the ConversAItion: Season 5 Episode 31

AI and Online Dating: Tinder’s Match Made in Heaven

In today’s episode, we’re joined by Dr. Jennifer Flashman, Director of Analytics at Tinder, the world’s most popular app for meeting new people. Most people have heard of Tinder, but they may not necessarily be familiar with the technology that powers its hallmark matchmaking capabilities. Today, Jennifer will give us a behind-the-scenes look at the role of AI in optimizing personalized matches, how the pandemic turned the modern dating scene on its head, and how Tinder has evolved over the last year to meet the brand-new generation of daters.
Listen on Apple Podcasts badge
“Tinder gives people an opportunity to meet people who they wouldn’t otherwise meet in real life. Before, I would meet someone in a bar or through a friend or in my neighborhood or at work. There’s a set of places where I could meet someone, and that was constrained by where I was and my social milieu. Then, Tinder broke that open. None of those things matter now, because we can give you an opportunity to meet anyone.”
Jennifer Flashman

About Jennifer Flashman

Dr. Jennifer Flashman is the Director of Analytics at Tinder, where she oversees the AI and data science teams working to optimize connections made on the app. Prior to Tinder, she held data science roles at Microsoft and Hulu. A sociologist by trade, Jennifer earned her bachelor’s degree from Reed College and PhD in sociology from UCA. Jennifer can be found on LinkedIn here.

Short on time? Here are 4 quick takeaways:

  1. Jennifer was drawn to data science—and subsequently Tinder—because they helped her make an immediate impact on people’s lives.

    Prior to her time at Tinder, Jennifer was a sociologist by trade, and published more than 20 research papers. In fact, her research focus on friendship networks—how people choose friends and how those friendships impacted their behavior—is ultimately what drew her to the company. She was particularly intrigued by how Tinder gave users visibility into not only who they ended up with, but also who they could’ve ended up with. 

    This work at Tinder, and more broadly in the data science field, enabled Jennifer to instantly see the results of her work. Covering experimental design, user discovery experience and long-term strategy, Jennifer and her team are building algorithms that have the potential to democratize the entire dating process by expanding the pool of options—thereby bringing tangible value to people’s lives. 

  2. Tinder’s matching algorithms help users make meaningful connections.

    To date, Tinder has created over 60 billion matches. But Jennifer’s team isn’t only focused on creating a high volume of matches for users; they’re focused on creating meaningful ones. So in addition to user data—like photos, preferences and bios—Tinder’s algorithms keep track of who they “like” (swipe right on) or “nope” (swipe left on), as well as how many people “like” or “nope” them, to drive better potential matches. They also take into account how recently each user has been on the app, to ensure that new matches can have an immediate conversation. Most recently, Tinder introduced a new feature, Hot Takes, that allows users to message someone prior to swiping on them so that they can assess whether there’s an instant connection. Taken together, these AI-powered strategies help Tinder foster a high quality of matches at scale.  

  3. The pandemic changed the way we think about dating. 

    In spring 2020, Tinder took a close look at usage metrics in northern Italy, where COVID-19 was concentrated at the time, to understand how the pandemic was impacting dater behavior. The company noticed a dramatic drop in engagement metrics there—but that drop was largely driven by non-Italians: those who were studying abroad or on vacation in the country who left as the pandemic ramped up. Meanwhile, resident Italians—especially younger members and women—were actually driving a large upswing in overall engagement. This trend was quickly replicated across countries; as people spent more time at home, they focused on the hyper-local online dating sphere. The phenomenon was a fast lesson for Tinder in not simply taking data at face value, especially during a time of rapid change and uncertainty, and digging deeper into context. 

    In response to these new trends, Tinder made its Passport feature—which enabled users to place their location pin anywhere in the world, and swipe in that location—free to all. Tinder saw a hugely positive response, as users (virtually) explored the world again and scratched their travel itch.

  4. AI is poised to make online dating—and the ability to find the right match—more accessible.

    While there’s already been ample innovation in the online dating space over the past several years, Jennifer believes there’s even more to come in terms of accessibility. Prior to online dating, we could only meet people within our social circle, whether at work or school, or through a mutual friend. In contrast, Tinder gave us the opportunity to meet people who we wouldn’t otherwise encounter in our day-to-day lives. Jennifer expects that AI will continue to give users the ability to see an even broader pool of potential connections and help more people find their match. 

Read the transcript



Jim Freeze Hi! And welcome. I’m Jim Freeze, and this is The ConversAItion, a podcast airing viewpoints on the impact of artificial intelligence on business and society. 


In today’s episode, we’re joined by Dr. Jennifer Flashman, Director of Analytics at Tinder, the world’s most popular app for meeting new people. Most people have heard of Tinder, but they may not necessarily be familiar with the technology that powers its hallmark matchmaking capabilities. Today, Jennifer will give us a behind-the-scenes look at the role of AI in optimizing personalized matches, how the pandemic turned the modern dating scene on its head, and how Tinder has evolved over the last year to meet the brand-new generation of daters.

Jennifer, thanks so much for joining us today!
Jennifer Flashman Thanks for having me.

Jim Freeze So Jennifer, I understand you’re actually a sociologist, I should say, by trade and a well-published one to boot. Can you give us a sense as to why you pivoted to data analytics?

Jennifer Flashman Sure. I’d like to say that it was very intentional, but it was actually very accidental. I was a little bit disillusioned with the academic world, had a bit of a two body problem, cause my husband is a philosopher and also an academic. And I sort of randomly saw on a Listserv for my undergrad, this job that I thought I was fully unqualified for, and I reached out to the person cause I was like, well, I’ll get some information. He’s an alum, he’ll tell me a little bit about this. And he was like, oh, apply for the job. And I continued to insist that I was unqualified and he was like, ‘no, no, no’ apply, apply. So I applied and I got the job and I was like, well, I should try it. And it turned out to be this great fit that solved all of my problems.

Like my problems with academia were that it was so slow. I felt like I didn’t impact anything. I liked my research. But you know, when you spend two, three, four years working on the same project, it gets old and a little isolating, right. You’re sort of sitting by yourself doing all of these things. And it was like all of a sudden I found data science solved all these problems where I could work really fast. Nothing ever took more than two weeks or something. And I was working with this whole team of people and there were always people to bounce ideas off of. And I think I lucked into a great first role where I had a lot of freedom and autonomy, and I think people listened and it was just fantastic. So I wish it was intentional, but it was a really happy accident.

Jim Freeze Were there specific things about the role at Tinder as you kind of looked at it from the outside in, that drew you there? I know you’ve been there since 2018. Can you walk us through your day-to-day role and the team you manage?

Jennifer Flashman Yeah. So initially I was really interested in Tinder because it kind of helped me get back to a lot of the academic work that I used to do. So in my previous life we’ll call it, I studied friendship networks and academic achievement education. And you’re like, what, how is that related to Tinder? I was really interested in how people choose their friends and then ultimately sort of how those friendships impacted their behavior, being able to sort of see that choice was really interesting to me and sort of understanding why they made that choice and Tinder is actually very similar, right? We have created this app that shows you a bunch of people and you get to make a choice, yes or no, on everyone that you see. And so in contrast to the real world where most of the time we just get to see the yeses, right? You know, you see who people end up with, but you don’t actually see who they could have ended up with, but didn’t. 

On Tinder, we actually get to see that sort of trace process, that was the most intriguing thing about it to me initially when I came here. There were lots of other great things, obviously as well. There’s great people and a really nice culture and sort of fun atmosphere, but that data really brought me in, in the first place and that problem space, which is really quite interesting. So day-to-day, I am one of the directors on the analytics team, so I kind of cover three areas of the app: trust and safety, which is always very important, the core experiences of the app, which includes recommendations, that sort of chat experience that happens after you match, the general discovery experience that you have sort of seeing people and deciding whether to like them or note them all of that. 

So I cover that plus then some like core foundational stuff around performance and a lot of stuff that happens underneath it all. And so my teams are responsible for a vast array of things from any experimentation that we do in designing those so that we can learn and understand, as well as exploration to help us figure out, strategy for what we do down the line, what is working for our members, what isn’t and helping to build with our members in mind. So a lot of what we do is sort of using our data to help inform the future product decisions, but also sort of what is working and not working for our members, because ultimately what we care about is that our members are having a good experience in the app.

Jim Freeze Yeah, absolutely. And it must be really cool to work for a company that puts swipe right into pop culture.

Jennifer Flashman Yeah. I have to say it is super fun. So the first job that I had in this industry was at a company that was acquired by Microsoft and I would try to explain to people what it was. It took a long time, it was hard and no one really had heard of it or had heard of it sort of tangentially. What’s great about working at Tinder, it’s like, I have to explain it to nobody. Right. Everyone knows about it and it’s quite fun. It feels like it is everywhere. It’s often like the New York Times crossword puzzle. That’s always a fun thing.

And I think in the office too, we know we’re Tinder, right. I don’t think we take ourselves so seriously. Right. It is meant to be a fun app, a fun experience. And I think we try to bring that into our work as well. Our work should be fun and it’s not rocket science that we’re doing, but I do think it really does influence people’s lives. And I think that’s sort of this nice balance of like both a meaningful and fun and playful that can come together at Tinder.

Jim Freeze Well, let’s actually talk a little bit about the rocket science element of it. The role of AI in Tinder’s powerful matching system, how has it impacted the user experience? AI in particular?

Jennifer Flashman Yeah. So what we’re trying to optimize for is not just people getting matches, which is important because right? That’s the beginning of a connection and connections are kind of the key to what we’re trying to do at Tinder, but not just a connection or many connections, but trying to create quality connections. And making sure that when people are matching that they are actually building something meaningful with those connections.

And I think we’re doing that, obviously with changes and tweaks to the algorithm and the AI. But I think we’re also doing that with a lot of the new features that we have been developing, and have contributed to sort of how our AI works.

So one of the things that’s most important for our algorithm is actually how recently someone has been on the app, right? Like that’s the key to having a good experience, it’s terrible to match with someone who doesn’t even seem to be on the app anymore and so that’s not a great experience for our members. And so that’s something that we’re always really cognizant of. And that was actually the sort of inspiration behind the new feature that we’ve released recently called hot takes, which actually lets you talk to someone prior to even swiping on them so that you get to kind of have the tech experience with someone who is online right now. So the idea being that like we know you’re online, they know you’re online. So you can have an immediate back and forth and get a kind of immediate gratification that sometimes is a little bit light, I think in the experience for our members in this sort of standard traditional way of discovery in the app.

Jim Freeze It’s really interesting. So, over the course of the past two years, and virtually every episode, we’ve talked about the pandemic and with social distancing and shelter in place orders, the pandemic has had a huge impact on everything, including dating and people turned on online apps more than ever. What trends did you and your team see and how did Tinder adapt to a new generation of daters?

Jennifer Flashman Yeah, so that’s actually really interesting, really early on in the pandemic, like just when it started to spread outside of China, we started looking really closely at our metrics and really locally trying to get a sense of what might happen. Right. I think everyone was concerned. And so like for example, we initially looked really closely at Italy, which was one of those first countries outside of China to lockdown. And we immediately saw these dramatic drops in engagement metrics, especially in Northern Italy where COVID was concentrated at that time.

And I think initially we were like, oh no, what does the future hold for us? This is actually a great example of why, and when one shouldn’t always take data at face value. We dug in a bit further and discovered that actually a lot of the drops had come from non-Italians. So these were presumably non Italians leaving Italy. So either they’re studying abroad or on vacation or something, and this big thing started to happen and they left. This really kind of drove down a lot of our engagement within Italy when we were looking really locally.

But when we look just at Italians, it actually was fairly steady, normal engagement. And so, I mean, I think what we sort of saw here with like right at the beginning of the pandemic, it kind of upended the way that people exist in the world. And in this case, it upended travel and we were actually seeing the result of that in, in our numbers. And so we’re seeing that shift from like a very global world to a really, really local hyper-local world. That was sort of our first look at things. And then over time, as people were spending more time at home, we saw a real upswing in engagement. So our members were generally swiping more messaging more, spending more time on the app. Those trends were actually driven largely by our younger members, like those 18 to 22 year olds and women in particular, which I thought was very interesting.

In response to a lot of these trends that we saw, we have a feature called passport, which basically lets you place the pin anywhere in the world and lets you swipe in that location as if you were there.

Jim Freeze Interesting.

Jennifer Flashman Yeah. So we made that free to everyone that April, right after the pandemic got moving and we got kind of a huge response to that, which was super interesting. With our small attempt to help our members scratch that itch to get out of their house and travel and at least kind of virtually meet people around the world.

Jim Freeze What a great idea. Yeah.

Jennifer Flashman Yeah. And we saw that in our numbers a lot.

Jim Freeze Just to kind of wrap up, how do you see the role of data and AI evolving at Tinder and maybe just more generally speaking, dating apps in general over the next four to five to 10 years.

Jennifer Flashman So the thing that I love most about Tinder is that it gives people this opportunity to meet people who they wouldn’t otherwise meet in real life. So, it kind of opened up this world where, before Tinder, right? Who was I potentially going to meet for a date? Well, someone who I met in a bar or a friend of a friend or, in my neighborhood, at work. Right. So there’s sort of like a set, set of places where I was going to potentially meet someone, at school. Right. And, that was sort of constrained by where I was and my social milieu and all of that. And then Tinder kind of broke that open and said, Hey, none of those things matter, we will give you an opportunity to meet anyone.

And you know, obviously you can put preferences around exactly who you’re meeting in the sense of how far away they are from you, what gender they are, that sort of thing. But it’s still like creating this really neat experience where you actually have this huge opportunity that really never existed before. So when I think about what role AI plays at Tinder, or just more generally, I want to make sure that it’s continuing to fuel those opportunities. So not kind of, sort of narrowing those opportunities, but really giving everyone the opportunity to see that broader world or meet the people who you wouldn’t necessarily have the opportunity to meet if you walked outside of the door of your house.

Jim Freeze That’s an inspiring future. I think for many people.

Jennifer Flashman For sure. And I think every time, like we think about how we optimize our algorithm or how we use different, the information that people provide us in their profiles, for instance, in that algorithm, that’s something that I think we need to have in the front of our minds, making sure that we are still enabling that access and that opportunity.

Jim Freeze Absolutely. Jennifer, thank you very much for joining us. We loved hearing about your own journey into analytics and kind of the evolving role of AI in the online dating space. It’s been a pleasure having you on today’s episode

Jennifer Flashman It was great being here. Thanks for having me.


That’s a wrap for this episode of The ConversAItion. In our next episode, we’ll be shaking things up a bit with Dr. Susan Hura, Interactions’ own, Director of Conversation Design Services. Susan and I will sit down virtually to discuss the growing role of AI in customer experience today, and what it takes to build a truly conversational experience.

This episode of The ConversAItion podcast was produced by Interactions, a Boston-area conversational AI company. I’m Jim Freeze, and we’ll see you next time. 



Check out more episodes of The ConversAItion.