Social Media Analytics Outperformed the Pollsters

By on

Click here to learn more about author Mark Green.

Social Media Analytics Outperformed the Pollsters in the Presidential Election

By Mark Green, COO of 4C (

The polls have been proven wrong twice in 2016 – first in the Brexit vote, and more recently in the U.S. Presidential Election, the two most closely watched and hotly contested political events this year. Yet, our 4C analysis of social media engagement identified that both the Trump campaigns and the Leave camp were going to win prior to the votes.

Polls are complicated and in unusual elections prone to getting their projection assumptions wrong.

The takeaway is clear. Polls could benefit from new ways of measuring voter sentiment – and social media proves the most comprehensive metric today. While the debate rages on about how to address the disconnect between polling numbers and actual results, social data is demonstrating to be an effective way of measuring behavior. Going forward, it has to become a key component in predicting outcomes.

In the U.S. elections, despite dizzying gyrations in the polling numbers over the course of the election, the polls projected that Hillary Clinton was going to win. But 4C’s analysis of social media told a different story. In the weeks leading up to the election, we tracked significant changes in social media engagement particularly in battleground states that had been predominantly Democrat and turned out Republican in 2016.

This disconnect between the polls and the outcome is even more striking because it’s the second time this year that it has played out so publicly. During Britain’s heated battle over the vote to exit the European Union, 4C’s analysis predicted a #Leave victory although the majority of British polls forecast that the Remain vote would carry the day.

The power of social media is that it captures a “limitless focus group” at a massive scale. Analyzed correctly, that focus group provides novel ways of gathering crucial data and pulling out insights that are increasingly in tune with public opinion. It’s understandable when you consider that more than two-thirds of the U.S. population is now on social media. Twitter, Facebook and other platforms better reflect people’s voices and provide a much larger sample size than any poll possibly can.

Of course, the trick is in the analytics. Our elections research focused on engagement and sentiment, not simply likes and followers. This provides a much clearer view of actual behavior and strips out metrics that can be misinterpreted in much the same way the election polls were.

At 4C, we collect all the social engagement data points across social media platforms, tracking the relationship, or affinity, between these data points to see what different topics have in common. We analyze the relationships between these topics over time and shifts as they occur, tracking the frequency and the intensity of the activity in these topics, or how many people are engaged and talking about them and how intense that engagement is. We also look at positive and negative sentiment. By analyzing this activity on social media 24/7 and monitoring anomalies, you can get a baseline of what the real trend lines are and a robust, relevant read of that trend line. As we did with the U.S. Election and the Brexit vote.

There’s been a lot of post-election discussion about how much Twitter and Facebook are creating echo chambers for the candidates supporters. What isn’t in doubt is that these social media platforms provide people with a forum to contribute unvarnished points of view on live events, in real-time. When you study what they’re saying and doing in aggregate using intelligent analytics, you gain insights that weren’t possible before—and that traditional polls are missing.

We use technologies such as cookies to understand how you use our site and to provide a better user experience. This includes personalizing content, using analytics and improving site operations. We may share your information about your use of our site with third parties in accordance with our Privacy Policy. You can change your cookie settings as described here at any time, but parts of our site may not function correctly without them. By continuing to use our site, you agree that we can save cookies on your device, unless you have disabled cookies.
I Accept