Love has always been big business. Valentine’s Day spending this year was estimated at a record-breaking $19.7 billion, according to the National Retail Federation’s Valentine’s Day Consumer Spending Survey conducted by Prosper Insights and Analytics, almost a billion dollars more than in 2015. And as more and more people are turning to the internet to find their Valentine —a new study from Pew Research estimates that online dating has nearly tripled for singles 18-24, and more than doubled for singles 55-64 since 2013, with 15% of American adults online dating— dating sites have evolved into a $2 billion dollar industry. Beyond finances, online dating sites were calculated to have led to one third of marriages as of 2013, and there is research showing that those marriages are happier than ones that got their start offline.

So how do these love machines work? Analytics are the key to the success of online dating giants like Match.com and OkCupid, which use a bevy of questionnaires and predictive algorithms to match its users with a compatibility percentage. Match.com estimates that it has that it has over 70 terabytes (70,000 gigabytes) of data about its customers, which it uses to constantly fine-tune its predictions.

Christian Rudder, co-founder of OkCupid, recently spoke at Northeastern University about love and analytics, the topics behind his best-selling book Dataclysm. He informed the audience that OkCupid’s algorithm matches users based on four metrics: attractiveness, matchiness (based on responses to personal questions), frequency of responsiveness, and randomness. The last one may seem like an outlier, but it isn’t: just as there is randomness in meeting people in highly-populated real life, there is randomness in whom OkCupid chooses to display to certain users. Rudder’s thesis is that online habits on dating sites reflect offline habits. OkCupid’s data proves this in numerous examples: everyone is interested in the most attractive people, the attractiveness of users is a bell shape (where most users fall somewhere in the middle, and fewer on the extremes), and women are harsher than men when it comes to critiquing love interests.  

Screen Shot 2016-03-01 at 12.37.39 PM
Christian Rudder, co-founder of OkCupid, at the Northeastern speaker series

Some newer companies are taking the analytical approach to another extreme. The online prediction tool Nanaya bills itself as “the first app to scientifically predict your love life.” The brainchild of Rashied Amini, a former systems engineer at NASA, the browser-based webapp uses an algorithm to predict person’s romantic chances using a combination of data, scientific method, and results from a personality test. Their “Romance Report” includes a chart predicting your odds of finding love over the next seven years, Romance Scores that analyze what type of relationship matches your personality, and locations and groups where you are most likely to find your ideal match. Nanaya launched on Valentine’s Day 2016, and already has over 130,000 users.

Algorithms are not the sole answer for something as old fashioned as love, but as online matchmaking expands and evolves, our odds of finding that special someone only increase.