PageRank and Matching Algorithms in Dating Apps
https://mashable.com/article/tinder-bumble-hinge-okcupid-grindr-dating-app-algorithms
https://www.vox.com/2019/2/7/18210998/tinder-algorithm-swiping-tips-dating-app-science
Apps designed to help you find the perfect someone, so that you can quit the app. That’s what most of today’s dating apps advertise themselves as. Tinder, Bumble, Grindr and Hinge are among the most popular, drawing in hundreds of thousands of users. Hinge’s slogan is even “designed to be deleted”. But are these apps really designed to help you find your perfect match? To investigate, let’s take a look at their ranking algorithms.
At least until 2019, Tinder used the Elo rating system to designate which users showed up on your page. Elo was originally designed for ranking chess players. It uses your wins (right swipes on your profile) and losses (left swipes on your profile) to create an Elo score for you. This score effects how often you show up on other people’s pages. Importantly, the popularity of the people who swipe right or left on you also impacts your score. This is a similar, but not identical algorithm to PageRank, the ranking algorithm used by Google. PageRank orders the importance of websites based upon their number of visits, and the number of other websites those visitors visit. The more websites a user visits, the higher their score, and those scores in turn boost the score of the website. In this way, Tinder’s old ranking system is similar to PageRank. The more popular users/sites are shown first, and as you swipe/scroll through subsequent user/sites shown have a lower score.
However, nowadays Tinder allegedly uses a different rating system than Elo, due to having a higher number of users. Now that their collection of user data is of significant size, they are able to curate a list of profiles based off what kind of users are most likely to swipe right on each other. This system of matching becomes more like an auction, in which users have different calculated valuations of other users. These users are then “matched” with others, although the difference is that one user will be shown multiple other profiles, instead of just being shown one profile. Overall, dating apps such as Tinder rely heavily on economic concepts of auctions, rankings, and matchings.