Hinge CPTO on Building an App Designed to Be Deleted, Merging the CPO and CTO Roles, Chasing $1B in Revenue Without Chasing Engagement | Ben Celebicic | E312

0:00I can see a world in the future where if I am able to provide a perfect query input to, let's say, Cloud Code or Cursor or Codex or any of these tools, I will probably get a perfect output.

0:11Carlos: It's kind of an anti-metric. You want people to stop using your product because you want them to find success as soon as possible. As looking at the market cap of your holding company, Match Group, it's around 25 % up here today. You look at SaaS companies, most of them are down. So Hinge is on the mission to achieve$1 billion in revenue in 2027 next year. Ben Celopechik, Chief Product and Technology Officer at Hinge. You're making me think about the dark side of AI now. People creating fake images, trying to look better. People who can fake conversations to try to get more data. Trust and safety lives under me.

0:44And it's the third of our Hinge organization is trust and safety. All of that has been done through AI for ages.

0:50Carlos: Do you see a future where the lines are truly blurring and we can have more non-technical people really shipping into production. Hey, this is Carlos, CEO at Product School and your host on the Product Podcast. My guest today is Ben Chelebechik, Chief Product and Technology Officer at Hinge. He has been there 11 years, since the 2015 relaunch, and he got the product title added to the tech one. So now he runs a team of about 300 and negotiates the CPO versus CTO tension inside his own head. Hinge is on track for a billion dollars in revenue next year, and their North Star metric is getting you to leave.

1:21Carlos: Here are the things we'll cover. The app designed to be deleted and why great dates is the metric. They only charge for what they cannot give away free. A third of the whole company is trust and safety. His conservative take on vibe coding and where he draws the line, the perfect query problem and why he thinks it's two years out. Let's get into it. Welcome to the product podcast, Ben. Thank you, Carlos. Thank you for having me. Ben, you've been at Hinge for almost 11 years. You're a real OG, right? And then most recently you got promoted from chief technology officer to chief product and technology officer.

1:52Carlos: So I'm very curious about that transition that seems to be a trend for other companies as well. Yeah, 100%. So you're right. I've been with Hinge for 11 years now, pretty much since the beginning when we rebooted Hinge in 2015. And I led our technology org from the get-go. But I was in the room, you know, with the exec team when any product decisions or any marketing decisions for that matter, any company decisions were made. So it's very, very close to products. And when our old chief product officer, Stefan, left about two and a half years ago, it was actually an idea from Angel, who was our chief operating officer at the time, you know, why don't we let Ben do this?

2:33It wasn't my idea at all. And, you know, Justin, who was our founder and CEO at the time, said like, look, you've been with me for all these years. You've been in these rooms. Let's try to see how it goes. And honestly, we never looked back. It's been, you know, I'll be honest, it's been really tough in the beginning, especially, you know, for a company that is a product company like Hinges to take additional responsibilities beyond just building technology. But I think there are like some bright spots around, you know, there's always a tension between CPO and CTO around resourcing, around like what to concentrate on, what to build, where to spend time.

3:13CPO has always kind of tendency is to concentrate on product and innovation and build as quickly as possible on that front. While the chief technology officer always wants to make sure that, you know, we maintain high quality, there is, we pay debt on the tech side, we build infrastructure that will help us and enable us to build cool things in the future. So there's always this tension between the two. And I think it's a very healthy tension to have. So I had to have the tension with myself since then. And, but, you know, I would say luckily for me, you know, I have a really strong team under me and we still debate a lot of these decisions when we kind of prioritize work.

3:56So it's been really tough in the beginning, as I said. It's a lot of responsibilities to handle both. But I think with time and team building that I made, I'm now at a place where I feel this was the right decision.

4:09Carlos: What is the total size of the team that you oversee? Yeah, so Hinge is about overall, I think about 360 or so. my team is very close to 300. Most of those are on the tech side, and the rest is product design and research. So other than the obvious benefit of having one person that consolidates both technology and product, what are other decisions that you have to make around the org to make sure that you still have that healthy friction and that the teams collaborate even without you in their room. Yeah, 100%. So, you know, I think for me, work structure heavily depends on the problem you're trying to solve and the strategy you have in place.

4:49Like work structure is always a reflection of those things. And it doesn't really work in the opposite direction. So what you want to have is a very clear vision, what you're trying to achieve, a clear mission for the company, and some kind of a strategy that is helping you moving towards that vision and mission, and then put a team in place that will help you achieve that. So every year, as we build our strategy for the following year, we really look at the org and how it's organized and what makes sense, what doesn't make sense, and adjust from there on. From this tension perspective, I think having leaders on engineering and AI and data and then also having leaders on product and design and research in the room when we have these discussions is tremendously helpful.

5:36You know, when there's really no debate around resourcing, around prioritization, whether we spend time doing more of a product work or more of a tech work, that kind of lights a bulb in my head that something is wrong. So I really prompt the team to really think deeper about it. So you have to have a team underneath you that is going to have perspective and is going to push you and is going to make sure that like nothing really falls through cracks. Especially now in the age of AI, I would say we're still all trying to figure out what this technology is going to enable in the long run. I think having infrastructure built and ready once we kind of decided what makes sense to build is tremendously important.

6:16So having people really pushing for spending time building infrastructure, building, preparing for the future is very, very necessary. On the other hand, we also have to do good for our users and build features and build improvements to the product in a way that is going to get them achieved success on our platform. So that tension.

6:37Carlos: Let's talk about that, Ben, because to be honest, I mean, I'm married, so I cannot speak as a user. But the reality is that the online editing space, I think when Tinder came out, they pioneered this new behavior, right? Swipe left, swipe right. But I haven't seen much more innovation since then. It feels like an industry that hasn't produced any new of those type of breakthroughs. So I'm curious to know from your own perspective, what is it that you are trying to do to maintain that level of high innovation? Yeah, I think that's a very interesting question. In the very beginning, the industry tried to solve for access.

7:11The industry was born to kind of get you in contact and connect you with people who you otherwise would not meet in real life. And that problem was solved. But the problem that was created by solving the problem is the problem of access. All of a sudden, you have access to too many people. You have this paradox of choice. And it's really hard to make progress when you always think there's like something else out there that I can look through. So Hinge, you know, purposely decided to concentrate on one thing, which is what actually happens when people leave Hinge. So we are measuring our success and our North Star is great dates.

7:46So we don't really try to keep you engaged with the platform. We really try for users to spend time. We really do everything we can to push them off of the platform and really spend time in real life. So I think when you optimize for something like that, that's where innovation is born. Because it really makes you think hard. It's counterintuitive that we're trying to actually get rid of people. And Hinge is known as the app that is designed to be deleted. We really want people to come to us, find success, and leave Hinge. And when you kind of optimize like that, it hasn't been done before. So you really have to think outside of the box and innovate quite a bit together.

8:22Carlos: I like that. And you're right. It's kind of an anti-metric. You want people to stop using your product because you want them to find success as soon as possible. But that also leads me to the next point, which is your business model, right? Usually, like in a social media world or other types of products, you want to keep people as engaged as possible because then you monetize through ads and other mechanisms. So in your case, how do you actually make money? Yeah, it's a great question. So most of people on Hinge don't actually pay. I think probably less than 15 % of users on Hinge pay for something.

8:50And our core principle from monetization perspective is the free experience is sacred. You really want the free experience to be built in a way that everyone can achieve success. And then you charge for things that either break constraints or help you achieve success a little bit faster. So the second principle we have is we only charge for things that we cannot give away for free. Because if we gave them away for free, either they would mess up the ecosystem, and I'll explain what it means, or they would mean nothing, you know, where scarcity is actually a value that you provide. So if you think about things like Boost on Hinge or Rose that kind of helps you stand out in the crowd, those things would, you know, mean nothing if we gave away those for free for everyone.

9:32On the other hand, you know, for dating app to really work well, think about it as like this great party. You're this great party. Everyone is excited to be there. Everyone wants to be there. Everyone wants to speak with other people. And the more people you have in the room, the better the party is. So like providing access to this party comes down to not charging as well. And then if you really want to break some constraints, you want two people to go and talk with each other separately, or you want to have a leg up in how quickly you achieve success, those constraints can be broken through subscription or maybe paying some a la carte, things that you can stand out with.

10:13So our monetization model is very simple. We only charge for things we otherwise cannot give away for free, and our free product is sacred. We want this great party to continue going.

10:24Carlos: As you think about the volume of users that are already on your platform, the age of those users. You have Gen Zers who are probably starting to date for the first time. You probably have a ton of millennials. And so I'm curious to know, how do you go about segmenting users and what type of experiences you think are different for different types of populations? Yeah. Yeah. You can look at it from that perspective. You can also look at it from a perspective of like, I'm a brand new to dating and I need some help figuring things out, or I just fresh out of divorce and maybe I need maybe a little bit different type of dating experience.

10:56You know, the way Hinge operates is we always look at the dating culture. So we have a team that we call Culture and Consumer Insights that are constantly out there talking with daters, understanding their problems, understanding where the needs are, whether those are brand new daters coming onto the scene now or daters who have been using dating apps for a while or not finding success, and then deeply understanding, you know, what problems we need to solve for those. Once we kind of align what problems we need to solve for those, then we have hypotheses that we utilize to really provide solution to those problems.

11:31The key thing to kind of mention here is, as I said originally, Hinge is really measured through getting people off of the app. So if anything that we build here does not increase the chance of two people leaving Hinge and meeting in real life, we don't ship it. So everything goes through that lens. So we kind of prioritize all the work we have, all the problems that we hear from people through the lens of like, is this increasing the chance of vast majority of people meeting in real life or not? And that's what also helps us keep the product very simple and very clean.

12:05Carlos: I think that's part of the magic jet, the challenge, right? It's a clean, simple product that has a lot of complexity because you want to personalize the user experience depending on the situation. You mentioned someone who is divorced versus someone who's newly new to date. So how do you create that type of one-to-one personal experience so someone gets to that aha moment as soon as possible? That's a great question. I think that is what AI is enabling us to do recently. That wasn't really possible if you look like two or three years ago. It wasn't really possible to provide the personalized experience to everyone.

12:40You might have maybe provided an experience that would fit one or two different needs, but experience that is truly personalized for individual needs, and those can be many, was really impossible to do without AI. So if you think about what AI can do is can really quickly recognize, based on some behavioral patterns and how people behave on the product, what would be the next best step for them and help them achieve that. So think about, as you create your profile, you might be really quick about either selecting your photos or choosing prompts you want to answer about yourself. and you might be really good about it.

13:14Or if you're new to dating, Hinge might actually, you know, push you to think a little bit deeper or go a little bit deeper as you answer things about yourself. And we kind of recognize that through how you answer questions on the profile. And as I said, that was really not possible before generative AI.

13:32Carlos: You're making me think about the dark side of AI now. Applied to dating, right? Like people creating fake images, trying to look better than they are, or like, I don't know, these clickers or people who can fake conversations to try to get more data. So how do you go about identifying those type of situations and kind of prevent them to ensure the health of the ecosystem? Yeah, that's a really great question. You know, I think, first of all, trust and safety. So trust and safety lives under me. And it's the third of the entire hinge organization is trust and safety. So we deeply care about trust and safety of users and spend a lot of effort and resources finding those fake users or bad actors on the platform.

14:12So all of that has been done through AI for ages. You know, it's not that generative AI, the recent development generative AI have been game-changing in this space. What has been game-changing is AI generally and simple, traditional machine learning models that we use forever. Now, if you really think about, you know, people coming onto our platform and purpose of Hinge, of getting them off of the platform, You really want to make sure that you are as authentic as you can be as a user. So when you meet other person in real life, you know, you show up as they expect you to show up. You're not going to achieve success if you show up as something completely different.

14:52So what we're really seeing is people are presenting themselves in true light and authentic light. Otherwise, success is almost impossible to achieve.

15:00Carlos: Totally. And look, speaking of AI, as looking at the market cap of your holding company match group, It's around 25 % up year to date. If you look at SaaS companies, most of them are down, right? So there's clearly an AI narrative that is working for you, right? So I'm curious to know, what is it, what do you think it is that is helping you be so successful in the public markets? Yeah, I mean, it's probably hard for me to speak generally about Match Group and what they're doing. I can speak specifically about Hinge. So Hinge is on the way, on the mission to achieve$1 billion in revenue in 2027 next year.

15:40And we're well, you know, on a way to do so. We've more or less grown double digits year over year over the last, you know, 10 years, more or less. And I think it's very much tied to our obsession about really talking with users, understanding their core needs and core problems, looking at the next generation of data all the time, and building features and products that are solving their core needs. That is like the very simple, I would say, formula for success. As you're like obsessive about really constantly thinking about your users and solving their problems, your business is going to have this flywheel of getting better and better and better.

16:20And that is from a Hinge perspective. Match Group, as I said, cannot really speak more broadly because it includes Tinder and many, many other apps under the same umbrella.

16:31Carlos: Yeah, it's really fascinating. I was looking at this around, Match Group has 45 global dating social discovery brands all under the same umbrella. It seems like Hinge is number two in terms of market share. Number one is Tinder. Correct. Just curious to know from your own perspective, like how do you go out positioning your own product to make sure there's a clear value prop that is not competing against other companies even within your same group? Yeah, I mean, if you look at most of those, you know, 45, you mentioned dating products out there, some of them are very specialized. So you have products that are specialized for maybe older generation of users, or maybe for people looking for Black Love, BLK, or Cheese Puffer.

17:14So they have multiple specialized products. And then you have probably out of the big ones, you have Tinder and Hinge. Hinge is, I would describe Hinge as a product that is more intentional. So people coming to us are looking for really intentional connections and getting off of the app as quickly as possible. So our positioning is different than any of these other dating apps. And therefore, you know, we have our lane to kind of play in. Other apps have their other positionings. And of course, it's impossible to achieve very clear swim lanes. Everyone is kind of going to play in each other's.

17:48But it's interesting also to understand that this category is multi-use product category. So users don't necessarily use a single product. They're going to use Tinder and Hinge and maybe some others as well. And I think that works in our favor.

18:03Carlos: That's a super interesting story to me. That's one I keep looking at closely. A friend of mine was the CPO at Tinder. And he told me that he actually joined the company because his kid now is in the nature of starting to date. and he wanted to figure out, you know, how the new generation of people go about their product. And he told me that, yes, one of the things was that they were multi-product, right? Like the same way they would be using that app, they could be using others. And that seems to be like an acceptable behavior. Yeah, 100%. You know, most users use multiple apps. And it's similar, you know, in other cases.

18:37You know, if you're streaming videos or movies or whatever, like most people will use Netflix and maybe Hulu or some other products as well. So this is not special to the dating category. It is just the nature of how people utilize these digital products.

18:52Carlos: So as we look now under the hood in terms of how you're building, how your team is shipping now fast, better, prettier, right? Like tell me more about what is it fundamentally different today that you maybe weren't doing a few months ago? And I'm assuming you're looking for how AI is actually helping us do so, right? But for real. For real, for real. So when I think about AI, you know, at my CPO pod, at my CTO pod, I talk with individuals in New York that we're all kind of really asking the same question, like, how is this really helping us achieve success? And everyone is, you know, still questioning the value.

19:32What I can tell you at Hinge specifically, the biggest value I see is the feedback loop. How quickly can I learn whether something is valuable to invest in or not? So what does that actually mean on the ground? I can really, without any engineering support, prototype an idea that is as close as possible to final product. So it really works with the APIs that connects to our back end. And I can put it in the hands of users and really gain either conviction that this is valuable building and putting in product, or this is not worth pursuing and leave it behind. So this feedback loop of how quickly I really learn about whether something is worth building or not is real.

20:13I really learn much, much quicker. That's one. The second thing is, you know, product managers now can wipe code a quick, you know, visualization of something they're trying to create. So instead of, you know, having this like product spec of 20 pages describing what this feature is supposed to do, me as a product manager can say, okay, please use our design system and build something for me that I can actually put in front of people and show them what I'm really thinking about without them having to read all the products back to understand what the idea is. So like that kind of thing is much, much easier.

20:50Then doing research. So user research has become also a little bit easier. You can, through some AI tools out there, reach more people, number one, and also do more user research. You do more interviews with users. And then I would say the last piece, were starting to slowly move towards what I would describe as mobile engineering. So Hinge is a native product. So you have iOS and Android built separately, natively. And some people on the ground that are picking up some of these tools are already starting to say, like, look, I'm building this feature on iOS. Why don't I just tell the tool to build it for me on Android?

21:28And then all of a sudden, I've done a work of two people in one sitting. So those kind of things we're already seeing happening on the ground. So like in multiple pockets, we see a lot of improvements and help of how these AI tools are making our life easier. But I think if I really look holistically, it does not translate to what probably the narrative in the media out there is. We are moving faster. We're learning faster. We have more tools at our disposal. But I believe, you know, we still haven't achieved the full potential. So there's probably a couple of years left, I think, before we really start to see true benefits of how quickly we move, how better we build, how higher quality the products we build are, and things like those.

22:11Carlos: You used to know where you draw the line for what PMs can do, given that you come from an engineering background, right? And I can see two camps now. We see a lot of native companies saying, hey, everybody's a builder, PMs, designers, basically any non-engineer now is empowered to go and ship into production. but I also see sometimes engineers saying, hey, we have a background in engineering. There's other things that we need to take into consideration. So what is your take on this dichotomy? I'm taking a little bit more conservative approach to this personally and not because I'm coming from engineering background but because I really want to see the proof that product managers or designers can actually provide and build technology that is going to be scalable and maintainable in the long run.

22:58and in order for me to really maintain that I draw the line at the build so for us engineers are still building the product and there's still those who are designing the architecture around it and setting up standards for the quality of our product I haven't really made a leap of having product managers for designers actually contribute to the code base they do actually use these tools to as I said before, to create prototypes, to learn quicker whether these things are worth pursuing or not, to convince others through visualizing their ideas in a sense. So like we are learning much, much quicker, but we're not in a place where non-engineers are building products.

23:43Carlos: Got it. I think you are the first CPO, CPTO that actually shares that. Everybody else seems to be more on the everybody can ship, right? So it's refreshing to hear that. But I'm also curious to know, as you mentioned, there's a lot to be realized in the upcoming years. Do you see a future where the lines are truly blurring and we can have more non-technical people really shipping into production? So I can see a world in the future where if I am able to provide a perfect query input to let's say cloud code or cursor or codex or any of these tools, I will probably get a perfect output. So if I work backwards from like what does it actually take to provide this perfect query, it probably starts with having a really high quality bulletproof product spec in some sense and design system built in a way that is really repeatable.

24:42So as this query is generated, it can pull out from the design system, design elements that are necessary to build a feature. From product spec, it will have all the cases covered, all the edge cases covered in a way. So I can see the future. We're not there yet. And that's what I said before, probably two years at least, away from having non-engineers build products for us. Because I haven't personally been able to achieve this perfect query myself. I always have to go and adjust something or change something or tweak something on the output side. And I think you need engineering skills. So I'm hopeful that we're going to get there.

25:20And some people that I'm talking with, which are taking a little bit less conservative approach, more aggressive approach, believe that even if you create some tech debt now and if you lower your quality right now in exchange for moving faster, these tools are going to get better. They're going to be able to fix all those things for you. I tend not to trust that at this point. So I'm like, really, you know, let me take slower, more conservative approach. And then if the tools do achieve that level of quality and perfection, then I'm going to jump on the train. So we're not there yet. Let's see what happens.

25:56Carlos: And it's been a pleasure to have you on the pod. Thanks so much for joining us and for your honest takes. I think it's also very refreshing to have someone who hasn't been doing the rounds in the podcast who's really grounded in reality. Yeah, of course. Thank you for having me.