We’ve all heard about Facebook‘s ill-conceived “emotional experiment” and OKCupid‘s even better follow-up. While Facebook’s research only concerned slight tweaks in the algorithm that determines which stories show up in users’ news feeds, OKCupid experimented on total strangers who would later meet each other and go on what we call “dates.”
We’re interested in the story primarily because Facebook’s response was simply a blog post that didn’t serve as a very effective piece of self-defense. OKCupid co-founder Christian Rudder, on the other hand, has gone all out to defend his company’s practices as the kind of thing we deal with every day as connected individuals — whether we know it or not.
Last week, to follow up on his “yes, we experimented on people, now get over it” blog post, he gave an interview to TLDR, a podcast associated with the excellent NPR show On the Media (which we encountered via the also-excellent Press Think blog).
The fourteen-minute segment is well worth a listen–especially for anyone with clients in social media.
Some key quotes and takeaways after the jump in case you can’t listen or don’t have time.
The point of the whole thing was to determine whether the site’s definitive “matching” algorithm is really worth a damn: would people who data called a great match actually get along with each other in the real world?
To the OKCupid users who trusted the site to deliver matches, Rudder says:
“I understand…how it might seem that we’ve violated that trust.”
But, he argues, conducting experiments to ensure the validity of the underlying algorithm is actually a case of upholding that trust rather than violating it. Also:
“For the 500 or so people who sent messages [and were all notified afterwards], the worst possible outcome was a stale conversation.”
Was it really? An NPR staffer eventually interrupts the conversation to ask Rudder: are you a scientist or a salesman? Pick one. Rudder defends himself by calling attention back to the practice that brought us to this point in the first place: advertising.
“Oil of Olay’s advertising is there to make people feel insecure…to make women feel old…why is that kind of emotional manipulation OK?”
The difference is that the relationship between the audience and advertiser — unhealthy as it may be — is established in that case. The OKCupid and Facebook experiments, on the other hand, upend that supposed relationship between user and provider. As the staffer puts it:
“I’m not [as comfortable] being a guinea pig as you are making me a guinea pig.”
Rudder responds that users should understand that this is something that just comes with using the service and that it will work better for everyone because of experiments like these:
“In 20 years, everyone will be like ‘yeah, they’re just running an experiment.’”
Maybe. But in the present day, we don’t think OKCupid’s damage control strategy is really any more effective than Facebook’s.