Spencer Rascoff on turning around Tinder: from holding company to operating company, with AI sprint cutting product timelines from 12 months to 2

Speaker 1: Well, congratulations on OCC approval, and thanks for stopping by the show.

Speaker 5: Great to see you. We'll talk to you soon. Thank you for fighting the fight. Have a good one. Dollar. Yes. Cheers.

Speaker 1: Keep the dollar strong. Great to

Speaker 5: see you dude.

Speaker 1: Goodbye. Up next, we have Spencer Rascoff from Match Group. Welcome to the show. Hi. Fantastic entrance. Very quick. Sometimes I have to do a whole ramble

Speaker 11: Very sneaky.

Speaker 13: You are.

Speaker 1: But for those who don't know you, please introduce yourself a bit. I'd love to go back a little bit about your career and then I want to see where Match Group is going.

Speaker 11: Yeah. Great to be here and thanks for what you do for technology and also for the LA tech community. Yeah. Is. Fun. It's very

Speaker 5: long L. A. Of us to

Speaker 11: do this here. I appreciate it. I grew up here and I've I moved back here about ten years ago and have been all about promoting the L. A. Tech community. I started a company called Dot LA which is a media company to promote Oh, wait. LA

Speaker 3: You started that. That's

Speaker 11: Yeah. So, yeah. Quickly about my background. I did some time on Wall Street

Speaker 7: Yeah.

Speaker 11: And then broke free from that.

Speaker 1: Thank you for your service.

Speaker 11: Thanks for your

Speaker 5: not enough people.

Speaker 11: Wants to do something more entrepreneurial. My first startup was hotwire.com which we sold to Expedia. Yeah. Then I was at Expedia briefly and then I left to start Zillow. I was the CEO and and co founder of Zillow. Ran it for fifteen, sixteen years.

Speaker 3: Yeah.

Speaker 11: And then kind of retired for a couple years. Angel invested in new startups, incubated new startups mostly with former Zillow Yeah. And then stepped in as CEO of Match Group about a year and a

Speaker 10: half ago.

Speaker 1: What does that recruiting process look like? Like, you know, it's like they could have gone with someone who was actively a CEO. They tried to pull you off the beach in some ways. Like, what's that conversation like? Why you I mean, you're clearly capable and

Speaker 11: Well, was on yeah. I was on the board. Okay. And I'd been on the board for about a year.

Speaker 5: Okay. Okay.

Speaker 11: And when the opportunity presented itself, the board turned to me and said, you know, what do you think? And I immediately jumped at it. So I had been watching the company for more than a year from the boardroom. It's an incredible company doing really important work. I actually I feel very personally motivated by the mission of the company. Mhmm. There is a loneliness epidemic. Yep. It is a real global problem. It's existential. Yeah. And it impacts longevity and mental health and like everything about how society works. And Match Group is one of the very few companies as the leading dating app company that can do something about And so that's why I got

Speaker 1: So you're on the board you're watching. When you choose to step into the CEO role, are you do you have a ninety day plan? Do you have a big transformation plan? Like, what were you going into? Because it's Yeah. Presumably a of showing up tabula rasa here.

Speaker 5: Yeah. A lot of your best ideas. Hopefully, you had been Yeah.

Speaker 1: You've been like, hey, I told you guys to, you know, acquire this company or, you know, change this or whatever.

Speaker 11: I have this this kind of mantra and I I write it on the board. I put it up. There are stickers around the company of it, posters, etcetera. It's great people, properly motivated and properly organized Mhmm. Build great products that are informed by user research that then when properly marketed generate a lot of audience and revenue and profit and then shareholder value. So So the entire business. So it starts with people. Yeah. It starts with people and motivation and organization. So that's where I started. Yeah. That was the beginning of of the ninety day plan. Who's here? Why are they here? How are they organized? And so for example, I immediately started by breaking down silos. Yep. Match Group was basically a roll up.

Speaker 1: I was gonna ask you about it. Feels like the ultimate siloed organization.

Speaker 11: It was. It was. It was basically a holding company and I've transitioned it to being an operating company.

Speaker 5: Okay.

Speaker 11: So Barry Diller's ISC bought the cat know, bought a lot of the companies in the category. Started with match.com in Dallas, then Plenty of Fish in Vancouver, OkCupid in New York, Tinder in LA, Hinge in in New York, Pairs in Tokyo, Meetic in France and on and on. And there are 25 brands that have been assembled. And then it spun out to IAC shareholders about six or seven years ago and a lot of the hard work around integration was never done. So that was where I first focused my attention was on people, organization, motivation, employee engagement, and then we moved on to product. And at that point, I took on the role of running Tinder myself.

Speaker 5: Yeah.

Speaker 11: Because Tinder well, obviously, with a with a great team, but being essentially CEO of Tinder. Yeah. Tinder's the number one dating app.

Speaker 5: Guys, I got this.

Speaker 11: It be? Right? So it's the number one app in a 160 countries.

Speaker 1: And writing marketing copy.

Speaker 11: I haven't gone I haven't gone that far. I I that's that's where I team. We still have a great team. But look, it's it's it's almost 2,000,000,000 of our 3 and a half billion of revenue Yeah. And it's by far the largest dating app worldwide. Yet it had sort of lost its way in terms of innovation. Sure. It hadn't been founder led for a long time. Yeah. It hadn't prioritized user outcomes. It hadn't had an innovative product road map. Mhmm. And so we recast all of that about six months ago and we've made enormous progress in just six months. And I've shared a lot of the metrics last week on earnings. We are turning around Tinder and it feels great.

Speaker 1: That's amazing. How do you think about de siloing? Because I I I guess my big question is like how important is the are the individual brands? Because I imagine that's the last thing you desilo is like, oh, if you'rea match.com user, like now surprise, you're just it's the Tinder brand because you could do that. Like it's totally possible to acquire a company, change the name, maintain most of the user base. But you that feels like the last thing that you would do. But why?

Speaker 11: Yeah. What we've started by doing is integrating a lot of the back ends.

Speaker 1: I'm sure. Like finance, legal, HR, and then also databases.

Speaker 11: But even the back ends of the product. Sure. So so for example BLK, our app for black daters and Chispa, our app for Hispanic daters and match.com. Yep. Like all of these now have an integrated back end. Sure. So we launched a new feature Yep. And it goes out across a dozen or so

Speaker 1: different So like one cloud platform, one set of SREs and engineers

Speaker 11: What we haven't done yet Yeah. Is some of the liquidity sharing that I think maybe you're Sure. At which is, you know, should we show a Tinder user to an OkCupid user or a Plenty of Fish user to a BLK user. We've done a little bit of cross sell.

Speaker 1: Okay.

Speaker 11: And and that's been very effective. So you'd be on one of our apps and Yeah. You'll get a message that says, hey, you've been invited to join the league. Tap here and now your profile is kind of ported over into the league.

Speaker 1: Got

Speaker 11: it. And now you're searching in league. That's driven a lot of audience growth and revenue cross sell, but that's kind of where we started.

Speaker 1: What is the dynamic of being this roll up? Because is there some sort of adverse selection or some weird market dynamic where there are founders or entrepreneurs out there who basically say, oh, if I start a dating site and I get anything going, I'm gonna be able to sell to these guys because they want everything. Is that something that happens?

Speaker 11: Yes. It does. And I mean, I guess I would describe it this way. Every

Speaker 5: Has has anyone sold multiple companies?

Speaker 11: To us? No. No. Not that I'm aware of. There's no Nikita Beer of

Speaker 1: because there's a couple there's a couple of these serial entrepreneurs in Silicon Valley that do this type of thing. I mean, you'll see it in like cybersecurity or enterprise software where it's like, oh, he sold three companies to Google that do basically the same thing.

Speaker 11: Yeah. And then we haven't had any repeat repeat sellers. But look at Zillow Group Yeah. We bought 17 companies Yeah. During my time as CEO of Zillow Group. So far, we've bought kind of one and a half in my year and a half here. So we bought Her, which was the leading lesbian or sapphic app.

Speaker 8: Okay.

Speaker 11: And then we invested last week a $100,000,000 in Sniffies which is the number two app for non heterosexual men.

Speaker 1: Sure.

Speaker 11: With a right to buy the rest of that company. Sure.

Speaker 5: Sure.

Speaker 11: We made those two moves so far. But you know there are a lot of startups in this space. There always have been, there always will be because every 15 22 year old starts dating and they think, oh these dating apps are are not great. I can build a better one. And it's easier than ever to build a new app or or website obviously. So I think the competition will keep coming. The challenge of course is this is a network effects business and people tend not to want to use a new network because there's a cold start problem. So even if a new app has great functionality There's

Speaker 5: a network effect very clearly but at the same time is there also some advantage that new platforms have?

Speaker 11: It can be sort of a nightclub, know, of a hot goosebumps.

Speaker 5: You're the one there. The first one there, know, big fish in a

Speaker 11: I mean once once you start solving the cold start problem, then I think new brands can take on some momentum, but it's very difficult at the very beginning.

Speaker 2: Yeah.

Speaker 11: This is of course, you know, the the kind of the put and the take with a brand like Tinder, which is an older brand, has massive brand awareness. Everybody knows Tinder. Everyone also has an opinion about Tinder. And so now we have to work really hard to change the product, improve the marketing to try to regain resonance with Tinder. So features like Double Date, which are doing amazing. About a quarter of Gen Z users on Tinder now use Double Date where two friends kind of join up, create a joint account and then link their profiles. Now they're swiping on pairs and then there was four way

Speaker 1: grouper that was doing something similar to that with three and three.

Speaker 11: That's right.

Speaker 1: And I don't know where that wound up

Speaker 5: but So so We didn't buy fascinating. Every company you've mentioned so far seems very demographic focused. How are you thinking about new technology? I feel like there's a bunch of people that are thinking about how to use AI to create dating app experiences. I imagine you guys are already using a bunch of AI.

Speaker 11: We use AI constantly in the product. There are a lot of startups that are creating AI only products and we actually incubated one with Justin McLeod, the founder of Hinge and then spun it out. So we're the largest owner of a company called Overtone which hasn't launched yet but the founder of Hinge is is the founder there and we're the biggest shareholder.

Speaker 5: That's

Speaker 11: building an AI specific product.

Speaker 5: We're like, let's spin it out. We'll buy you back in a few

Speaker 11: So there may be that might be the the first repeat seller. What

Speaker 1: does AI only mean? It means you're dating AIs?

Speaker 11: No. No. No. No. No. We we we don't support that. It it means that your experience is one where kind of it's AI guided. So rather than you reviewing lots of people and holding your hand along

Speaker 3: the way.

Speaker 11: We use AI in all of our products in so many different ways.

Speaker 3: Course, even

Speaker 1: just to sort the stack of matches on Yes.

Speaker 11: The recommendation algorithm day one. And ML powered.

Speaker 2: Yeah. Of course.

Speaker 11: We use AI for profile creation. Sure. It can be very daunting to look at kind of a blank. Yeah. Tell us about yourself. And and so we use AI to tease out information about your profile.

Speaker 5: Have you gotten pretty good at detecting deepfakes?

Speaker 11: Yes. So that was one of the first things that I focused on is, something called face check where you have to record kind of like a video selfie of yourself and that's reduced interactions with fake accounts or deepfakes by about 60%. So we've rolled that out in almost every country globally across most of them, our brands.

Speaker 3: Yeah.

Speaker 1: I saw a fascinating video of how people are trying to get around it by having nine fake images, defaked images of some person and then the real person who's operating the account will like style transfer their face to look like a like a renaissance painting and you will just assume that it's like, oh, this hot person that looks exactly like, you know, it's nine AI photos. It's like, oh, they like painting and they happen to paint some random guy. But then that's person that IDs the paycheck. So I imagine it's a cat and mouse game.

Speaker 11: Yeah. It's a cat and mouse game. The bad guys are pretty good but we're better. Yeah. And you know, we have to stay a step ahead. We've made a lot of progress and and How

Speaker 1: valuable is like community feedback in that loop? Because I imagine that like if somebody if there's a if there's a system that you build, someone's always going to try and find some edge around. But like I've been surprised that like the community notes system on x has been pretty robust. Yeah. And it's been like you can always just ask the LLM, like, is this real? But the community notes often surface things in interesting ways.

Speaker 11: Absolutely. So user reports Yeah. Behavior or fake accounts, that's an important signal for us. And now we have a system that actually kind of cross bands. So if if we detect a fake account on one app, we take it off of all the accounts Mhmm. All the apps. So that also improves trust and safety across the whole Match

Speaker 1: Group portfolio.

Speaker 5: I have a buddy, a startup founder probably eight years ago. Thought he'd he'd be cute and try to market the startup that he was working on on one Mhmm. One of the Match group services got banned and then got banned from everything. And so he's like, I guess I'm destined to be alone

Speaker 11: in this cold world. Sorry about that. We don't don't take kindly to to promotional or fake accounts. Yeah.

Speaker 1: Yeah. What what have there been companies that have been trying to sell services on top of your platforms like auto swipers and stuff?

Speaker 11: There are yeah. There are some startups that that try to do that.

Speaker 1: I wouldn't even necessarily call them startups. They're like, You know, I guess sometimes they might be, but they might just be like operations.

Speaker 11: Yeah. There there are. Yeah. We don't allow kind of automated Yeah. Like, you know, scrapers or Sure. Auto, you know, auto swipers or those types of things. Sure. So when we find out about those, we take them out. Yeah. Yeah.

Speaker 5: Do you think the AI boom has hurt LA's prospects as a startup hub?

Speaker 11: Oh, boy. Yeah. It's I mean, we do

Speaker 5: It's like we were sitting around prior to starting the show. We're sitting here like we're we work in tech

Speaker 10: Yeah.

Speaker 5: But we don't want to leave LA.

Speaker 11: Yeah.

Speaker 5: What possible business could we start in LA? It's not hard tech.

Speaker 1: We have this idea, media.

Speaker 11: Media. No. This is this is kind of this is what makes LA LA. Mean, like, couple years ago when creator economy was booming, even even crypto was was, you know, LA had a crypto moment and NFT moment. And then also when direct to consumer and like celebrity food and beverage brands, those types of things when D to C was was hotter, I think LA was booming even more and clearly we've lost a little bit of market share and mind share to the Bay Area Sure. As this boom has happened. So it's a bit of a bummer for those of us in LA. But you know, we're we're it's

Speaker 5: a Yeah. Very simple In some ways, number the of LA which is like entertainment, the consumer Yes. Media, hard tech, right, just like sort of continues and it's very hard to But it but I I I think, you know, the the long LA contingent deserves credit in some part for El Segundo.

Speaker 11: Totally. A SpaceX or Andoril IPO will be massive. Service Titan was a was a nice moment. Obviously, the Honey exit was a nice moment. Yeah. But, you know, once SpaceX gets public, I think that will be pretty amazing for the ecosystem here. Obviously, much of some of it's moved to Texas, but there's a lot of of pent up wealth which will become recycled into angel investments

Speaker 1: Sure.

Speaker 11: Into the rest of the LA tech ecosystem.

Speaker 1: Yeah. If you look through or you imagine looking through like all the KPIs that you look at over a quarter or a year, are there any metrics that have shown the clear, like, pre AI, post AI, like, kink in the graph? Like, we were looking at the number of books published on amazon.com, and it's gone from 100,000 a month to 400,000 Clear AI effect. You can see it just grow. Now is that good or bad? It's probably fine because they have recommendation algorithms that can filter that stuff out. But I'm just wondering if there's any like, okay, we can I imagine you can like see COVID in your chart?

Speaker 11: Oh, for Yeah. You saw COVID and then you saw kind of the the COVID hangover.

Speaker 1: So so COVID was a spike for you because everyone was sitting at home, they're

Speaker 11: on their couch, getting their phones and then lonely and

Speaker 8: Yep. Out of connection.

Speaker 11: Yeah. And then people went back to work

Speaker 2: Yeah.

Speaker 11: And kind of dating apps fell out of favor. Well, one internal metric is just pull requests now that I'm we're doing so much internal AI coding. Yeah. Pull requests are up they were up 40% year over year a couple months ago. Now they're up 60% year over year. So Yeah. Basically the amount of code that our employees are writing is Yeah.

Speaker 1: Really accidental. The question I have for all the CEOs that are seeing incredible gains with, like, AI agents and coding agents, like, the PRs don't pay the bills, you know? Like, they don't Yeah. Like and I and I'm always wondering, like, you know, Meta was spending, like, billions of dollars token maxing, and I think it's great. But when I open up Instagram, I'm I'm like, is it better? Like, I I haven't seen a new button.

Speaker 11: It's a

Speaker 1: great question. And so so I imagine as CEO, you have to sort of grapple with this. I'm super optimistic, obviously, but

Speaker 11: We're all we're all going through this right now. So so our approach to it is we we have to pay for these AI tools some Yeah. Course. You know, for us, it's 5 to $10,000,000 a year that we're spending on AI tools from basically nothing Yeah.

Speaker 1: Eighteen months ago. New opex.

Speaker 11: So we're slowing down hiring somewhat. Yeah. And that's how we're choosing to fund it. Sure. But in terms of when do we start to see the actual benefit, you know, CEOs and CFOs are are starting to ask that question more and more. I'm asking it more and more. I I I think we're benefiting from it. It's hard to it's hard to feel it.

Speaker 1: Yeah. Feel like the the the optimistic cases that in a few quarters we're seeing case studies like

Speaker 11: Yeah.

Speaker 1: Okay. Yeah. We had an engineer run a bunch of a b tests, automated, running overnight. And we found out that like the pricing on this page was better this way and it actually created a lift in revenue or something like that.

Speaker 5: Yeah. Or or a SaaS company says, hey, we added this product that was originally going to be we were gonna build it in three years. We built it now and it's ramping revenue and paying for all of this token.

Speaker 11: Our our best example would be, the January, we decided to to as part of the Tinder turnaround to focus on in real life events. We know that young people in particular really want to meet people in person. They want to get off their phones, get off the apps. And so January, we made that decision. March 12, we had a product event Mhmm. Here in LA that demoed all these new changes to Tinder. And between the January and March 12, so in just a couple months, we went from a standing start because of AI Mhmm. To shipping

Speaker 1: Yeah.

Speaker 11: In app, in real life events Yeah. In LA, and we never could have pulled that off without AI. Yeah. Was an AI sprint that made a product launch in Yeah. Two or three months, which would have taken six or twelve traditionally.

Speaker 1: Is that productivity improvement changing your or potentially changing your thoughts around like incubating new properties? Like I'm imagining you could go much more niche. You mentioned a few of the niches. Yeah. But there's like famously like Farmers Only, which I don't know if you own.

Speaker 5: Like We

Speaker 11: don't own that one.

Speaker 1: We don't own that one. The

Speaker 5: one that got away. But

Speaker 1: that's an example of like a much more niche community Yes. That the maintenance and the OpEx associated with running it at Match Group probably too high. But if you drop that by 10 x with AI, then maybe you can go after like, you know, like Hollywood media executive dating site or something.

Speaker 11: Yeah. Mean, Upward is an example of this. So Upward is our our Christian brand but also kind of traditional values, which is something that obviously has a huge addressable market

Speaker 1: Yeah.

Speaker 11: But it's not a 100% of the population mainstream.

Speaker 12: Yeah. Yeah.

Speaker 11: Hinge, for example

Speaker 5: Sure.

Speaker 11: Which we own goes after kind

Speaker 1: of Yeah. Yeah. All of And was that homegrown?

Speaker 11: And Upward was an incubation Hinge. But now because of AI Yep. We're able to put more resources into it Yep. And accomplish a lot with five, ten, 15 people

Speaker 1: Yeah. Which And then it just becomes an audience acquisition problem which you have solved because basically everyone

Speaker 11: Yes. We can cross sell people into it but we also can have a pretty innovative road map Yep. With a small team Yep. Because of AI.

Speaker 1: Yeah. So you can actually roll out

Speaker 5: all of Is Instagram Hinge's biggest competitor?

Speaker 11: No. I think well, if by Instagram you mean people meeting each other on Instagram? No. If by Instagram you mean people sitting on their couch and Oh. Doom scrolling and choosing not to get out there and date. Yes. So I mean our biggest competitor at Match Group is is just inertia. It's just people's loneliness and you know, sort of doom scrolling on TikTok or Instagram or other social media, even Netflix or YouTube. Yeah. We need people to put their phone down and put themself out there and go try to meet new people. That to us is a you know, would be a flip of a headwind into a tailwind. And generationally, that's a little difficult. A lot of Gen Z is is pretty nervous about meeting new people or or isn't sure it's worth the return on their time or or maybe in some cases didn't develop the social skills because the pandemic caused them to miss a couple critical years and so it doesn't come as naturally. So this is starting to change and the double date is an important part of this. In real life events are an important part of it. That's why they've been so central to the Tinder turnaround.

Speaker 5: Have you thought about making products in the wedding category at all? Would you ever do wedding registries? Prenups.

Speaker 1: We We got the pipeline.

Speaker 6: So Dabores is a new member.

Speaker 11: Welcome back. A new member of our board of directors is is Reyna Moskowitz who's the CEO of the Knot Worldwide. Yeah.

Speaker 5: You go.

Speaker 11: You So so no, we're Matchgroup isn't doing it but we're very lucky to have Reyna's expertise as as the leading

Speaker 1: Yeah.

Speaker 5: If you reactivate an account on any of the platforms after ten years, divorce support.

Speaker 1: Add on. What about advertising? I mean, I know many of the platforms have pro plans or paid plans. But what is unique about you have a lot of attention, you have a lot of eyeballs. What's the story been with

Speaker 11: with We a pretty relatively speaking small advertising business considering how big our audience is. Yeah. We make less than a $100,000,000 on advertising Yeah. Out of our almost 4,000,000,000 of revenue.

Speaker 5: It's We love advertising. I'm sorry.

Speaker 11: You know, I mean the way we make most of our money, almost all of our money is subscription revenue where users pay for extra features, extra You know, some of our competitors have really leaned into advertising and I think advertising can improve the user experience but sometimes it can't.

Speaker 1: You need to be very careful. There's two classic deals on ad monetization. Like one is that, you know, like Twitter never got good at ads because people aren't in that like lean back shopping mentality. And then the other take is like it's an engineering problem. And if they just had the meta engineers building that matching algorithm, do you see one of those being more important than the other?

Speaker 11: I mean, I think from my Zillow experience, when we had a pretty big ad business at Zillow. Yeah. And when we had endemic advertisers Yeah. The, you know, the the moving companies Yeah. The mortgage companies, the title insurance companies Yeah. Tightly integrated into the product, that was, you know, that was ads as product

Speaker 1: Got it.

Speaker 11: Ads as improving user experience. In this category, we also have done some of that. Sure. And the ads, for example, on Tinder are integrated in so you right swipe on an ad, you left swipe on an ad the same way you would right swipe or left swipe on a user. So it because it's endemic in that way and it's it's pretty tightly integrated, the ads perform very well. But we're much more focused on driving user outcomes and getting people out on dates. Yep. And so the subscription revenue

Speaker 1: More aligned with

Speaker 11: the user. Is more aligned with user Yeah.

Speaker 5: Did you ever invest or do anything around the housing problem we have in this country given that you're at Zillow?

Speaker 1: Quickly, can you solve the housing crisis for us?

Speaker 5: No. No. I mean, I I I would like I would like you to take a run at that at some point. Yeah. You know how much demand there is.

Speaker 11: You know, if there's obviously an affordability crisis, there's mortgage rate lock in. If we could find a way to make mortgages transferable Oh,

Speaker 1: yeah.

Speaker 11: That would be massive. Tons of people are kind of stuck in their homes at two to 3% mortgages. They're not in the right home that's right for them anymore

Speaker 1: Yeah.

Speaker 11: They can't afford to fit two to 5%

Speaker 5: or six Right?

Speaker 1: If you bring the mortgage with you?

Speaker 11: I mean, it's you basically can't do it. You have to there are some startups that have tried. I looked at incubating some startups to do it. But you all the stars have to align. The buyer has to match your credit. It can't be done essentially in the current regulatory environment. That would be a massive unlock. More real estate development, so just more builder friendly legislation would be very important. That's something that some other states, not California, have done a better job of. California has been a laggard in that way. You know, I have incubated and invested in a lot of startups that help real estate agents use AI to be more productive.

Speaker 1: Mhmm.

Speaker 11: I do think that's a bit of a limiter as well. There's one called House Whisper which I started with several former Zillow people which makes an AI assistant basically for real estate agents and and gives them incredible AI superpowers. So that's a part of the unlock as well. But, know, we don't have a demand issue. We have a supply issue in The US. So we need more supply. We need more houses built. We're missing about a million homes

Speaker 1: Mhmm.

Speaker 11: That basically weren't built after the financial crisis. This goes all the way back to to '20 to 2008. Yeah. Builders were building a million homes a year and they went down to two or 300,000 for four or five years.