Inside Match.com

It’s all about the algorithm.

What makes a good match?

Entering the offices of Match.com is a bit like strutting into a disco. Coloured lights flash from the ceilings, workers lounge on circular banquettes, dance music plays from hidden speakers. Despite being in a mid-rise office tower overlooking a turnpike in the dry, landlocked city of Dallas, Texas, the Match offices are evocative of a racier environment, where anything might happen.

On a hazy Monday in June, I came to meet Mandy Ginsberg, the president of Match.com US, the world’s largest online dating site. Petite, preppy and freckled, with long brown hair, Ginsberg was wearing sandals, tight black jeans and a loose blouse. Her jewellery was limited to a diamond bracelet and a wedding band. Confident and casual, she seemed as good a person as any to be the face of online dating. We sat in a conference room ­overlooking a floor full of computer engineers gazing at their monitors, and with a PowerPoint presentation, she endeavoured to show me how Match uses cutting-edge ­technology to cultivate age-old emotions.

With the number of paying subscribers using Match approaching 1.8 million, the ­company has had to develop ever more ­sophisticated programs to manage, sort and pair the world’s singles. Central to this effort has been the development, over the past two years, of an improved matchmaking algorithm. “We had to get more ­intelligent,” Ginsberg says. “If you say you want a guy between 30 and 35 in New York who has a master’s degree, you’re going to get thousands of matches.”

Codenamed “Synapse”, the Match algorithm uses a variety of factors to suggest possible mates. While taking into account a user’s stated ­preferences, such as desired age range, hair colour and body type, it also learns from their actions on the site. So, if a woman says she doesn’t want to date anyone older than 26, but often looks at ­profiles of thirty-somethings, Match will know she is in fact open to meeting older men. Synapse also uses “triangulation”. That is, the algorithm looks at the behaviour of similar users and factors in that ­information, too.

Until Ginsberg joined IAC, which owns Match, in 2006, she worked at i2 Technologies, a supply-chain management company, also based in Dallas. She was promoted to her current post earlier this year, after former Match president Gregg Blatt was made chief executive officer of IAC. Besides having the right résumé for the job, Ginsberg had enough experience in love to know that ­finding the right partner is tough.

After a divorce shortly out of college, she tried JDate, a site for Jewish singles, but kept coming up short. Then, while still at i2, she became involved with an engineer at the company who was born halfway across the world. They soon married. “If I had laid out a criteria for what I was looking for, it would not have been a guy from south India,” she told me. “People are complex. You’re constantly making trade-offs about who’s too tall, too short, too smart and too dumb. People come in and tell us a bit about what they’re looking for. But what you say and what you do can be different.”

Academics call this “dissonance”. “It’s a theme that runs through social psychological literature,” says Andrew Fiore, a visiting assistant professor at Michigan State University, who works on ­computer-mediated communication. “We don’t know ourselves very well on a descriptive level.”

The same is true for the millions of Match users, says Ginsberg, and she tried to incorporate dissonance into the algorithm. “I might come in and say I’m looking for a nice Catholic guy between 30 and 40 who is non-married,” she says. “But after weeks of looking at people, I might get an e-mail from a guy who has kids, and I might accept that. It’s all about behaviour modelling. All that data goes into algorithms and affects who we put in front of you.”

To sort expressed ideals from actual desires, Ginsberg realised she would need some technical help. After becoming executive vice-president and general manager of Match’s North American operations in 2008, Ginsberg initially looked to her old employer, i2, for assistance. “I brought over a bunch of people who I thought could help solve one of the most difficult problems out there, which is how to model human attraction,” she says.

A key recruit was Amarnath Thombre, a soft-spoken engineer from Pune, India. Thombre had attended the prestigious Indian Institute of Technology Bombay, then taken an advanced degree in chemical engineering at the University of Arizona. Like his boss, he met the love of his life offline. His wife is also Indian, and they were introduced through family.

Yet Thombre says his experience at i2, where he spent years finding ways to move products around the country more efficiently, was perfect preparation for the online dating industry. And once at Match, he, Ginsberg and a team of nine maths ­whizzes hired by Thombre, set about updating the Match algorithm. “The one thing I knew was numbers and analytics, so we started building a numbers team here,” he told me. “It’s just supply and demand. The same principles work, no matter what kind of numerical problem you’re solving.”

The way the Match algorithm learns, he says, is similar to the way the human brain learns. “When you give it stimuli, it forms neural pathways,” he says. “If you stop liking something, those shut off. It’s learning as you go.” The same principles are powering the recommendation engines at popular sites around the web. Amazon uses similar ­technology to recommend new products for people to buy, Pandora learns from likes and dislikes to customise its internet radio stations, and Netflix famously offered $1m to anyone who could improve the effectiveness of its algorithm by 10 per cent.

“With Netflix, people are constantly rating movies,” Thombre told me. “But there’s only one The Godfather, and you rate it once.” Predicting preferences in human beings is altogether more complicated. “Even if you like The Godfather, The Godfather doesn’t have to like you back,” he said. “The whole problem of mutual matching makes the problem 10 times more complicated.”

It is a subtle shift, but one with profound implications. “Before, matches were based on the criteria you set. You meet her criteria, and she meets yours, so you’re a good match,” Thombre explained. “But when we researched the data the whole idea of dissonance came into focus. People were doing something very different from the things they said they wanted on their profile.”

As a result, Match began “weighting” variables differently, according to how users behaved. For example, if conservative users were actually looking at profiles of liberals, the algorithm would learn from that and recommend more liberal users to them. Indeed, says Thombre, “the politics one is quite interesting. Conservatives are far more open to reaching out to someone with a different point of view than a liberal is.” That is, when it comes to looking for love, conservatives are more open-minded than liberals.

With a mountain of data in its servers from the 75 million users it has had since it was founded, Match has been able to uncover a series of curious trends. Some findings are obvious. Women are less likely to e-mail with men who live far away, men who are older than they are, and men who are short. Other findings are more nuanced. Catholic women are especially unlikely to e-mail a Hindu or atheist male. While men are most particular about hair colour, a woman’s income is less important to them. “We are so focused on behaviour rather than stated preferences because we find people break from their stated preferences so often,” Thombre says.

The Match algorithm is constantly at work behind the scenes, scouring terabytes of data and working to find possible matches. Likely candidates are suggested when users ask to see “more like this” and are also put forward through the “Daily5”, a selection of profiles e-mailed to users each day.

But it is not enough for Match simply to suggest dates without gauging the effectiveness of its efforts. When each Daily5 profile is viewed, the user has to “rate” that profile before he or she can see the next one. The site asks users if they are interested in the suggested match, and gets a reply of “Yes”, “No” or “Maybe”. Each answer is recorded and logged with the user’s profile, becoming another data point for the algorithm to work with.

It’s not known how many dates the algorithm has resulted in. Match can’t know what happens offline. Yet it is clear that changes to the algorithm orchestrated by Ginsberg and Thombre have had an impact on Match users’ engagement with the site. Since the introduction of the improved ­algorithm, the “Yes” ratings on the Daily5 have increased over 100 per cent. More than half the e-mails sent on Match now originate from ­recommended matches (chiefly Daily5). And during the past year users have logged more than 416 million Daily5 ratings. Says Ginsberg: “If we can get more people going out on dates, it will have a profound impact on our success rate.”

On the evening of April 5, 2010, Jonathan Cambry, a muscular ­professional pianist living in Chicago, switched on his ­computer and logged on to Match.com. Cambry, then 28, had joined Match a few months earlier. He had recently ­separated from a girlfriend, but had grown weary of spending time and money trying to find dates at local bars. “That wasn’t something I was interested in, and it gets expensive,” Cambry explains.

So, for about $20 a month, Cambry maintained an account on Match under the alias “Wrigley-Pianist”, where he could browse the online profiles of thousands of women in the Chicago area, and communicate with them through e-mail and instant message. And on this evening, as an ­unseasonably late hailstorm kept most Chicagoans indoors, he was notified by Match that a user named “soubrette1988” was interested in him, having seen his profile in her Daily5.

Viewing the profile, Cambry, who is black, saw a pretty young white woman who lived nearby and seemed to share his interest in music. He sent her a short e-mail to say hello, and within a day received an e-mail from Karrah O’Daniel, an opera singer. Their first date was a flop, but they made it to a second date, and soon Cambry and O’Daniel were getting serious. It turns out they had attended the same music school, but never met there. They took to playing pieces by Franz Liszt together, recording videos that they would post on YouTube. Six months later, he proposed. The two are to wed on October 1 at a church in Minnesota.

As often happens in love, the woman Cambry fell for is not exactly the woman he thought he wanted. “I wasn’t expecting that the person I was going to marry would be a white woman from Inver Grove Heights, Minnesota,” he says. Nor did Cambry fit neatly into O’Daniel’s idea of who she would marry. According to her Match profile, she was looking for a man between the ages of 21 and 26. Cambry, by O’Daniel’s own standards, was too old for her. In fact, Cambry and O’Daniel never really searched for one another at all. They were introduced by the algorithm.

Cambry was included in O’Daniel’s Daily5 because he was similar to 11 other men that she had already indicated she was interested in with a “Yes” rating, Thombre told me. So even though he was older than O’Daniel wanted, and described himself as “stocky”, while O’Daniel wanted a man with a body type that was “about average” or “athletic and toned”, Match correctly assumed she might be interested in him. “We didn’t match, but you can’t really sum up a person in a check box,” says O’Daniel. “Women change their hair colour every month.”

Online dating has come of age. Once a seedy corner of the internet, digital romance is today nearly as commonplace as e-commerce. Of the 87 million singles in the US, nearly half of them, or 40 million, have tried online dating, according to the US Census. Some surveys estimate that one in five new relationships, and one in six new ­marriages, begins online. “This is one of those businesses where scale really ­matters,” says Ginsberg, noting that Match has facilitated 1.2 billion e-mails since 2005, and 110 million virtual winks [a way for members to “break the ice” before e-mailing] in the past six months. She also says there is no reason to expect growth to stall any time soon, as online dating becomes more mainstream and new singles of all ages come online. “With the divorce rate in this country being 50 per cent, we’re a ­reflection of that society,” she says.

Internet dating has also matured into a robust business. Match is owned by IAC, the digital media conglomerate. Last year it and IAC’s other online dating sites generated $401m, or nearly a quarter of all revenues for the group. Paying users were up by 30 per cent last year, to 1.78 million. Most revenues come from subscriptions, with some extra cash coming from advertising. Diet ads, such as WeightWatchers for Men, are popular on the site.

Online dating is also an international phenomenon. Native language sites flourish in countries around the globe. Match UK is successful, though operated independently. And while Match is the best-known and largest online dating site on the web, it has plenty of competition. Chemistry.com, another IAC property, is growing. eHarmony, a rival, has proved popular with an older crowd looking for serious relationships. Niche sites cater to specific nationalities and religions, such as Shaadi for Indians and JDate for the Jewish crowd. OkCupid, launched in 2004, is free to use and has caught on with the young, ­hipster subset. Its success led IAC to purchase it for $50m earlier this year.

Pressure from new competitors has made Ginsberg and Thombre’s work all the more critical for their company’s success. “Match’s fundamental offering is more vulnerable today than at any previous point,” says Frances Haugen, a software engineer at Google who studied Match while at Harvard Business School. “With the advent of improved recommendation algorithms and the implicit compatibility information encoded in social networks, Match must act now or put itself at risk for disruption in the next five years.”

So far, Match has not been knocked askance by the advent of social networking. In fact, advertising on Facebook has become a great recruiting tool for Match. “Facebook is about people you know, and Match is about connecting to people you don’t know,” says Ginsberg. And while there is indeed more competition than ever before, 16 years after Match.com was founded as one of the original online dating sites, it is still positioned as the industry’s frontrunner.

Match.com was founded in 1995 by Gary Kremen, an entrepreneur who saw the potential of the web early on. Single at the time, Kremen was using 1-900 number dating hotlines he found in the classified sections of newspapers to meet women. “I noticed I was paying a lot of money for those numbers, and those big bills got me thinking that maybe people would pay the same online,” he says.

Kremen was right. He founded a company called Electric Classifieds in 1993, and two years later unveiled Match. It was one of the first sites to use the internet to facilitate dates, and among the first to charge money for a service. “That was the original idea, to do classified ads but make it electric,” said Kremen. “I always knew a lot of women; I’ve done a lot of dating in my life.”

Kremen says he designed the site with women in mind. “You have to design the whole system for women, not men,” he said. “Who cares what men think? So things like security and anonymity were important. And little things, like talking about body types, not pounds. Never ask a woman her weight.”

Yet Kremen was faced with an early problem. In 1995 most people weren’t online, and those that were weren’t finding dates there. So Kremen got everyone he knew to sign up for Match. He had all Match employees create profiles, and even though he was in a relationship, he signed up and had his girlfriend sign up, too. There was early success. A critical mass started using the site and online dating in the internet era was born. But it backfired in one important respect. Kremen’s own girlfriend met another man through Match and left him. It was a painful lesson, but at least he knew the site worked.

Kremen and his board had a falling out soon after Match got going, and the company was sold in 1998 for $7m to Cendant. A year later, Cendant sold it to IAC for $50m. Since then, IAC has grown Match to be the most dominant dating site on the web.

Not all digital romance is as wholesome and picture-perfect as the love between Cambry and O’Daniel, however. There is a dark underbelly to online dating that attracts spammers, con artists and those not suited for modern love. A recent lawsuit filed against Match levelled the claim that more than half of its profiles were inactive or fake, an accusation the company denies. Real-life users can be problematic, too. Stories of dates gone awry abound, ranging from the merely awkward to the truly creepy to the tragically abusive. A California woman sued Match after a sex offender she met on the site allegedly raped her. Match has since begun screening new members against the national sex offender registry.

Nor is the online dating experience universally positive. Plenty of users give up on the service after one too many bad dates. “The Match algorithm should have figured out that I don’t want a 45-year-old from New Jersey,” said one frustrated thirty-something professional woman from Manhattan. “Every time I log on I feel faintly insulted.”

However, every corner of the web has its unsavoury aspects. “I don’t know how bad an underbelly it is compared to the rest of the internet,” says Andrew Fiore. “There are good and bad operators in every sector.” And that people are often disappointed with their dates is no surprise to those who have studied the industry. “People tend to like their dates less on average once they’ve met them face to face,” he says. “They tend to like the online version better.”

An online profile, of course, is not an accurate reflection of someone, but a template for them to project their ideal self-image. “Once you meet them in person, it’s harder to have as many optimistic illusions about them,” says Fiore. “We engage in this kind of idealisation when we’re faced with limited information about people online. We fill in the gaps optimistically.” Fiore calls this “an illusion of specificity”. “It’s a way to give someone a sense of control,” he says.

And even Fiore acknowledges that for all the utility online dating provides, reducing potential soulmates to pixelated thumbnails and fields of information can be a draining experience. “It can feel a lot like shopping for a blender on Amazon.com,” he says. “But these are people we’re talking about, not blenders.” Not even the most potent computers in the world, it seems, can engineer a panacea for lonely hearts.

Despite these concerns, it is becoming accepted wisdom that any lingering shame around online dating is gone. Familiarity with the internet, a more casual dating culture and verifiable success stories have all helped. By now, most of us are not far removed from a couple who met online. “There’s a tipping point happening,” says Ginsberg. “There used to be this stigma, or it was ‘good for my friends but not for me’. People don’t realise how pervasive online dating is.”

We don’t know another industry that can change people’s lives so profoundly, except the medical one

And it is an industry that has evolved. Were Match still the site that Kremen founded in 1995, Cambry and O’Daniel would never have met. While plenty of online hookups still happen the old-fashioned way – by searching based on criteria such as location, age and interests – an increasing amount of digital matchmaking is being powered by sophisticated algorithms like the ones Ginsberg and Thombre conjured up.

With their algorithm, Ginsberg and Thombre have taken the allure of online dating and amplified it. Instead of simply creating a digital disco where it is easy to find lots of potential dates, they have put forward a tantalising promise. By evaluating your stated preferences, mapping your site behaviour and using triangulation, Match.com will get to know you, and what you want, better than you know yourself.

It’s not a promise Match can keep with all of its users. But for some, like Cambry and O’Daniel, it can prove transformative. “We don’t know another industry that can change people’s lives so profoundly, except maybe the medical industry,” says Ginsberg. “We often deal with the maths and the statistics, and we have to keep reminding ourselves that this is about helping people find love. There’s not that many businesses that can say that.”

This article originally appeared in Financial Times. Click here to read more coverage from the Weekend FT.