How Discord Completely Redesigned its Mobile Apps

This wasn't just about Discord's mobile app. This was just about Discord in general. If you were to ask somebody off the the street, they might say something along the lines of this. Yeah, I downloaded it once because somebody asked me to join or I clicked this link, but it was really complicated, so I deleted it. And we kept hearing that over and over again. You work on like small iterative things to optimize growth. Well, at some point in the cycle of every growth team, you hit a wall where it's like, we've optimized everything that's possible here. we need to change basically where we're optimizing from. One principle that we set out with is we don't want to build anything that will alienate our existing users. So the idea was like okay take the next group of people as a supererset of the current set where it's like whatever we build has to fit the full set. Welcome to Subversive, a podcast

dedicated to sharing stories from the best consumer subscription apps in the world. We'll bring you lessons for how to grow your consumer subscription business, including insights and inflection points that led to exponential growth from leaders at category defining companies and innovative startups. Let's get into the show. This episode is brought to you by RevenueCat. The world's best apps use RevenueCat to power purchases, manage customer data, and grow revenue on iOS, Android, and the web. Join more than 100,000 apps that are already using RevenueCat to grow their businesses and track more than $10 billion in spend, including category leaders like OpenAI, Notion, and Ladder. Learn more at revenuecat.com/subversive.

My guest today is Franchesco Pitzy, former head of core experience at Discord, who oversaw a complete overhaul of the company's mobile apps that expanded Discord's appeal beyond gamers to a much broader audience and made it better for on the-go usage. In this episode, I talked to Franchesco about the key product changes his team made as part of Discord's mobile app redesign, about the impact these changes had on Discord's users and business metrics, and about the company's unusual decision to layer subscriptions with premium features on top of a free social network. All right, welcome Franchesco to the Subversive podcast. Really excited to have you on today to talk about Discord. We actually met I don't know several years ago now when you were at Photomath and I was at Quizlet and I remember comparing notes with you on building edtech apps. Uh but today we're

going to be talking about Discord. So why don't you start by just telling us a little bit more about your professional background and how you made your way to Discord and maybe a little bit about what Discord does for those who don't already know. >> Yeah. Uh well it's great to be here. my professional background. I studied to be a software engineer and I found out pretty quickly that uh I it was not my calling, not the thing that I was good at and not the thing that I was particularly interested in either. I found that like the decisions behind how a product gets built were the thing that felt more important to me than h than what is actually built. Um and that led me to finding product management. It led me to a career in consumer software that spent or spent basically all of my career. um started it uh Mozilla working on the Firefox team in actually as a growth PM and then worked my way over

time Photomath and then Discord and more recently I was at a company called Out School which is a big edtech marketplace u and then Discord for people that don't know it's basically a chat app but primarily for gamers. It started as kind of it actually started as a game studio. Uh and what they realized they built this game that had really awesome chat features and they said at some point um one of the first engineers his name was Stan. I actually worked really closely with Stan. So we'll talk a bit about some of the work we did together. But Stan realized like hey there's this awesome stuff that we built as like infrastructure within this game. I think I want to spin it out. Like maybe we should spin it out and create a platform that's just about chat. And he pitched it to Jason who is the CEO. Stan actually wasn't a founder, but was is considered a co-founder of the company because that was the start of Discord.

And so now it's it's effectively you can think of it like Slack. Um their servers and and group DMs and things like that. >> Well, it's funny you bring up Slack because I feel like it had a very similar origin story. Uh didn't they start as a game and then pivot into messaging as well? So it's it's funny that both those companies follow that trajectory. U let's jump into Discord in

a moment, but before we do that, you mentioned that you started as a software engineer and transitioned into product management, and that's something that quite a few people end up doing in their careers. What drove you to want to make that transition and and what have you found you love so much about product management and product leadership? >> It was that realization that, you know, I was spending a lot of my time think and it's funny because I'm going to describe to you a problem that I think is is dying. It's like no longer existing. So I think if I had if I had shown up in my career today I might have made a very different decision but at the time it's like okay as an engineer you basically get these tasks and at the end of the day you're required like the thing that is the core of your job is to figure out how to take a thing that you want to build and build it in the best way possible with the least bugs the

fastest right the most scalable that as a core responsibility seemed less interesting to me than like what is the list of things that we're going to build >> that felt way more way more interesting to me because it was like, you know, the first task I get that I'm like spending a lot of hours on that that that I don't feel is right. It's like, okay, I would rather engage in debating that. And of course, you you you have things like that you don't enjoy with any job. But I realized actually the first job that I had in tech, which was at Mozilla, I saw that there's this role called product management. I had never heard of it. And uh and all of a sudden it's like, wow, okay, maybe that's what I want to do. um because I find myself in these rooms doing the thing that they seem to be doing in some way.

>> Yeah. It's sort of if I'm understanding you correctly, it's the distinction between focusing on the why, the what, and the for whom as opposed to the how. And engineers need to be very good at figuring out the how. But if they're working with a good product manager, then hopefully a lot of the why, the what, and the for whom has already been determined for them. And so it's just a different set of skills and and requirements. Um, it's such a great transition into the primary topic we want to talk about today, which is this major app redesign that you and your team did while you were at Discord. But briefly before we get there, so you you were a product manager at Photomath. And then when Photomath got acquired by Google, you ultimately were the head of the Photomath product within Google is my understanding. And so walk me through what that acquisition looked like and then what made you decide to make the

leap to Discord. Yeah, I so I left Photomath before the acquisition, but um I kind of I was there through most of the story of Photomath being built. Um I was at the time that I joined Photomath, I was a PM, a growth PM at Dropbox, kind of a new grad um finding my footing in the industry and I got reached out to by a recruiter and I talked to this team. They were in Croatia of all places and they had it was a serial founder. Um he had built the first ISP in Croatia in like the '9s and he built this app that was kind of like a side quest for the main product he was building. They were building like B2B OCR for like document validation. Think like Airbnb needs to validate a document for like uh ensuring that this is actually the person that's saying they're signing up. That's what they were doing. But what he realized is like he's a dad and he he had uh math homework with his kids and he he you

know it's been years since he was an engineer. It had been years since he had studied math. And so he said, "Look, like I think I with this OCR tech that we built, like I think I could pretty realistically like marry together a very basic math problem solver and like OCR." >> They built this app. It it kind of went viral. I showed up. I was employee number like 20, but it was still very early. It was the second employee in the US. So they were all in Croatia other than this. They had just raised like a seed round. And so they were kind of getting their footing and saying let like we want to hire like Silicon Valley PM to help us grow. Um and yeah, I mean that company was kind of the classic startup story. I joined they were asking for donations as a as a business model when I joined. When I left they were doing tens of millions in revenue. We had tens of millions of MAU. Like over half of US high school students were

using the app. um toward the end, you know, the this was before chat GPT. And so if you're familiar at all with with uh with the the homework help kind of category, chat is totally I mean it decimated shag. It's kind of like really disrupted it. >> So we were doing this before that technology really took off. >> And um at the time we had built like a pretty strong machine learning team um and we're kind of gaining some traction in the industry. People were starting to know about us. we raised to series B and um we were starting to have conversations about like well where do we go like do you expand beyond math the technology just wasn't there at the time to to move much beyond the like deterministically solvable math problems and there's some solutions they just didn't scale well >> and so we said look like there are these frontier labs and at the time Google was building Gemini it's like okay we have

kind of the option of we go all in and say like we're better as a part of a bigger organization um or we try to rough it on our And and what we realized in having the conversations and seeing the stuff that was being built, it's like it's clear that that's the route we needed to go. >> And that was when it's like, okay, my job is done as head of product. Like, I built the thing. We're like clear that we're going to sell this. Like, okay, it's time to move on to greater pastures.

>> Yeah. Well, that's such a great example of timing is everything. I definitely think you guys made the right decision. having seen what Jim and I and Chachi and collad have done to some of these edtech companies in hindsight although still a lot of really exciting innovation happening uh within the edte space despite the big labs >> right and they're there now I mean the team is roughly they they are Google Zagreb that is like the Google office in Croatia is the photom map team and they're still building on Gemini so they're contributing to that >> pushing that forward you know >> which must be a lot of fun uh but let's transition into Discord so I believe you joined Discord in 2021 one and you jumped in and one of the things that you

did during your 2 and a half year tenure there was lead a complete redesign of the iPhone and Android app experience and so we're going to talk about this from a few different angles but as a starting point what did Discord's mobile app experience look like when you joined in 20121 and what triggered the company's decision to want to overhaul the mobile apps >> yeah I did join in 2021 I actually joined the growth team which is a very interesting team to be joining to eventually lead the core experience teams like how do you how do you make that transition? Well, the story was it's 2021 and and you know as we all know co was happening and forces everybody inside. Um and Discord is kind of interesting in being a social product in that it kind of rides the waves of growth of other products.

>> Um and so as an example like uh Among Us was a huge boon for growth. Um there was a WhatsApp outage that shifted a bunch of people. But anyway, all of these like external factors, my point is that they were all pushing people online. >> And you know, then when you're spending more time online, it's like, okay, your friends that are gamers, they're pulling people into Discord servers and they're gaming together. They're hanging out together. And uh crypto was also booming at the time. So, a lot of crypto servers showing big crypto communities showing up. This wasn't just about Discord's mobile app. This was just about Discord in general. like the the predominant opinion if you if you were to ask somebody off the the street, hey you know is this scored like what's your opinion? They might say something along the lines of this. I had this conversation many times. Yeah, I downloaded it once uh because somebody

asked me to join or I clicked this link but it was really complicated so I deleted it. Yeah. And we kept hearing that over and over again. And uh you know how growth teams work is like you you work on like small iterative things to optimize growth. Well, at some point in the cycle of every growth team, you hit a wall where it's like we've optimized everything that's possible here. >> Um we need some big ch we need to change basically where we're optimizing from.

And I think we realized like, hey, we're having this conversation over and over again. Like that like standard funnel optimization stuff's not going to get you out of this product's complicated. Y >> and that's when, you know, it was like a hallway conversation with Stan. I was like we we sat down in these chairs in in Discord's office and he was kind of uh talking to me about the, you know, term inshitification and how it's kind of really top of mind for him. you know, he's written about this and it's like, uh, you know, that resonated with me and it was like, okay, I I actually think that this is something we need to do is we need to probably redesign the product. The reason we landed on mobile being the focus was because that was where a lot of people were spending their time and a lot of like the non-gamer people were actually joining Discord was in mobile and so we said, hey, like we do think that we need to

redesign the product and that was the start of the team. >> Yeah. Well, I'm excited to talk more about this because I was one of the first product managers that I bought. Uh, this was back in 2016 and we went through this very similar transition. I didn't I didn't join as a core PM or a growth PM. I was one of the very first PMs and so I was working on a little bit of everything, but I remember a very similar dynamic whereas like we would test the signup page and the onboarding flow and we would test various aspects about the core user experience, but they just weren't moving the needle as much as we thought they should. And so it was very similar. I I for in my case it was a designer named Kevin Carile and it was a very similar hallway conversation and we we said you know what if we just took a more fundamental step back and reimagined how this app needs to look and then we took it to the CEO and the

CTO and it just so happened that it coincided with a big inflection point in the company's trajectory where we were moving from offline grocery rebates to mobile commerce in addition to offline uh receipt uploads. And so it meant that we were going to need to reimagine the navigation and the information architecture of the apps anyways. And it turned into this like six-month project. And by the end of it, the whole company was involved and it and ended up being this massive redesign and and was ultimately very successful. But these projects can be very uh risky cuz it's like you're you're changing so many things about the experience and they don't always lead to a lot of impact. So let's jump into what what you all ended up doing. you decided that it was time to sort of step back and reimagine things, not just optimize on the margins. What were some of the biggest

problems with the old mobile app experience and how did you go about validating what those problems were? >> Yeah, one of the biggest problems that there there are a few categories but one of the biggest problems that we faced is some of the like terminology of discord was very hard to understand. >> Um so basic things like what is a server? I mean, if you think about the word like what what is it rooted in? It's rooted in this idea that there's a physical machine sitting somewhere that this thing is hosted on. And that's just not I mean, it's not true. It's like everything's in AWS or what you know what like that's just not how the internet works anymore. But that gamers understand that because okay that's how people talk about game servers and um you know team speak server was a thing before discord existed and so okay so this is like a legacy term that people outside of gaming might not understand.

>> We we didn't tackle like renaming servers but what we did tackle is like how do we make that feel more familiar less >> less unfamiliar to people. And so one one example of something we did was we split out DMs. So there's a there was a tab for messaging. Um there's another problem which is we had a thing uh with every username on Discord called a discriminator. It's like a hashtag with four numbers at the end of the username.

>> Well um it's a common pattern in gaming. It allows everyone basically 9,999 people to all have the same name. They just have a discriminator at the end. That's not a familiar pattern for people that are not gamers, right? And so when somebody that's not a gamer shows up on Discord because their friend says, "Hey, like add me." >> And they and they send them the name uh they will type in the name but not the discriminator and then they can't find their friend and it's like this is like a growth this is a growth problem, you know, like people can't find their friends on Discord. So that was another problem that we solved. There were a huge list of of these things, but I would say they all kind of fell into the categories of quality, usability, and familiarity of the product.

>> Yep. Quality, usability, and familiarity. But it also sounds like specifically there was this throughine around wanting to expand the appeal of Discord beyond the early adopter base, which was primarily gamers. And as somebody who has spent a lot of my life as a gamer, I could totally understand how the parlance and the culture is just very different from the mainstream into this broader ecosystem of people who just want to get online and chat about anything, particularly during the pandemic, right? Because you have so many people coming online and looking for alternatives to traditional messaging services. very similar to what I was describing in Ibata where we were moving from offline grocery reboots using receipt uploads which is one use case to mobile commerce purchases which is a totally different checkout flow different target user different um different partners uh and so what were

some of the steps that you took to figure out how to get from A to B with this new design? Did you talk to users? Did you run online surveys? Did you use tools like user testing.com? I'm curious what the process looked like. So, it's me, me and my engineering director and then Stan. And it was like one one it was one or two weeks where I personally what I did was I blew up my calendar. I said I'm not going to take any more meat. Well, I I'll say before we did that, we we tried to do like the standard we're setting up a team thing and like let's write a strategy doc and and and basically uh my peer and I turned that into S. He's like no no no this is not what we're doing. We actually his vision was we wanted to incubate the culture that we want Discord to turn into with this team.

>> And um I think that he kind of saw some of these quality challenges and the things that we, you know, were were seeing as opportunity as things that were like ingrained that we needed to to kind of course correct and we need we wanted to optimize a company. And so then then I said, "Okay, well, I need to do something very different." I blew up my calendar. I said, I'm not going to take any meetings the next week or two, >> and what I'm going to do is basically just go use as and familiarize myself as much as I can with as many chat products as I can.

>> And so I was I have throwaway I have I don't know how many throwaway accounts. I'm probably eating up like dollars worth of spend. Uh ju just like throwaway accounts on Skype and Telegram and WhatsApp and Signal and and Messenger and iMessage. and all of my friends and family loathed me for those two weeks cuz I was testing all features. >> But basically what we put together is like a master list of like all of the things that we see as problems. And that was the starting point of like okay we just need to brain dump like what is all the what's the full set of things that are like in this sphere and then okay then we need to like go group this and figure out teams and all the rest of that stuff. That makes a lot of sense.

You looked at your competitors out there or adjacent companies in the messaging space broadly defined and you took that as inspiration to build this laundry list of gaps or deficiencies in your existing product experience. Um, did you

supplement that at all with talking to existing users or was that not as much of a focus for you guys as you were developing solutions? Discord's kind of an interesting product and I would say the only corlary to this that really exists in the world is Reddit >> um in that it's a social product that has users that spend hours and hours a day um with their closest friends. And when you have both of those things together, you have very vocal people um with it's very easy for users to form an opinion about how they think the thing should work.

>> And so we have very active Reddit community. Um, we have active feedback servers where we get feedback from people. And so we we did a design sprint where we basically put together a vision and then we put together a server where we got feedback from power users. The thing that looking back I wish that we had more clarity on is really drilling into some of the tension that existed. Like one principle that we set out with is we don't want to build anything that will alienate our existing users. So the idea was like, okay, take the next group of people as a supererset of the current set.

>> It's like whatever we build has to fit the full set. >> Um, we're not moving away from where we are. >> I think we did not properly investigate or pinpoint like what is the tension. Like for example, the usernames change was a big one. If you Google this, you'll see a lot of a lot of hubhub about the username. >> Yeah. >> Everyone on Discord had to change their username. I mean, that's a like, you know, objectively a big deal. Mhm.

>> But I think what we didn't what we maybe underestimated or didn't understand enough was how much something like that like was really going to come to a head or really what the tension was going to be between this group we want to go to and the group that we have today. >> Yeah, that makes sense. And I can imagine the trade-off there. I mean, people get very emotionally attached to their usernames, especially gamers. But then on the flip side, and we'll talk more about this in a bit, I can imagine how the discriminators just made it really hard to remember who your friend's username was because you're not going to remember a random four-digit number.

>> Um, so let's use that as a transition into talking about some of the most important changes you made and the impacts those changes were able to deliver. And let's just start with um

what what were some of the new features or changes in existing features that you all ultimately ended up focusing on as part of the app relaunch? Uh, I'll start by telling you about one which is probably my favorite. Um, one of the PMs on my team um came to me one day and said, "You know, my nephew loves to send voice messages to his friends and we don't have that feature on on Discord." And um, there was an anecdote floating around from another business of like, "Yeah, well, they did it and it didn't really work." And uh, >> and she wasn't very convinced by that.

So I said, "Okay, like um we had this principle uh of how do we get to like the minimum viable like deniable situation, right?" So I said like, "Well, what's the what's the minimum thing that you could do to try to go validate that?" >> And she ran a really quick survey and we got we got a sub a subset of uh of Discord users. It was something like I mean it was an overwhelming majority. It was something like 70 to 80% of them reported using voice messages weekly.

And that was even higher in certain geos like South America, the Middle East, like much higher. It's like oh my god, this is insane. Like all like this is a clearly a generational gap where like I don't send voice messages. Kids send voice messages a lot. >> So okay. So like that's that was a feature we built and it ended up being great. It like increased overall communication on the platform by some single-digit percent which is huge for a platform with you know hundreds of millions of people.

>> Yeah. >> Um so that that that was one example. There are a couple things I love about that and then let's get into a couple other examples. But one, >> you know, I'm somebody who has always loved surveys. I know they can be controversial and there's always a risk of bias, but in my mind, it's just such a fast way to get a directional indicator on how users are feeling about your product. Now, promoter score, product market fit score, willingness to pay, sentiment analysis, and you can do it in a day if you're savvy about it. uh maybe a week if you're going after a larger sample size with a more sophisticated set of statistical techniques. Um but it it can just be so powerful and I think some people object to surveys because they make perfect the enemy of good. It's like yeah, it's not going to be perfect. There's going to be some bias, but if all you're looking for is a directional indicator, in this case

on voice chat, it can just be such a rapid way of getting a quick answer. Um, and then the second thing I love about it is that you didn't take the fact that voice messages on another similar chat platform hadn't worked, right? Because like so often when you try to copy paste features or experiences from other products, even other products in the same space, it just doesn't translate because your target customer, your product are different. And Discord in particular is a product that has so much personality to it. And it's not for everyone, >> but it also has some very passionate power users. And in this case, it turns out they love voice chat, which again, as a former gamer, doesn't surprise me at all. Um, so that's a great example.

And um, and and what sort of impact did voice chat have either on overall messaging volume or other parts of the product experience or business KPIs? >> It increased overall communication of the platform, which was great. And uh, some of the some of the ways people were using voice messages are hilarious. I get like I I now get Tik Tok sometimes of like people like playing a voice message in a Discord server. It's just kind of very funny serendipitous thing where it's like yeah I know how that I knew how that came to be. One one additional point that I'll make that I think was really kind of the root of the culture that we were trying to build. We had a strong belief that quality and speed were not at odds. I think that in the world of AI that's like becoming increasingly clear. But the parallel I see to like vibe coding and and what it's done for your ability to like iterate and learn. That was really our

focus was like how do we get to the first thing that you can see and feel together. That is a piece of information that helps you kind of decide like what do you need to do next? The reason that a survey worked in this case was because it was like a very discreet question. How often do you do this thing that already exists somewhere else? Like that that's like that's like the best kind of survey that you can run. I think the reason that people find I think the reason surveys get a bad rap is because not all problems a survey can solve and in many cases it's very easy to design a survey that like doesn't tell you anything. Like you see the data and you're like I don't know what I expected actually to to to get from this thing.

So >> I think that's a really good point. I mean one you got to distill what your key research objectives are up front before you design the survey. both to ensure you avoid the problem you just described where you run the survey and you're like wait a minute we didn't actually learn anything from this but also to keep the survey very short because every incremental question you add to a survey is going to reduce the response rate two I think as much as possible if you can ask survey questions that are not hypothetical or ambiguous and in this case it's you know how often are you already using voice chat across other platforms like there's there should be no bias there maybe there's a little bit of people over underestimating how much they actually use voice chat but like you're asking a very straightforward question with a very straight straightforward answer, but one that's very valuable to solving

the specific research goal that you were after. In this case, so that's a great

example. What are a couple other examples of key product changes you made with the app redesign and the impacts that they delivered for the business? >> We completely redesigned um how how Hangouts like video chats work um which was awesome. It was like a a great uh exploration also very hard technical problem to solve. Like we I remember we worked on it for many months. Uh, but video chats were like notoriously very buggy on mobile. Like I remember I would be in like uh uh all hands and I'd like zoom in on like the screen when I'm on my phone like commuting to work like walking to work >> and it would like get stuck and I couldn't zoom out and I'd have to close the app and reopen it. So there are things like that where like when you build a product that big that scaled that quickly like Discord did, >> inevitably you're going to have bugs. I mean, this was a very lean team building like a highly scaled platform that

millions of people use every day. And so, we found a lot of the nooks and crannies of Discord and like clean things up. Like video chat as an example, like increase video minutes between friends, which is huge. I mean, for a social product, like video chat is one of the more intimate ways that you can connect with somebody else that you know. >> There was another feature that I uh really loved which was we redesigned the profile tab. Um, so we made it uh clearer that like the profile is effectively like how other people see you. Um, it was kind of this weird state where you have to like unlock and then relock the profile. Very hard to get to the actual customization features and now natively you can kind of see all of your profile effects. So at the time uh there there didn't exist these like profile customization effects um as as um maturely as they exist now. Actually one of the PMs that I hired now leads

that that product. But now you can get all sorts of stuff. Animations on top of your profile. You can get like header that's animated. Um, so things like that that's like a monetization feature that only really became m fully mature uh because of the foundations of building this on mobile. >> This episode is brought to you by Paddle. Today's fastest growing apps are multi-channel. They use web monetization to reach new audiences, improve attribution, and boost margins. But selling on the web gets complicated fast. Managing currencies, taxes, fraud, and disputes gets increasingly difficult as companies scale. Paddle solves this.

As the merchant of record for thousands of apps, including Runna, Goodnotes, Codeway, and HubX, they take on the complexity of web monetization so you can focus on growing your business. Learn more at paddle.com/subverive. And then you mentioned the username change previously and all the controversy that came with that. Um, but what sort of impact did that have? Did it did it make it easier for people to connect with people they already knew because it was easier to recall the username and what sort of impact did that have on network effects and communication volumes?

>> We saw um a pretty crazy increase in the percentage of outbound friend requests that were accepted. Um which is not not super surprising. I I will say that like overall it it's very hard to attach something like that to a growth metric because you can't have hold out groups in the standard way. Uh just the effect of like people need to choose a new username and if somebody chooses it before you then you don't get it. So like everyone had to get this all at once. Um and the other thing is like it's very hard to know longitudinally like over many months like what happens. I would say that like my intuition is before that change, you know, if somebody was like sufficiently motivated, they would have found their friend eventually. It's like, hey, you know, you didn't accept my friend request. I think there were probably many people on the fringes, though. And whenever you have a product this big, that can add up to hundreds of

thousands, millions of people over the course of many months or years. >> Yeah, absolutely. Well, and in particular for new users where that extra little bit of friction can make such a big difference >> in whether or not they get to the activation moment >> where they decide to stick around and keep using the product. >> Yep. >> Can have massive implications, right? Like >> Facebook's whole seven friends in 10 days activation moment is sort of canonical at this point. I don't know what the activation metric was for Discord, but I imagine it was similar, right? It's a social network, so it has to be related in some way to the number of friends you have on the platform and how often or how deeply you're engaging with them. And so being able to quickly find friend requests or connect to the right people who you already know on Discord feels pretty important for a new user. And then it's kind of binary like

either you activate or you don't. And if you don't, you're never coming back. If you do, you might be, you know, a user for years and years and and then ultimately you might subscribe. Which gets us into the last topic I want to talk about and and that's around monetization. So, let's transition to

Discord's monetization strategy because Discord is the unusual company that has built this massively popular social network with millions and millions of MAUs. I don't know the exact number, but I imagine it's tens if not hundreds of millions of MAUs. And layered on top of that, a pair of premium subscription plans, Nitro and Nitro Basic. Um, which is unusual. Like Twitter now X has a similar model, though they didn't for many years. Um, LinkedIn obviously has subscription plans, but they're a very different social network because they've always been so professionally driven.

Um, so what prompted Discord's decision to build premium subscriptions as opposed to monetizing through ads, which is what most social networks do? Chatbased products are tough products to monetize. Notoriously, they're tough. Um, if you look at like Facebook and Messenger, they haven't monetized Messenger. Uh, now, okay, you can argue like Messenger for business and the B2B play and like same thing with Telegram. Yeah, like those things exist. Um, if you think about an ads business and what it what it incentivizes and then also like the types of mechanics that have to exist to to build a really great ads business, you really need places like that you can put like you need inventory locations, inventory spots. It's hard to find real estate in a product that's mostly happening between two people or three people. Um, now this is a little bit of a misnomer about Discord. Of course, communities is like where many

people find Discord, but I would say like the meat of what happens on Discord happens in small groups. That's like where the magic is. Um, small groups gaming together in a small server or in a group DM or one-on-one to build ads into that feels a little icky. It feels like maybe maybe it wouldn't actually be a super huge business. Now, that's like the traditional display ads path. They're forging a new path, which I think is hugely innovative and very different through reward ads, which are called quests. And there's a whole digital ecosystem around them and like integration with gaming and entertainment, which I think is awesome.

Um, we we talk about that a little bit. It's a little bit after my time. I think they pivoted that direction after I left. Um, but then if you disqualify the display ads as a direction, it's like the other route that makes sense is okay, there are these like power user features, Discord has a very strong power user curve of people that are using the the platform for, you know, 8 hours a day. There are some people that are on there basically 24/7, right? And it's like, okay, there are features they care about like file upload size, custom emojis, things like this that >> they're not going to affect the core curve of like a new user onboarding.

It's not like uh file upload sign's going to prevent me from becoming a new user if my friend invites me to a server. >> Um but they will help power users. And so I think that there was that principle at play of like okay it kind of makes sense for how people use the product that we build a premium subscription on top of it. >> Yeah I love that. So just to recap you said the decision not to monetize through ads was largely just a pragmatic one. It's not like it was an ethical decision of well we don't want to monetize eyeballs. It was more in a messaging app where people are largely interacting in onetoone or small group forums with maybe two, three, four people. There's just not a whole lot of obvious surface area real estate to be able to insert ads that are going to drive any real amount of volume. And then secondly, and this is where I'd love to go deeper because I think this is such an important point, you were

able to build premium features into Nitro and Nitro Basic that didn't disrupt the core growth loops that were responsible for driving Discord's user growth. So, let's double click there because I think this is so important for premium subscription apps to get right. You never ever ever want to put features behind the payball that are critical for driving your core growth loops. Whether that's word of mouth and verality or it's content driven growth like Quizlet uh benefited from where people are creating content that then gets indexed by Google or other search engines to drive SEO or increasingly now answer engine optimization on chatbt or claude or Gemini. So what were the core growth loops that were critical for Discord and how did you ensure that the features you were putting behind the payw wall didn't disrupt those growth loops? I would say that the most important growth loop was friends inviting friends to use Discord

for either for a server or to game together or to like chat. And so that that was like the main thing that you you don't want to disrupt that flow. >> The monetization principle that I think ultimately kind of gates that is the incentive. um which which so you know that I would say like at the scale that Discord was when I was there it's kind of clear what the monetization strategy is um and oftentimes I think a lot of a lot of times companies make that decision pretty early on I'll give you an example I when I was at Photomath um there was a similar company that uh started around the same time basically had the same value proposition the main difference is they didn't have OCR so you have to type the math problem but it's the same thing we we would uh constantly kind of battle with one another and the top charts. They so they were doing very well. They made the decision fairly early to monetize on the

number of scans that you had, the number of math problems you solve. So after a certain number of problems, you have to pay. You basically can't use the app for anything other than looking at the past problems you solved. At Photomouth, we realized like that affects the core loop. Like you can't really get any more value out of the app. And what was harder was we had to make the decision to monetize something else. But that ended up being massively successful. And if you look at the outcomes, I mean, you know, from the time I joined to the time I left, we're talking at like 100x outcome.

>> Mhm. And I'm guessing that this this product that I think it was eventually acquired, maybe they made a few million bucks off of it, but I'm guessing at least 10x lower, may maybe even more, maybe an order of magnitude more lower valuation because they decided to kill that core loop. And so I guess where I'm getting at is like at Discord, it was pretty clear like what the strate There there were like mechanisms in place that at that point it didn't really make sense for a PM to build a feature that would cut off a core loop. uh because uh the monetization features actually need to be valuable to somebody.

>> Yeah, it makes so much sense. And going back to the days when you and I were riffing at Quizlet and Photomath respectively, the equivalent for us was StudyBlue was a direct competitor. And they ended up making a decision relatively early on, call it 2015, 2016 if if my memory serves me correctly, to one be much more aggressive about forcing new users to register an account before they could use the product. And then two, they made different decisions with regard to pricing and packaging.

that on the margins impacted a non-subscribers's willingness to create a set which was the atomic unit of content on Quizlet and StudyBlue and it had a huge impact on their growth trajectory. Like up until that point, Quizlet and StudyBlue were pretty much neck andneck in terms of user volumes. And then Quizlet just leapfrogged StudyBlue significantly because this was before I joined, but the company continued to be very generous around letting non-registered users continue to access the product for things like reference use cases around getting help on a homework assignment. And Quizlet was very adamant about never putting features behind the payw wall that would disrupt content creation because creating sets of flash cards, digital flashards, was so critical for driving the core SEO loop that was responsible for more than 70% of Quizlet's user acquisition. And so it's almost like a a sacred cow of like you do not touch

those growth loops, particularly if you're a company that depends heavily on organic growth through word of mouth or SEO. Um, let's talk a little bit about

how the app redesign that your team led ultimately helped to impact monetization. I know that wasn't the primary focus of the redesign. Uh, but you mentioned when we were preparing for this podcast that there were a couple of bugs you fixed that did have a pretty significant revenue impact. So, walk me through what that looked like and and how that ultimately helped the business grow revenue. This is a one-off, but there was a bug on the profile where if you tapped the Nitro button, uh the the Nitro landing page wouldn't load in some cases or it would take like 30 seconds to load. So the these are things that like you know I have plenty of examples of this throughout my career mostly on growth teams where you find small stuff and like it actually has a huge impact on growth because there's so many people there right that like inevitably like you fix one thing that leads to a payment flow and it it leads to millions

of dollars. Um that was one example. I would say like the biggest thing was um two things. One is of course we like emphasized the profile a bit more so it's clear like how you're representing yourself and how you're expressing yourself. That was a direct learning from this like we're moving beyond gamers to like >> okay there are other teens being invited like what do teens care about? teens really care about, especially Gen Z, care about authenticity.

>> And that was a thing that like um yeah, like gamers care about authenticity too, but more in an anonymous way. I would say like we really leaned into customization and expression and that I think the foundations of that have have carried on to the monetization strategy that exists today with the shop. The second thing is the learning that like hey one big thing we realized through this whole play is like one foot in front of the other. I mean, we we had expanded our TAM during COVID, but I think we overestimated how how how permanent that change was. And um what we realized is like, yeah, gamers really are the bread and butter of this business. And uh and that that learning I think led directly into quests which is now the the kind of the primary leading the charge monetization strategy that that that I see externally being focused on which is effectively reward ads. And it's funny because we we talked

about like the value behind the way that a business monetizes. If you think about regular display ads like don't make sense for Discord. something like quests, which basically how it works is like you get uh they work with brands to say, "Hey, like uh we will put this content in front of users that are relevant and if the user completes the quest, they get orbs they can spend on profile customization features." >> Um well, the the quests are really interesting for people. I mean, they're like sneak previews of a game that you might want to try, that you might be talking about with your friends, and you might be hanging out in servers where that exists. And so this is like Instagram ads. Instagram ads provide a ton of value to me as a user. I f I find products all the time that I love through Instagram ads and like that that is an example of like a virtuous loop with the monetization strategy that I

think I personally think like quests are probably going to do well just just because of that thesis that I see as an outsider. >> Yeah, I love that point and I'm by no means an ads expert. I've spent very little of my career working on ads, but I have found in a couple cases, you know, Ibata had an ads team and Quizlet had an ads team. And so my observation of the ads that worked most successfully at those companies is they were really well integrated into the core user experience in a way that meant that in a best case scenario the ads were actually a net positive experience for the end user. Obviously that's not always going to be the case but in a best case scenario you can achieve that. And so in Ibata's case that was you know the the user there to save money. People using Ibata many of them are from the middle of the country you know maybe they make the median household income in the US.

Uh, but the money they save on IBATA is the difference between having money left over at the end of the year for Christmas gifts or to go on a family vacation in the summer. And so there were these ads that were video ads that a user had to watch in order to unlock additional savings. But they're happy to do it because they're there to unlock savings. And a 30-cond ad spot to unlock a dollar in savings is is a pretty good return on investment for these people.

Um, but then two, because Ibata has all of this data on the products you're purchasing and the savings you're redeeming, it means they can be very targeted but about which ads they send you. And so like your example with Instagram, often times the ads are are really valuable for these users because they're they're showing products that they're genuinely interested in. Um, and so I think in the best case scenario, you're not just serving a, you know, a dumb banner ad. You're serving something that feels like a very native ad. You know, Instagram is certainly a great

example of that. Uh, well, this has been an awesome discussion. Let's jump into our lightning round. And I always ask the same question first, which is just what did you find to be the best part about working at Discord? And then what's one thing you would have changed? >> The best part by far was how whimsical Discord was and that's why I joined the the the company. I mean, just like everything about it, the brand, the way that the team works, the like signs in the office were whimsical. It was just so awesome. Uh the the mascot being something called Olympus. I mean, that's just like just like brings a lot of joy to your day. Um, something I would have changed I yeah I think I look back and like I would have changed probably the clarity to which we like understood the the the disconnect between the thing that we were wanting to do and the core core power users of the product. I think we discounted it a bit more than we

should have and and hindsight's 2020. It's impossible to know sometimes until you ship stuff. So that's one thing I would have changed. >> Yeah, it's a great point though. I mean, there it's such a fine line to walk between continuing to double and triple down on your core and make sure your power users are happy while still broadening your appeal over time to a a broader swath of users. And it sounds like that was a big part of what you were trying to achieve with the redesign. And in some cases, you succeeded and in some cases less so. Uh, but it's always tricky because the more you water down your experience to make it broadly appealing, the more you can actually cut against the grain of what the power users loved about about the app in the first place.

>> Yeah. Um well, next question here is what's the weirdest Discord server that you personally have ever been a part of? >> I um this this is like more a funny story, maybe less weird, but I was a part of a a staff server where we were testing a new feature. Uh I think I think we were testing voice messages and uh the way basically the way the server worked is like we enabled the feature for this server only. And so a bunch of employees we were just like messing around sending test stuff, just being silly. And uh one day we show up to work and our phones are blowing up and it's like a 100 people have joined to this server and when you click in like these are not bots. These are like real like real users that somehow like some way they found a way into the server. The link had been leaked. I don't know. It got posted in a server somewhere. We never figured out what happened. We just like kicked everybody and closed the

server down. But it was a very very fun experience. >> Yeah. Yeah. Well, I can imagine a feedback server, given how passionate your user base was, if if users were able to find their way into it, >> they will find a way. It's like water, you know. >> It's a little bit of a it's a little bit of a tangent, but we've worked with several clients who've done either big redesigns recently or they, you know, changed their subscription packaging and put features that were previously free behind the payw wall. And it's such a controversial thing to do, but what I've found over and over again having having worked on a number of these redesigns at Quizlet and then and then with clients now is you're always going to have a very passionate minority that is very vocal in the first week or two after a big launch.

>> But you got to be very careful not to overreact to that initial backlash because people generally don't like change. And so if you change anything, you're going to get some really angry people. But you almost have to force yourself to wait at least a few weeks just to yeah to use your words let the dust settle and then see how much of it really lasts and compare the anecdotal you know feedback often negative feedback on social media that comes early on versus the staying power of the impact it ultimately has on the metrics and it's particularly true of price changes because anytime you increase price people aren't going to be happy but ultimately it can be very powerful for the business. Okay, so next question here is you've been a CTO twice, chief product officer. What's the biggest piece of advice you have for new product leaders who are making the leap from being an individual contributor PM to

leading a product team? Whether that's as a group product manager or director or ultimately as as a chief product officer as as you've been at a couple cases. >> There's a lot of advice out there in being a new manager, being a new product leader about how to properly delegate. And so I I think that's right. I think that you need to build an awesome team. Don't discount the idea that you also will need to make hard decisions for people sometimes. Like that's the that's the flip side of that which is like sometimes people need you to make decisions for them. And actually that's better for both of you. And like that's the sign of it. That's the hallmark of a good relationship. The hallmark of good leadership is can you be candid about when you need to take the reigns and just be clear about how you do that and how you approach it and uh don't be shy to to try to do that whenever you feel you really need to.

>> Yep. I think that's spot on. And there are a couple different dimensions on which I think that's so important. One is the nature of the person you're managing, right? There's sort of a two I forget what it's called, but there's this 2x2 matrix of >> new employee versus tenured employee and then low to medium performing employee versus high performing employee. And the way you treat a highly tenured employee who's one of your top performers is going to be very different than the way you treat a new employee or somebody who's struggling. And generally, you're going to give a lot more autonomy and deference to somebody who's really proven themselves versus somebody who's new or hasn't been performing very well.

And then the second dimension is just how important the decision is, right? If it's a one-way door and if it's a controversial decision, then ultimately I think it's your responsibility as the leader to just jump on the grenade and say, "I'm going to own this one whether it goes well or not, but we got to move forward and take that pressure off of the IC on your team versus if it's a two-way door or if it's low stakes and you can easily revert the change, then often times I like to delegate those to more junior employees because it's a good way for them to learn and to get a sense for how good their judgment is and then you can always coach them on the way out." Great. So, a couple more here.

What's a lesson you've learned about building apps for high school and college students? Because you're somebody who's worked at Photomath, worked at Out School, worked at Discord. All three of those products cater to younger users in their teens or early 20s. And so, what's what's a lesson that you've learned about serving that particular customer demograph? >> Yeah. So, I I thought about this question. I think it relates to the thing we were talking about with servers where kids find a way. Uh it's hard to guard somebody in basketball that doesn't know how to play basketball.

>> Mhm. And um I don't play basketball, but that resonates with me. The idea that like somebody that doesn't know the rules can kind of find new ways of doing things that are it's hard to guard against, but it's also hard to um know what they're going to do. It's easy for them to find things that that work that you wouldn't have thought of. >> And so I think that that relates to um something that is really important to me, which is curiosity. Kids are innately curious because of that. And um so it affects everything about how you build a product. It affects how you uh do trust and safety. It affects how you monetize and it affects the rules and assumptions you think like you know photomath was a a big disruptor in edtech in in terms of like homework and what's the purpose of homework and what role does it play in the classroom like kids do not care about that question right they care about learning and like

what value do I get out of this class and so I would say yeah that that's my biggest thing is like kids don't necessarily play by the same set of rules that adults do and that's probably good and bad in some ways. Yeah, I couldn't agree more. Having worked on a couple of edtech products myself, the fact that you brought up, which is students are always finding new and different ways to approach things and the gener generational gaps that exist between millennials versus Gen Z versus increasingly now you have Gen Alpha uh starting to play with some of these products is pretty material. And so things like, you know, voice messages, something that you and I probably didn't do a lot of, but a lot of our users who are 5 or 10 years younger, that's all that's all they do. It's really important to uh keep that in mind. Um, and then I think that's even more important now in the AI era because I think you're seeing cases where, you

know, like we were probably among the generation that were native uh social networking users, right? Like I graduated college in 2008, right around the time that Facebook was really starting to take off. But then there's sort of this new generation now that are growing up with AI from a very young age. And so they just it's more intuitive to them and they use it in different ways in some cases which means they can leapfrog much older and more tenured uh members of a team because they just think about things from a different angle and so keeping that in mind is really important. Well, let's end on this question. And so you recently left your executive role at Out School to start your own startup. And I know it's in stealth mode. So I don't want you to share anything you're not comfortable with, but what can you say about the uh the product or the problem space you're working on?

>> Yeah. So I I can talk about the problem space, not the product yet, but the problem space that's really exciting me. It it roots in a thesis that I have about AI, which is I think AI capability is going to far outpace adoption for quite a long time. Mhm. >> And when I when I say adoption, I don't necessarily mean like the average the average consumer using AI. I think a lot of people use AI. My parents have used AI.

>> But I think the adoption of AI in ways that meaningfully change somebody somebody's life uh is is going to is is going to be outpaced by the ability of AI especially in consu in consumer. >> Mhm. >> And so that's the problem space I'm focused on. And um one one layer deeper than that that I'll share and then I'll stop is when I was at Photomath, we really focused on this idea of um the moment of being stuck. And the thing that I would always picture is it's like 1000 p.m. You're sitting at a table, maybe your parents with you, you're like on the verge of tears because you're so frustrated. And then there's this app and oh my god, it's magic. It solves this math problem for you.

I think that we need to find moments like that for the average consumer and and find ways to solve solve problems for them. I think that for many things that like where getting unstuck is very trivial like solving a math problem. AI like out of the box models do a good job. I think for life's hardest problems that's not true. And for life's hardest problems that's also where people are honestly the most motivated and and maybe the most likely to give up uh in the face of challenges too. So that's that's what I'm excited about. That's what I'm focusing on. I don't know yet.

Like I can't share all of the details of exactly the product, but that's what I'm working on. >> Yeah. Well, I think it's a fascinating problem space and one that is so important, especially as we talked about, you've got this entirely new generation who's coming out of high school, coming out of college, trying to find a first job. And many of those opportunities, those entry- level jobs that you and I took for granted aren't going to exist pretty soon. and and they've already started to atrophy because AI is taking away those jobs and so helping people particularly young people just become very proficient in AI very quickly uh is going to be very important for them as individuals and also for society.

>> Um so I I very much wish you the best of luck and it sounds like an exciting startup to be working on. Um this has been awesome. I knew it would be. Before we part ways, any parting words you want to give to the audience before we wrap up? >> Go out and make the world a better place. I think that that software gives us this awesome opportunity to do it and AI in particular is is really powerful thing and so I would encourage you to think about how you can apply that to your career. That's that's my parting wisdom.

>> Yeah, I love that. Very inspiring note to end on and uh couldn't agree more. I mean there's a lot out there about some of the negative impacts of AI, but there's unbounded upside as well and it's a pretty exciting time to be building. So thanks Franchesco. This has been a real pleasure. >> Thanks Bill. >> Thanks again to RevenueCat for sponsoring this episode. If you're building a subscription app, you can use RevenueCat to power purchases, optimize pay walls, offer discounts, handle refunds, manage customer data, run AP tests, generate insights, and unlock more revenue. I work with dozens of subscription apps, and most of them are already using RevenueCat to power their businesses. Learn more at revenuecat.com/subversive.