It’s 2021, Why Are Dating App Algorithms Still so Bad?
Posted on July 1, 2021
WIRED – July 1 – There's been a post-pandemic boom in app downloads, but success rates remain depressingly low. Lockdown was boom time for dating apps. Hinge had a 63% spike in downloads since 2019 and tripled revenue in 2020. May 2021 saw 6.5M+ download Tinder. Tinder and Grindr rank in the top 10 of apps most likely to leave users feeling sad. Most dating apps rely on a combination of stated preferences and collaborative filtering i.e. patterns in who users have said yes or no to. A 2016 study of match rates on Tinder (Queen Mary University) revealed low rates of reciprocal interest: the match rate was 10.5% for women and 0.6% for men. "Machines can only work with what you give them," says Samantha Joel of Western University in Canada, who researches romantic decisions. These complex preferences aren't transparent: people might be embarrassed by something they're drawn to, or even entirely unaware that the preference is there, even as it shapes their behaviour. Compatibility and desirability can be poles apart. We don't always know what we like until it's put in front of us.
Mark Brooks: The science of compatibility profiling and matchmaking is entirely, still, a virgin science. Until we open up our databases to academics and data scientists to do proper experimentation, we will continue to clutch at straws when it comes to really provide more than marginal improvements to long term matchmaking. However, there's a LOT of value in just getting showstopper questions out of the way right upfront. i.e. sexuality preferences, marriage status, marriage-minded?, want kids?, got kids? etc.
