Social listening — monitoring social media for comments, in this case, about your candidate and opponent — is one of the new techniques making its way into political campaigns and lots of screws are being pounded into lots of boards as folks learn how to use the tool.
Such analyses can be quite useful in understanding the perceptions of the online world, but problems come in trying to translate those findings into assessments of the offline world.
(I reject dividing between the online world and the “real world.” The online world is real, it’s just different than the offline world.)
If positive sentiments about Candidate A are greater than those about Candidate B, they presume A is ahead. Or if negative tweets or posts about B jump, they believe B is experiencing difficulty in the offline world as well.
The evidence, however, suggests the online and offline worlds are distinct and importantly disconnected.
One recent study probed online and offline conversations about 500 brands during 2016 along several metrics.
There was a modest correlation between the volume of discussion about brands on- and offline. Beats Electronics, however, generated lots of online discussion, but little offline. The amount of discussion in one venue said little about the volume in the other.
Sentiments, the positive and negative views expressed on- and offline, were positively correlated in less than 1 percent of cases.
The difficulty in translating between the online and offline world is evident in politics.
Facebook reported that Sen. Bernie Sanders (I-Vt.) dominated the conversation among Iowans just prior to the 2016 caucuses, garnering 42 percent of the comments to just 13 percent for Hillary Clinton. The two tied on caucus night. On the GOP side, Sen. Ted Cruz (R-Texas) generated only half the conversation Donald Trump did, yet it was Cruz who prevailed.
Indeed, sometimes social media analyses aren’t even correlated with each other.
After Dick’s Sporting Goods announced it would no longer sell assault rifles, one analysis found 79 percent of Twitter reactions were positive, while another, using a different methodology to listen to a broader range of social media, concluded that the online conversation about the company’s decision was largely negative.
And what happened to the company in the aftermath? Online sales jumped 24 percent, while in-store sales dropped 2.5 percent.
It shouldn’t be surprising that online conversations do not necessarily reflect those offline. Online discussions represent a fraction of the conversations we have.
Moreover, when it comes to politics, the number of people commenting is small.
A Pew survey leading up to the 2016 elections found fewer than 14 percent of Americans posting about politics on Facebook and less than 4 percent doing so on Twitter.
More recent data suggest that about a third of Americans have used social media to take action on issues important to them, which is not quite the same as posting on campaign politics.
But whatever the exact number, a clear minority of Americans take to social media to talk about politics. Almost by definition, those who do are more engaged than the average citizen.
Because social media comments represent a small, nonrandom sample of the public, there is no particular reason to expect that conversations on social media necessarily reflect those taking place offline.
None of this is to say that listening to social media is a worthless exercise. Far from it. Significant numbers of Americans participate in those discussions and an even larger number are exposed to them.
But we need to recognize that online and offline are separate, sometimes intersecting, sometimes diverging worlds. No one should assume what is happening in one space is being reflected in the other.
Mellman is president of The Mellman Group and has helped elect 30 U.S. senators, 12 governors and dozens of House members. Mellman served as pollster to Senate Democratic leaders for over 20 years and as president of the American Association of Political Consultants.