Online daters navigate algorithms that drive apps

Some choose to take matters into their own hands

Lilah Jones, left, and her fiancee Dani Martinez met through Match in 2016. Jones ventured beyond the feed of potential dates, played around with the app's filters and searched Match profiles on her own to find Martinez. (Michelle Kanaar/for the Chicago Tribune)
Lilah Jones, left, and her fiancee Dani Martinez met through Match in 2016. Jones ventured beyond the feed of potential dates, played around with the app's filters and searched Match profiles on her own to find Martinez. (Michelle Kanaar/for the Chicago Tribune)

CHICAGO-A front-row seat in a crash course on app-based dating was the perfect place for JoAnn Thissen.

Online dating takes a lot of nerve, and the 68-year-old retired marine geologist was working up her courage. She's dabbled on dating websites and apps, and even asked for a subscription to dating site Match for Christmas. She hasn't had any luck yet, but she's still determined.

That's why she was there, sitting in a hotel among dozens of other attendees interested in crafting the perfect online dating profile. There were men and women, millennials and baby boomers, singles and people in relationships.

The holidays are peak dating season, and the love lives of tens of thousands hinge on how algorithms behind popular dating apps like Tinder, Hinge and Match piece together their data. Even a decade ago, 1 in 3 marriages started online, one study suggested, and dependence on dating apps has increased. Some users fret over creating the perfect profile to rope in the ideal mate. Others work to outsmart the algorithms behind the services they use.

"There's a lot (about) meeting another person that can't be determined by an algorithm," Thissen said. "They take your info and they crunch the numbers and they come up with something. How do you get them to uncrunch the numbers?"

That's where Bela Gandhi and Smart Dating Academy come in. The date-coaching company, which Gandhi founded in 2009, hosted the dating-app workshop Thissen attended this fall.

The changing nature of the dating scene has caused Smart Dating Academy to change how it teaches people to approach online dating.

The increasingly digital world has changed expectations, Gandhi said. In the past, she made sure clients' hopes weren't built around Hollywood romances. Now she must preach that online dating isn't quite the same as online shopping.

"Our brains are wired," Gandhi said. "It's like, 'I sent an email to this guy, I want him personally to arrive to my doorstep with a dozen roses tonight.' It's like an Amazon Prime mentality to mate search."

Flitting attention spans make app dating a delicate dance, Gandhi told the crowd at her crash course.

You have "about 3 milliseconds" to make a first impression online, Gandhi said. No pressure.

One attendee, Kelli Murphy, 35, said she has noticed how quickly people lose interest in potential matches. She's not expecting instantaneous results-she's been using dating apps long enough to know that's not realistic-but she has crafted her approach based on other users' actions.

"It's best to plan a date within a couple of days or else people will forget about you," Murphy said.

Still, Gandhi loves dating sites and apps. More than one-third of marriages between 2005 and 2012 started online, according to a University of Chicago study commissioned by online dating site eHarmony. Gandhi said that will only increase.

Almost half of Americans are single. Prospects are good for digital daters, especially this time of year.

Dating season peaks between Dec. 26 and Valentine's Day, according to data from online dating company Match Group, which owns Match, Tinder and OkCupid, among others. More than 60 million messages are sent on the Match app during that time, and more than 750,000 dates occur.

Match named the first Sunday of the new year "Dating Sunday" and predicts there will be a 69 percent increase in new singles coming to the app. People resolve to find love in the new year, Gandhi said.

Meanwhile, all those people clicking and swiping in search of a potential partner are good for the bottom line.

For example, Tinder's third-quarter revenues were double what they were the previous year, according to parent company Match's most recent earnings report. That increase was driven in part by Tinder Gold, a premium service to which 60 percent of Tinder's 4.1 million users subscribe. Match also bought a 51 percent stake in Hinge earlier this year. Facebook is looking to cash in too, rolling out a dating service in some countries.

But there certainly is an underbelly to the technology, Gandhi said. For better or worse, people expect to be able to plug exactly who they want into an algorithm and have that person in no time.

"The problem is, people think they know what they want, but they don't know what they actually need," Gandhi said.

In the beginning, online dating was not built on algorithms. Match got its start in 1995 with online personal ads. Singles searched through the site's active profiles to find matches.

Then came the matchmaking era in the 2000s. Psychologists and self-help gurus got behind big online dating services. "Dr. Phil" McGraw dished out dating advice through Match.com, and psychologist Neil Clark Warren founded eHarmony, where users answered a list of questions in search of a soul mate.

"The idea was: 'You don't know what you want; you have no idea. You're going to marry the wrong person. Let us solve that for you,' " said Sam Yagan, the Chicago-based co-founder of OkCupid. "Thus begins algorithmic dating."

OkCupid used data differently when it launched in 2004, Yagan said. Its approach was less about narrowing it down to one soulmate and more about making sure dates weren't a waste of time.

How the algorithms work can be a mystery to users, and they can change at any time. New York-based Hinge, for example, got its start pairing users with friends of Facebook friends, but last summer it ditched the requirement to log in with Facebook accounts.

Match introduced a rating system for users in 2010 that gathers data on customers that the app's algorithm can learn from, said Dushyant Saraph, vice president of product at Match Group.

"We aren't trying to solve for marriages or predicting who is going to fall in love with whom," Saraph said in an email. "But putting two people in front of each other that will strike up a conversation on the app is something we can clearly measure."

Algorithms learn from users' preferences. They gather data on users and how they interact, and calculate which profiles will appear in feeds or as matches. If a user tends not to engage with people with tattoos, the app may stop showing that person people with tattoos, for example.

That worries some users, like Thissen from the app dating crash course. What if they miss someone special because of how an algorithm processed their data?

Yagan, who is also a Match Group board member, believes people generally know what they want, and apps do a good job satisfying those desires. And if they don't, people will just try a different dating app.

"Ultimately, you have to be happy with the product," he said.

Lilah Jones and Dani Martinez are getting married. They're still deciding on the date-summer 2019 or 2020-but they're sure it'll be in Spain, Martinez's home country.

The two went on their first date in January 2016, and Jones proposed in August after the couple scaled Mount Kilimanjaro in Tanzania. They met through Match, but Jones had to dig a little.

Jones, a Smart Dating Academy client, initially trusted the algorithms to do most of the matchmaking. She works at Google and has seen artificial intelligence and algorithms work well. Then she noticed a trend among the people popping up in her Match feed.

"I tended to be matched with people of the same color, people who were professional just like me, who live in certain areas," the Old Irving Park neighborhood resident said. "It was a lot of kind of the same. The person who I ended up with did not come up in my feed."

She ventured beyond the feed, played around with her filters and searched Match profiles on her own to find Martinez, 37. She noticed his photos first. He looked fun, and his profile was positive. The two have vastly different backgrounds, are of different races and grew up in different countries, but they hit it off, said Jones, 43. They each have a child, and their kids get along well too.

People have come to depend too much on dating apps' algorithms to find love for them, Jones said. Sometimes it takes a little work to find the right match.

"You're expecting this algorithm to find your next husband or find your next boyfriend or find your next hookup," Jones said. "This is life, and you get out of life the effort you put into it."

Upcoming Events