Hacking dating websites looking for love
Despite a long history of being blamed for coming between couples, more recently computers and online dating sites are proving to be the leading way to meet people, especially if you know how to get the best out of them. Hannah Critchlow explores further...
Hannah - Online dating. Have you tried it? if so, you're not alone. According to the data crunching website the Statistic Brain, of the 50 million Americans currently single, a whopping 75% of these have gone online looking for love. It's a big buck business with 1.25 billion-dollar yearly turnover. Why? Well, it meets with some success. 17% of American marriages in the last year report having originated from online love. Back over this side of the pond, myself and friends also have given it a go. Have we met with success?...largely, no.
Well first up, I wanted to find out exactly how these online dating algorithms are meant to work...
Matea - My name is Matea Yamnik and I'm a senior lecturer at the computer lab at the University of Cambridge. So, online dating algorithms gather information about individuals, about what those individuals like and don't like, and what they would expect from potential candidate matches. They assign weights to these characteristics and then they put all of that information into big equation and fit in those numbers into a big formula and come out with a percentage which is a potential match.
Hannah - Okay, so how to maximise the chances of online love. I spoke with data analysts and mathematicians who've used their tricks of the trade to do just that. Amy Webb works with data to help predict business's future and she describes how she tackled online dating.
Amy - At the end of one particularly horrible date, I went home and I was really upset and I called my sister. She had mentioned to me the beginning of Mary Poppins, the two kids going through nannies over and over again, and they can't find one. And so, they make a list of everything they could possibly want a nanny. And then miraculously, Mary Poppins shows up. And so, my sister said to me, "Why don't you make a list of every single possible thing you could want and then maybe Mr. Magical wonderful person will appear for you." So, I sat down and I started making this list and ultimately, I wound up a 72-different data points. I needed to figure out a way to prioritise that so I assigned weighted values to each one. And so, I had a top tier and a secondary tier and created a formula. But basically, what I was trying to do was to quantify, these are the things that are the most important and unless somebody meets a minimum number of points then it's just going to be wasting my time. so, I'd come up with those crazy formula and I sort of figured out what it was that I was looking for. It took about 3 weeks but at long last, I found a profile that met the minimum number of points and we started chatting and I didn't go out with him right away. So, we decided we should take a couple of weeks and I wanted to really make sure I was scoring the criteria correctly. When we met in person, that wound up being the last first date I ever went on and that's the guy that I married.
Hannah - Next, I spoke with Chris McKinlay. He joined a dating site whilst finishing his math PhD. These two projects kind of merged. Dissatisfied with how the dating site worked, Chris wrote code to extract data from over 20,000 female profiles and he applied algorithms to the data to figure out what was popular and what would match well with the ladies that he likes. And from that data, Chris was able to create his very own super profile.
Chris - So, it become the most popular male profile in Los Angeles. Before I did this, I maybe had 50 or so people that I matched at 90% or higher and after that, the number just in Los Angeles is about 30,000. I started going out on lots of dates and I kept them pretty short - maybe 20 and 25-minute coffee dates because these were dates with people who'd answered several hundred multiple choice questions in a statistically significantly similar fashion. I find the other similarities would creep out. Like at one point, 8 out of 9 people in a row were the exact same specific kind of funny coffee drink, ((inaudible))(32:12). People would ask the same questions in the same order, kind of have the same colour like iPhone case, similar affectations. There is definitely a groundhog day feel to the whole thing. It got a little bit alienating for me for awhile. I would go on these dates and a lot of them would be really similar but for them, oftentimes, I'd be the highest match on the website that they'd ever seen. Sometimes after several years of being on the site, they were generally really interested in meeting and invested in like having a really good first date experience. I got very good at them because they were very similar to one another. Went to like lean in and show interest, went to like nod. There was very little at stake for me because I usually had 7 or 8 dates setup after.
Hannah - But it ended well, I believe.
Chris - Yeah, it ended really well. I met someone who was just like completely, blew me away and that happened on about the 90th date when I met my fiancée Kristine.
Hannah - What do you think are the chances are of you meeting your future wife, your fiancée, had you not done all of this reverse engineering?
Chris - This particular woman, zero. I think that for any person, there's probably a couple other people on the planet that they would be really, really happy with for the rest of their lives and I'm really lucky to have met one of them.
Hannah - Back to Matea Yamnik on how analysis of the matchmaking skills of online dating sites.
Matea - When people have done statistics about how successful they are, whether they lead to true love and long term relationship. The criticism is that often there are no more reliable than just randomly picking people. What they do provide is a mechanism of meeting people which nowadays, busy lives, it's becoming increasingly more difficult.
Kat - Matea Yamnik and before her, successful online daters Chris McKinlay and Amy Webb.