This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Find out more here

FEATURE24 February 2015

Data date


Data analysis has changed the course of marketing, but Christian Rudder uses it to change the course of relationships. By Jane Bainbridge

Here’s a bit of advice: use unusual salutations and avoid references to God if you want to get a response when looking for romance online. While these may sound like arbitrary self-help musings they are, in actual fact, supported by hard data and detailed analysis of more than 500,000 first contacts on an online dating site. Apologies to those who prefer sentiment to science to unlock the secret of true love.

But who would analyse data in such a way, and for an online dating site of all things? Meet Christian Rudder, co-founder and president, and data cruncher extraordinaire for OKCupid and author of the book Dataclysm: Who We Are (When We Think No One’s Looking).

A Harvard maths graduate, he – along with his co-founders and fellow mathematicians – turned his maths knowledge into algorithms and insight, to add a touch of analytics to the world of match-making.

It’s proven a lucrative venture. In 2011 bought OKCupid for $50 million. But Rudder’s early ambition when setting up the site was relatively modest.

“We wanted to do something that involved people and how to bring people together was an interesting problem. That was a challenge. And it wasn’t putting people out of business, we weren’t into disruption particularly,” he says.

Formula for Love

Rudder is the embodiment of how a digital world has changed the career prospects for so many people. There was a time, when being a maths graduate would lead to some very specific career paths. But now, as data and analysis has become a bigger part of the world, with even the more traditionally creative industries needing these requirements, mathematicians are called upon in a much wider sphere.

Using algorithms to create matches between people signing up to the site has been central to the success of OKCupid. But Rudder is honest about this always being a work-in-progress.

“Whenever algorithms are used – for any recommendation sites such as Netflix or Facebook for instance – it’s never exactly right. It can always get better. There’s no provable way to recommend so the algorithm is always in flux. What we use now is better than it was 10 years ago and it’ll probably be better in another 10 years. Also we don’t think of a matching algorithm on its own – it’s not just the recommendation – it’s also how we display photos, the visual and verbal cues,” he says.

So I wonder if Rudder has experienced any backlash to his venture, that people didn’t want a maths formula to have a role in falling in love?

“OKCupid has never got any user complaints, people know what we do and we are transparent about the process,” he retorts.

As well as leading OKCupid’s analytics team, Rudder writes the blog OKTrends, sharing research and insights based on statistical observations from OKCupid data. This, along with his book Dataclysm has thrown up some fascinating – and sometimes rather disturbing – insights into online dating behaviour and people’s preferences when looking for a soul mate. 

Helping to order people

For instance, when looking at the age of men a woman finds most attractive, it increases pretty much in line with the age of the woman, give or take a few years either direction. For men, however, no matter whether they are 20 or 50 years old, they will point to women in their early 20s.

To make matches, people signed up to OKCupid are asked questions, with users answering on average about 300. Which questions are most important is in the hands of the responders, not the site. “We wanted to make it so users decide what appeals to them,” says Rudder. As people indicate answers that they consider a good match there is the danger of it leading to a very prescriptive ideal.

However Rudder points to questions such as ‘Do you like scary movies?’ and ‘Have you ever travelled alone to another country?’ as being particularly good at predicting the success of a match. Apparently for about three quarters of long-term couples brought together via OKCupid, both parties answered these two questions in the same way.

So what, in terms of the bigger picture, has OKCupid taught Rudder about love and romance? “I’ve learnt first-hand how judgemental people can be. I’m married now, but when I was single you move from one person to the next, all involving judgements and decision-making. I think back to the people I talked to and now I see it online,” he says.

Real world relationships

But he is keen to point out that once the initial ‘sorting’ is done online, it’s in the real world that relationships are made. “OKCupid is about love and sex and that happens in person, not online,” he says.

“There’s no mathematical secret [to a good relationship], it’s different for each person – you see the dos and don’ts in magazines but there’s no outside source that can tell you how to run your life. What we do is help order people.

“We don’t explicitly use neuroscience – my point is there’s no good or bad match – users choose for themselves, it’s just a ranking. Someone with a 25% match could be perfect for you. It’s a way to compare people so the goal is always that they decide themselves,” he adds.

Rudder has become a champion of data interpolation and the way in which insight can be gleaned from it – and he’s not afraid to take on the difficult topics sometimes associated with that.

So when in June last year the research and marketing community found some of its practices uncomfortably scrutinised following Facebook admitting to altering the tone of people’s news feeds and monitoring the impact on their subsequent posts, Rudder dealt with the issue head-on.

While some of the public were up in arms about this ‘emotional manipulation’ and the research community debated furiously about the ethics of such an exercise, Rudder wrote a post on OKTrends entitled ‘We Experiment on Human Beings’. In it he said “We noticed recently that people didn’t like it when Facebook ‘experimented’ with their news feed. Even the FTC is getting involved. But guess what, everybody: if you use the Internet, you’re the subject of hundreds of experiments at any given time, on every site. That’s how websites work.”

Talking about the controversy now, Rudder says “it sucks that people misunderstood it” but that “one positive upshot is that people are more aware”.

More Choice

“For both Facebook and us – it was so distorted by people motivated, and whose job it is, to generate moral outrage – we were just changing a couple of numbers on a page. Say we notice the algorithm may not be working, we’d make adjustments,” he says.

But experimentation aside, online dating – be it aided by mathematical algorithms or managed in a less data-orientated way – has become as much a part of meeting a potential partner as going to a bar. So will the rise of online dating have much bigger implications for society in the long run? For instance if, as Rudder’s data shows, looks matter more online than in real life and that men go for younger women for example – will online ultimately have real effects on the way people get together?

“It is changing the way people get together – it’s giving them more choice. You could see this as a bad thing and say when there’s more choice, there’s always another option and that online dating exacerbates it. Or you could see it as a positive thing, and say people end up with more solid relationships because they have more choice to select their true match from,” he says.

If it does have an impact, it’ll be beyond Rudder’s lifetime before it becomes apparent. “Only time will tell and human urges develop over millennia,” he says.


2 years ago

I had reviewed over 55 compatibility matching engines intended for serious dating since 2003, when I had discovered "the online dating sound barrier" problem. The Online Dating Industry does not need a 10% improvement, a 50% improvement or a 100% improvement. It does need "a 100 times better improvement" The Online Dating Industry needs a very powerful algorithm like the "Teller Ulam design". In this case 100 times more powerful than actual matching algorithms. With less than USD10 Million you can copycat eHarmony or innovate and revolutionize the Online Dating Industry, killing those old & obsolete sites forever. The key to long-lasting romance is STRICT PERSONALITY SIMILARITY and not "meet other people with similar interests". Breaking "the online dating sound barrier" is to achieve at least: 3 most compatible persons in a 100,000 persons database. 12 most compatible persons in a 1,000,000 persons database. 48 most compatible persons in a 10,000,000 persons database. 100 times better than Compatibility Matching Algorithms used by actual online dating sites! The only way to achieve that is: - using the 16PF5 normative personality test, available in different languages to assess personality of members, or a proprietary test with exactly the same traits of the 16PF5. The ensemble of the 16PF5 is: 10E16, big number as All World Population is nearly 7.0 * 10E9 - expressing compatibility with eight decimals, like The pattern is 92.55033557% +/- 0.00000001% similar to the pattern Using a quantized pattern comparison method (part of pattern recognition by cross-correlation) to calculate similarity between prospective mates.

Like Report

2 years ago

Fernando, you are right on all points, except one. The people in this industry are not really concerned with the accuracy matching other than claiming some successes based on people linking up for a period or the occasional marriage. Their objective is to just get the ball rolling as it were, so that basic "likes matching" is all that is needed for at least the first night on a date, which is why so many people keep trying because at least something was appealing. Its like gambling, the small payouts hook people for the longer run. There seems to be a maths opportunity in that last sentence. You had me intrigued with the Teller-Ulam story!

Like Report