflix. And I remember, you know, back in the day and, you know, Patty was with me too and we were doing these things. But, you know, when Patty and the team and I were like getting rid of vacation, right, it was like, you can't get rid of vacation. Nobody's going to take vacation. It's like, you know, and everybody was sort of freaking out.