Jan
25

JoePa’s Final Lesson: The Raw Power and Inherent Limitations of Social Media Monitoring Research

Post by Ted Donnelly

Legendary football coach, Joe Paterno, died on Sunday morning.  He was 85.  However, in the social media world, he expired one day earlier.  On Saturday night, a student run news blog had erroneously reported that Coach Paterno had lost his battle with lung cancer and that the football team had been notified via email.  The story gained momentum in the social media stratosphere and ultimately was run by several national news outlets, including CBS Sports.  By late Saturday evening, #Joepa was the top trending topic in the United States on Twitter and #2 in the world as an outpouring of condolences was expressed via tweet.  A number of current Nittany Lion players were included in this mix, which lent further credibility to the story (they later acknowledged getting caught up in the social media buzz and had not received any formal notification).  In the waning hours of the evening, the family refuted the story and his two sons indicated on Twitter (naturally), that their father was still fighting.

It was fitting that social media was responsible for propagating the false report as Joe was known to eschew technology.  He preferred a rotary dial phone because it gave him a chance to think about what he was going to say when the recipient answered the call.  Famously, he never read any of the email that came through his university account and in trading barbs with the media, he once quipped, “What do you guys call those things? Tweedle-doo? Tweedle-dee? I haven’t got the slightest idea what you’re looking at."

It goes without saying; the false report caused his family great pain during the final hours they shared with him.  It also is a cautionary reminder about overreliance on social media monitoring in research.  The proliferation of social media usage has brought great opportunity to mine insight.  And what’s not to like?  What a powerful way to aggregate a previously unmanageable quantity of data.  If anything, the criticism has focused on the overabundance of information and how to make sense of it all.  Additionally, there are concerns that the algorithms used to analyze content can fail in accurately portraying a statement as positive or negative in tone.  However, I would imagine these limitations will only improve over time.

The real lesson is accepting the position social media monitoring holds in the greater context of the research toolkit.  When new methodologies emerge, there is a tendency to speak in superlatives, to predict the downfall of other methods, to see the only way forward as absolute change.  Social media monitoring provides a new way to reveal insight, but it does not replace all other forms of research.  In the case of Paterno, if the objective of the research was to understand public sentiment toward him upon the announcement of his death, it would have been very effective, sampling error aside (and yes, there were some reveling in his demise).  However, it was based upon a false premise that he was, in fact, dead.  If the objective of the research was something of a more tangible nature, such as determining what the current health status was of Coach Paterno, obviously it would have failed miserably.

There is much to be gained by embracing social media monitoring in research as long as it is viewed for what it is: an observational research method.  It does not allow for interaction, hypothesis testing nor the exploration of underlying motives, attitudes and beliefs.  There is no way to direct the discourse to meet your research objectives, nor control when, whom or how many wish to discuss a topic.  As we learned Saturday, it is also subject to uncontrollable error.  As such, it should be positioned as a complementary method to traditional research rather than a replacement.  In the marketing research industry’s unending quest for actionable insight, we need to open our eyes to new methodologies without feeling compelled to close the door on what came before.