Perhaps the outcome of the 2016 presidential election serves as an excellent example of why we should not lean too heavily on quantitative research. Alternatively, why we should give qualitative more of a role in analytics and decision making. Sure, quantitative research plays an essential role in any researcher's arsenal, but sometimes, maybe even often, we become too reliant on the "math," which can be deceptively comforting. I say this as someone who holds a degree in economics and mathematics and conducts a lot of quantitative studies.
I started out last night's election-watch with Nate Silver's fivethirtyeight.com giving Clinton a 75% chance of winning. The charts, the stats, and the incredibly cool info-graphics made it all seem so certain. FiveThirtyEight prides itself on using "hard numbers and statistics to tell a compelling story." Several hours into the vote count, the probability had completely flipped with 538 now giving Trump a 75% chance of winning. In direct contrast to "predictive analytics," the math was continually readjusting to catch up to the reality - "reactive analytics."
It seems like the Clinton campaign should have done a little more deep listening and a lot less "scientific" polling from a distance. So much for the digital marketing experts and statisticians who relied more on their algorithms and data collection techniques than they did on good old fashioned listening.
The Donald kept saying that the reality he was experiencing was not consistent with the polls. "It's amazing; we're getting tens of thousands of people at my rallies." "I don't need Beyonce." "You should see the crowds." Trump was out there with the people in the thick of it while the Democrats were tweaking algorithms and cocooning in the comfort of their "hard numbers."
So, when clients ask whether we need to do some qualitative research, I'll be using the 2016 presidential election as an excellent example of why it pays for researchers and branders to dive into the thick of humanity and do some deep listening. Counting is not the same as listening and "hard numbers" do not tell the whole story.