Wait a second, that's not science.
Profile pictures tell all
So, what about the photos you use in your profile? OkCupid wants you to know: They're worth absolutely everything in your profile, and writing anything about yourself is a waste of time. When users were asked to rate potential dates on their personality and looks, more attractive people had higher personality scores. This really does us a favor by revealing how shallow we really are. Rudder makes his point by explaining that a person who displayed an image showing a lot of skin and looking pouty was rated as being in the 99th percentile for personality, meaning that she seemed to have a better personality than 99% of OkCupid users:
The catch here was that the profile contained absolutely no text to describe her personality. As Rudder explains, "[She's] obviously a really cool person to hang out and talk to and clutch driftwood with." This should probably give us all something to think about.
Here is where the experimenting turns rather shady. Granted, OkCupid needs to be able to test how well its matching algorithm works. But what happened in this experiment is that the site changed users' match levels, telling people that they were a 90% match when in reality they were a 30% match, and vice-versa. OkCupid did exactly what Facebook did: It toyed with people's expectations — and emotions. At this point, I'm not sure which experiment is worse.
What OkCupid officials did not reveal — and I doubt that it cared — was how users felt when nothing worked out and their time finding a mate was wasted. According to OKCupid's data, only about 20% of users exchanged more than four messages.
Online dating has become a big industry, worth more than $1.2 billion in 2013. Although OkCupid is largely a free service, it charges for "premium" features, like its A-List service, or the ability to filter results by body type. Despite all the pseudo-science data offered up in OK Cupid's explanation of what it did and why, I'm still wondering whether it's manipulated the results of its users outside of this experiment, and whether paying users were affected. If you're getting augmented results, does OkCupid deserve your money?
If it is impossible to know when the information you are being given has been altered, I would say no. Frankly, paying to get lied to is just a sketchy business practice. And it raises questions about whether this type of testing is happening on other sites owned by OkCupid's parent company, IAC, including Match.com, Chemistry.com, and Tinder. (I reached out to OkCupid for comment. I got no response.)
Sign up for Computerworld eNewsletters.