Hinge’s “good churn” connects 50,000 dates a week, and more unlikely startup lessons

Almost no company could give me a lesson on love and data better than Hinge, one of today’s most popular dating apps.

The modern day matchmaker has been very successful in digitizing an old fashioned way of meeting people: through mutual friends. But their popularity doesn’t preclude them from having a very complicated relationship with a metric most companies don’t want to see grow: churn.

Devin Markell, Hinge’s data lead, explains if you want to help your users and grow a product, KPIs like engagement, retention, and churn are table stakes. This probably doesn’t come as a surprise, but Devin emphasizes that interpreting these metrics is more complicated than people realize. The success of your app hinges (ahem) upon understanding the context of those metrics, and that’s where things can get tricky.

Hinge’s success, like most dating apps, is paradoxically linked to losing customers. After a user joins and finds love through Hinge, he or she typically then churns from the app. And yet, those “churned” users are also happy customers who refer Hinge to their friends when asked the classic new couple question: “So, how’d you meet?”

In fact, since launching in 2011, Hinge has been the Yenta to 50,000 dates per week, 3,000 of which turn into a relationship. Successful users churn out of the app every day. When it comes to fueling product growth, Hinge is proof that not all churn is bad churn.

Devin explains how products can have “good” and “bad” churn, and contextual data plays a crucial role when interpreting retention. For example, at Hinge, a higher churn rate wouldn’t necessarily be a reason to run to a pint of Ben & Jerry’s. If you’re a product manager or mobile developer, it’s time to commit: stay faithful to only reading metrics in context. Otherwise, you run the risk of misinterpreting the signs.

“Even if a user’s engagement is high on Hinge, but he or she isn’t ultimately connecting with new matches, we see that as a problem. At Hinge, we don’t prioritize mere engagement. We care about getting people connected, chatting, and swapping phone numbers,” said Devin.

“It’s important to remember that engagement numbers don’t always mean success. Product teams need a greater understanding of users’ journey to determine if they’re trending successfully, or not. That’s where additional data, like qualitative information, provides kernels of insights.”

It’s time we stop having black-or-white thinking around key metrics. The data behind customer acquisition or attrition is multifaceted, and understanding the numbers heavily relies on context, which often can’t be found with mere quantitative measures. No matter the company, data always requires an explanation and a context.

From Devin’s perspective, it’s time to embrace context around churn, retention, and other key metrics, instead of blindly falling in love with a single report or “magic” number. When trying to lock down a committed relationship with your users, Hinge proves that context is key.

In the world of data-driven product development and mobile apps, “engagement” metrics are key indicators for the health of your product. Reports on MAUs, retention, and churn inform major product decisions and strategy. However, high engagement rates (or, conversely, low churn rates) don’t always equate to success for your user.

For example, with a mobile gaming app, measuring and continuously increasing in-app time is crucial for success. For apps and products that connect the online and offline world, however – think dating, restaurant reservations, food delivery or ride sharing, etc. – in-app time may not be as important. Instead, success often means getting people off their phones and into the grips of an IRL (and untrackable) moment.

In fact, getting your users off your app and fulfilling their needs will likely increase their loyalty to the product. The reality is that some apps don’t need to be used all the time, but that doesn’t mean that they won’t be the top choice when a user needs them.

“If our users see Hinge as a dependable way to meet potential dates, then we’ve done our job,” said Devin.

“At Hinge, we measure almost everything because it’s how we track and understand how people are actually using our product. We call it success when we bring value into a user’s life, like connecting them with a match and giving them a place to message and get to know another person before meeting offline,” said Devin.

“Tracking everything is crucial, especially as you get started in building a product. Having metrics you can compare across different cohorts is important to know,” continued Devin. “What are the peak usage times? Are there points of seasonality or dynamism during the week or day? When are the big dips where people drop off in your onboarding flow?”

Metrics may seem somewhat straightforward to calculate, but there’s nuance. When you understand how your users behave, you can optimize the product so users keep coming back, over time. It’s important to track baseline behavior; otherwise, you can’t recognize how activity changes or improves as a team adjusts a product.

Sure, someone might only order delivery from an app once a month or check his dating profile during commute times, but the measure of users’ commitment isn’t in their obsessive usage; rather, it’s in their willingness to return.

With that in mind, it’s important for a team to define what the user’s ideal experience is and what an app’s role is in that user’s life.

Ultimately, your product is successful when your user is successful, even if the metrics don’t fit within a few neat-and-tidy reports.

“High activity on Hinge doesn’t necessarily equate to us being helpful for our users who are looking to date new people,” Devin said. “What we care about is facilitating connections, prompting conversations and seeing phone numbers exchanged, so people can build new connections in their life.”

“Ultimately we want to get to know our users as well as possible to give them matches that inspire them to continue to make more connections.”

This is where measurement gets blurry. When the effectiveness of a product like Hinge means people leave the app and enter into relationships, engagement and churn become hard to define.

The answer won’t be found by looking only at quantitative or qualitative data. In order to understand the user journey and whether users were successful, a product team must blend both.

Here’s some real-talk from Devin: “In actuality, there’s no single data source that will give you all the answers you need.”

According to Devin, Hinge never interprets its qualitative or quantitative data in silos. Instead, the team has made it a practice to compile all relevant data sources. This creates a context for what’s happening when users fire events in the app. Combining data also helps determine if a product is making traction against set goals.

“You want to have a flexible analytics tool to be able to go ask deeper questions, but you also have to have a system that’s going to start giving you answers right away,” he said. “Start with answering the big questions first, through event-based tracking, and as long as you’re able to go back to that event stream and input additional information, like survey data, that’ll help you achieve both immediate and long-term goals.”

Let’s take a hypothetical scenario: the quantitative data tells Hinge’s product team that 20% of users churned in a month. But then, a survey indicates that, of those who churned, 70% were people who started dating someone within the month through the app and thus deactivated. What looks initially like a negative is actually a positive since the app was successful for most churned users.

Knowing the strengths and weaknesses of both types of data helps you make more accurate decisions. Quantitative data, like event tracking, is the hard evidence of what users did in your product. Qualitative data, such as surveys, customer interviews, anecdotal feedback and metrics, like Net Promoter Score (NPS), tell you why they took a certain action.

“For us the ultimate goal is, how can we drive people offline and hopefully enter a successful relationship,” said Devin. So when churn amounts to matchmaking, Hinge chalks up this type of churn as a win.

When an app like Hinge has done its job well, users may not need the service anymore. However, those who leave the app happy often go on to give the company a high NPS rating, a huge indicator of a product’s virality. Perhaps most crucially for growth, though, “good churned” users can contribute to evangelizing a product through word of mouth.

When a user experiences “good churn” and shares their success story, they become a walking case study.

When talking to Mixpanel, Devin shared how Hinge highlights their “good churn” with a Wall of Love in their New York Offices. They celebrate the relationships forged through Hinge with photos of couples who have said thank you to the dating app.

Good churn is the starting point for evangelism and effective word of mouth marketing.

A user might’ve deactivated on Hinge, but he’s probably told several close friends about his success and encouraged his single friends to join the app. When channeled correctly, good churn is more beneficial for a product’s growth in the long run.

In a recent interview with Andrew Chen, the long-time growth hacker now at Uber, told us: “Inevitably for the really great products, word of mouth is such a big part of [growth]. A product can optimize what it can optimize, but these offline interactions are a big part of it.”

“You just have to accept that the really great products have tons of unattributable traffic and that’s actually a good thing,” Andrew continued. “It’s an honor to have that. And it’s not annoying, it’s actually really great. In fact, the products that, for example, buy all their customers and their traffic is all attributable, is a sign of a weak business.”

Hinge is happy to report that it has limited paid marketing and sees a lot of traction through word of mouth referrals. In the end, if you’re growing a product, you want “good churn.” It means a product resonates and works, and there’s a major demand in the market.

When users leave happy, no longer needing what a company is selling, it means you have a loyal fan base. Either they’ll come back for more down the line, or they’ll be a living and breathing billboard recommending the product to their friends.

But what about “bad churn”? Bad churn is just plain bad, right?

When building a product, the fact remains that you can’t satisfy everyone all the time. But here’s the silver linings playbook: just as you learn about what you don’t want in a relationship after one combusts (I’m thinking about the dumpster fire breakups we’ve all had), so too can product teams learn what factors lead to churn, and alternatively, what paths lead to ideal relationships and outcomes with their users.

When a product team unpacks the qualitative and quantitative data behind “bad churn”, they can uncover a rich story about how people use a product and the areas for growth.

“One way to uncover these areas for growth is to observe the behaviors of reactivated users,” said Devin. “Often times, returning users are power users.”

“We describe power users as the people reaching a certain outcome which is getting offline and off the app. Likely, power users have been through the process before so they’re higher up on the learning curve.”

After an extended period of inactivity (aka dating someone), Devin explains how (newly single) users return with what looks like an optimal strategy. They do everything “right”, right away. They rate all the potential matches in their batches. They initiate more conversations once they are matched with someone. These power users know that dates only come from those who actually make a move. Learning from power users and how they behave on an app is great insight on the most effective way to use a tool, and a good way to identify ways the product needs to iterate.

“For any app, you have to create happy paths for your users and build your product so users follow those paths, get the desired outcome they were looking for, and continue to return to your product or service because it delivers,” Devin advises product teams.

“Creating happy paths can get a bit complicated because there are a lot of paths to take in a product,” said Devin.

“Plus, the research gives more context on the circumstances that lead to a couple’s connection, and how that has compared to the average user. With Hinge, we mainly look at what some of these happy paths to the relationship are, and how do we make sure the product allows a user to do that as soon as possible.”

It begins with understanding why some users are successful – and others aren’t. This is where data comes in for the win. Recently, Hinge conducted a study of 1,000 couples who got into a relationship within two months of joining the dating app, to figure out how others could replicate their success.

In the behavioral analysis, Hinge found that 80% of couples had listed both job and education credentials in their profiles, and they also considered this crucial information to know when scouting out their prospects.

For heterosexual couples, successful men who found love on the app were actually 12% pickier with their “swipe rights” while successful women were 20% more open to their potential matches than the average user.

Devin also told us that couples in this study had previously messaged with an average of 16 different people before meeting their significant other. They also kept up with the conversation for about three days, swapping approximately 25 messages on average before exchanging numbers.

What a product team can learn from this study is not just how to online date but also how Hinge figured out the right context for future cohorts. Knowing these averages and the what it takes to meet a match, is what makes these happy paths, happy – and in turn, what makes Hinge’s product effective.

This research gave Hinge both a context and a model to follow to help others become successful, too. It’s not about just getting from Point A to Point B in the Funnel o’ Love but rather, what it takes to actually meet a good match in the end.

But product teams can’t expect all of their users to figure out those “happy paths” for themselves.

Your product needs to be designed so those happy paths are intrinsic to the user experience. It should feel effortless to go through a conversion funnel and get the goods.

User experience provides the guardrails for a product’s happy paths. In fact, a good product shouldn’t give your users a choice to follow anything else.

“We’ve changed the product to help optimize user outcome,” Devin told us. “According to the data, conversations typically start within 24 hours of a match. This has led us to implement the 14-day window around when you can connect with someone. Not only does this encourage people to say hello because of the expiration date, but successful conversations that end in a phone number exchange generally happen within two weeks, and quite often just within seven days.

When looking into my own profile (“For research!” I told my boyfriend), I noticed functionality constraints that were to my benefit. The home bar is simple, guiding a user’s actions through limited choices. Once my daily batch runs out, I have one of two choices: I can continue to swipe on preferences or things I’ve done (giving more data to Hinge’s recommendation engine), or I can dive into my messages and actually connect with another human being.

The UX’s simplicity is part of Hinge’s brilliance. Clear happy paths can be highways to success if a user follows through.

You learn a little more from each relationship. And in the case with Hinge, the app learns a little more with each user about what makes a successful connection. Hinge continually adds more context to the complex question lonely hearts seek to answer: what makes a good match?

“We try to make a product that’s not a game or a science experiment. We’re trying to reflect what happens in real life and dating, and to make that experience more seamless,” said Devin. “When users consider people honestly, give the app feedback with a rating or give us more information about their preferences, they’ll see the results.”

In other words, when a user invests in Hinge, the app invests back in the user. In the same vein, when product teams invest in understanding the broader context of complicated metrics, like churn, engagement and retention, they too will see a reward. As the adage goes, “What gets measured gets managed.”

Hinge’s story around data, and how they scored big in understanding churn, doesn’t just apply to dating apps but to many kinds of products. In fleshing out the complete journey of a user with quantitative and qualitative data, products delve into richer insights, bigger opportunities, and begin to see growth rise.

Ultimately, prioritizing the success of users is a strategy that will always be to a company’s benefit. Whether users commit to an LTR with your product or not, your product has to at least give them the opportunity to reap the benefits. In order to do so, the first step is to know your data and commit to measuring your metrics in the context of everything else.

What looks like rejection or failure in a report can, in fact, be the starting point for gaining a richer and more complex understanding of your users. Just as in healthy relationships, the strongest products are built on a foundation of deep understanding of their beloved users. Churn and engagement are complex, but it’s imperative to read those signals in the right context to catalyze growth and build product users love.