Friday, November 12, 2010

Inferring Cause -- the Case of Social Media

TV execs are budding research methodologists.

They wonder about inferring cause, such as whether social media actually add anything to their programming.  As all you research methodologists know, three factors must be present to infer cause (cause -> effect).  There must be time order (one precedes the other), they must covary (as one goes up or down, the other goes up or down), and that bug bear of all research bug bears, there must be no third variables that explain the effect.  That one always gets me.

Anyway, according to this story, TV execs wonder about whether social media matters, or at least whether the effects are more anecdotal than significant.  Or, as one exec says:
"If we can't actually create a direct correlation between social TV and ratings, we see what it is doing in a lot of other ways,"said Gayle Weiswasser, vice president of social media for the Discovery Channel. "Time will tell what the statistical impact is, but anecdotally it's pretty strong." 
Listening to the social buzz can even be harmful, another exec pointed out. 
He pointed to how CBS brought back a show it had formerly cancelled, called "Jericho," based on the fervent response to its cancellation on the message boards, only to have the show not do well on its second outing.

"That was a case of them listening a little too hard to a rabid minority," he said.

Never saw the show, but there's a good point buried in all the notion you do not want to rely too heavily on the anecdotal nature of social media.  On the flip side, there can be a wisdom of the crowd.  It's a tough business, a tougher research question, on when social media matter or when social media don't matter, or at least can mislead you into thinking tweets equals generalizable public opinion.

If rabid minorities had true influence, Firefly would have stayed on the air.  Unfortunately, it didn't, and it was a damn fine show.

No comments: