Conjoint at the Hip: Understanding what Dating App Users Really Want w/ RBC
A love story of features, preferences, and data
Hi all,
With Halloween close upon us, I thought I’d cover the scary topic of modern dating apps. Back in college I had the pleasure of taking Zhenling Jiang’s Data & Analysis for Marketing Decisions class with Chase Seklar, Alice Heyeh, and Carly Siegel. For our final project, we decided to focus on using ratings based conjoint (RBC) analysis to understand what features were most important to an individual’s decision in what dating apps to use and see how building an app with the most liked features would impact the overall market.
You can find the slide deck I’ve created to walk through our process and results here. You can find (and reuse if you so choose) our analysis and anonymized data here.
If you don’t have time to go through the short deck, here are the rough beats (read the bolded for an even faster tl;dr):
We sent out a survey to 117 individuals in their late teens to 40s to understand how they would rate different baskets of features (i.e. Free and limited likes, $7/mo and unlimited likes, etc.).
We considered 7 dimensions (Price per month, Unlimited likes, Superlikes per week, Like Visibility, Boost Per Month, Personalized Date Recommendations), finding Price Per Month to have the biggest impact on user rating (lower price = better).
We found Match Basic would’ve been the best app for the market (38% market share) as it was the only fully free app with an additional feature that would improve user app experience (Unlimited Likes).
Utilizing our RBC model to understand how users derive value from different features, we simulated the market with a few dating apps and the simulated value of the apps from our model. We calculated market share by taking the individual value of an app and dividing it by the sum of the total value of all dating apps in the market.
We then designed our dating app to have a new feature (Personalized Date Recommendations) in addition to being free and allowing users to see who liked them to make our app seem like a new entry into the market. Our app knocked Match Basic down to second place although both apps would hold large swaths of the market.
If you’ve enjoyed reading this so far, you should subscribe! I post articles like these every few weeks : )
This isn’t representative of the modern market however. Tinder is one of, if not the most, popular dating apps, yet its most popular form (the basic free app) was not most popular in our simulation of market share. One of the core areas I believe we missed was the importance of an existing user base for a platform business (a platform business monetizes its ability to facilitate and grow value from more interactions between multiple parties; think Facebook, Uber, and dating apps as a few).
Coined as the Cold Start Problem by Andrew Chen, the value of platforms comes from having more users to interact with. In the context of dating apps, if there aren’t many users, there’s a very limited pool of people to match with and, hopefully, go on a date with.
Both Tinder (75M MAU in 2022) and Hinge (23M MAU in 2022) have massive user bases which encourage more folks to join. Even in my own experience having tested some smaller and more niche dating apps, it was difficult to have any matches as there were too few people. Leaving this out of analysis left a large hole in our considerations.
Additionally, users can use multiple dating apps at one time (I had installed 10+ for this experiment; Please DO NOT use this many at once! I would not recommend it!). I don’t think a market share calculation was the best way to describe our end result.
A better way would be to think on how dating apps make money (mainly subscriptions and ads) and see how we could use say customer willingness to pay or time spent on app as a way to measure the value of a dating app not just to the end user but to the actual developer of the app as well.
As a representative example on how dating apps make money, this screenshot is from Match Group’s (owner of Tinder, Hinge, and many other dating apps) 10k in Feb 2023. Direct revenue is subscriber revenue and indirect revenue is ads.
Overall, this was a really fun project to work on and I hoped you enjoyed! A huge thank you again to my incredible team and the fantastic Dr. Jiang. It was a joy to learn from you all!
Thanks for reading folks (please subscribe if you enjoyed!) and have a happy Halloween!
Frank