Gone are the days when finding your soulmate online was filled with shame — a recent Pew Research Center report shared that the majority of Americans think that online dating is a good way to meet people. To show the transition in perspective, the online dating industry has grown by almost 12 percent in five years, with a revenue of $3 billion in 2018 alone.
And with the mobile revolution, swiping right (or left) has become a common trend in the dating world, as we increasingly trust our romantic life to our smartphones and let algorithms be the matchmakers. But how does it all work?
“As a dating app, you’re trying to take the place of a matchmaker for each individual person. Now we’re using AI and machine learning to help figure out who that compatible match is for the user on your dating app,” says Dig CEO Leigh Isaacson, a dating app for dog enthusiasts and owners.
Dating app users often theorize how these dating algorithms work. Do previous matches matter? Will you be punished for being too picky? Are the most popular profiles really prioritized over others?
Isaacson says that it truly is pretty objective. Your matches come down to machine learning and the information you’re willing to provide the app.
“The app will learn who you’re liking more and who you’re matching with to be able to show you people similar to that. It makes the process of matchmaking even more efficient” she says.
For example, Tinder gives every user an internal desirability rating based on how swipe-able you are. Others use a filtering system to match you with those that have the highest probability of clicking with you, or use the Gale-Shapley algorithm, a mathematics theory from 1962 (applied by dating app Hinge).
“As a dating app, you have to make that decision and the way you do that as you’re building a dating app is having constant communication with your users and just learning what’s most valuable to them,” says Isaacson.
“In the case of Dig, this algorithm is going to look very different than a dating app for cat people.”
For New Orleans-based Dig, this means matching single dog lovers by not only compatibility between the humans, but also their preferred dog lifestyle. The app, available nationwide, shows users five available matches near them each day.
Instead of swiping, users choose from “dig,” “really dig,” or “pass” on their profiles. Once someone digs you back, the app sends you pet-friendly date ideas.
“As a growing dating app, we’re starting from the beginning. The important thing is making sure that there are people for you to see in your area and get you to start clicking for the machine learning algorithm to learn more about you. We do this before we build other aspects on top of the algorithm,” says Isaacson.
“When publications post that you might do better on these dating apps if you like X, that’s because, over time, those apps have learned who their customers are, who their user base is, and what they’re looking for. Recently I read that if you speak Spanish, Zoosk might be the best dating app for you.”
The team at Dig uses both the feedback coming in from users and the observations of what they’re clicking on within the app to see if it matches up and create a better experience. For example, a user may say that they like large dogs, but continue to click and interact with matches that have smaller dogs.
“You want to talk to people directly to make sure there’s no social bias in the algorithm,” says Isaacson. That’s how they came to realize a need to add “non-binary” as a gender option on the app.
“You have to determine what is valuable information in terms of matchmaking; for example, things that you put in your bio may have nothing to do with what you’re looking for. We’re not going to use the school you went to in the matching algorithm, but may use age or distance to other users,” she says. That’s different from The League, which does use school and professional information for matchmaking.
For Dig, canine lifestyle is a big part of the algorithm. Do you let your dog sleep on the bed with you? How long are you comfortable leaving your dog alone at home? Do you take your dog to daycare, which might later spark a budget conversation?
Non-dog owners are welcome on the platform, but must answer questions about their own preferences to make sure they are matched with someone compatible.
“These big picture questions might give you a better idea of what type of dog owner and lifestyle the other person leads. There seems to be a big indicator that people are having these types of conversations and matching better. So we review it and decide to maybe make it more prevalent when signing up,” she says.
Isaacson and the team are still making decisions about their overall approach, such as how much time they want users to actually spend on their app. Dating apps are one of the only technology platforms where the desired successful outcome means the user will likely want to delete the app. With that in mind, how quickly do you aim to find users their best match?
Researchers explored whether or not online dating fulfills its mission to bring people together and found that the answer is yes. They observed that these platforms do expose people to prospective partners that they otherwise wouldn’t have met, with algorithms bringing matches together to provide a clearer sense of romantic potential than in-person first impressions.
“As a dating app, you have to make a decision of ‘do we want to go in with our own biases?’ and create an algorithm of a person that I, as a founder, would be attracted to?” asks Isaacson. “Or to let it be the wild, wild West and learn from your customers and take a little bit more slowly.”