Today’s post ties together a whole bunch of topics we have talked about recently. It starts with an article from Christian Rudder, one of the founders of OkCupid.
In this article, he crows about the experiments that OkCupid runs and pooh poohs the hubbub that arose after the Facebook reveal last month. Face it, he tells us. We are all the subjects of experiments on the Internet. Get used to it.
But as commenter Brandon says, A/B testing is one thing. Outright lying to your users is something else. And this is “not cool” as commenter “Not Cool” says.
That is how it ties to one previous article, here is how it ties to another. You might remember we talked about the algorithms that OkCupid uses to match people. It turns out that the algorithm may be a bunch of hogwash. And the reason, you guessed it, is priming, which ties us back to a whole bunch of discussions we have had on this site (here and here).
What OkCupid did was to reverse the algorithm. They found pairs that the algorithm suggested would be poor matches and told them that they were exceptionally good matches. Then they sat back and watched. These users viewed the profiles of the poor matches under the guise of good matches and the priming was enough to get them to bite. They were more likely to contact these matches than a control group. And they continued. They were also more likely to communicate back and forth four times (which apparently is their internal metric for a budding relationship) than the control group. In fact, being told that they were exceptional matches (despite being poor matches) had just as strong effect on a relationship developing as actually being exceptional matches but being told that they were weak matches.
Priming with a single word (“exceptional”) was just as powerful and effective as the high falutin’ algorithm they created. So save your money and just talk to yourself in the mirror a la Stuart Small only in reverse.