The US Presidential Election which concluded earlier this month, marked what many are terming ‘the first social media election.’ Millions of users across the world took to Twitter, Facebook, Instagram and many other social networks to air their views and offer commentary as events unfolded – culminating in a peak of 31 million users on Twitter alone on election night and, as results broke: 327,452 tweets per minute.
And it wasn’t just viewers and voters getting involved. A photograph posted from President Obama’s Twitter account has since reached a record level of viral engagement, becoming the most liked and retweeted post ever. The photograph, which shows the newly re-elected president embracing his wife, with the caption “4 more years,” has been retweeted over half a million times (and counting) and has been liked by over 3.23 million people.
Indeed, this level of social engagement very much speaks for the campaign as a whole. Both Obama and Romney were quick to recognise the importance of social media as they rolled out their respective campaigns in the lead up to the election, reaching an entire demographic via sites including Twitter – which 4 years ago, in the last election, was little more than a fledging community.
While it’s clear, however, that the conversations taking place online played a colossal part in setting the scene for the election, how accurately did these conversations gauge overall sentiment? Is it possible to analyse posts, in all their millions, to see if they could have predicted the end result? Or would an in-depth analysis reveal an altogether more random collection of posts, dominated by irrelevant puns and Big Bird memes?
It’s an interesting question, and one posed by this article on Mashable. It takes a closer look at Twitter activity in the lead-up to the election revealing that, although the breakdown of positive vs. negative tweets indicates a largely equal split between the two candidates, just prior to Election Day there was a comprehensive shift in Obama’s favour.
On a broader scale, social media sentiment analysis is important for PRs and communications professionals to take into consideration. While there are a number of tools available to monitor and report on social activity in terms of quantity – tweets, retweets or likes, for instance – there remain relatively few which are able to provide a more qualitative insight.
Frankly, this is unsurprising, given the vast amounts of social data being generated (each minute, 100,000 tweets are sent, Facebook users share 684,478 pieces of content and YouTube users upload 48 hours of new video, to give just a few examples)! Sentiment analysis itself is a complex science – tasked with grappling with thousands of key words and terms, not to mention a vast range of linguistic curve balls such as sarcasm. And as the amount of data only continues to proliferate, the task isn’t getting any easier.
Luckily however, the technology involved is getting smarter! Within just the last few months we’ve seen EDF light up the London Eye according to Twitter users’ sentiment around the London 2012 Olympics, while one particular tool has recently been launched which uses information from social media to drive investment decisions.
So what does the future hold in store? For PRs, the rise and rise of sentiment analysis will mean KPIs and deliverables are realigned and reassessed. As we begin to see social media ‘reach’ come into play when it comes to campaign planning, strategizing and reporting, the ability to incorporate automatic sentiment analysis across large volumes of posts will become more and more important. The worlds of ‘traditional’ and ‘digital’ communications continue to merge ever-closer – as so clearly demonstrated by the US Election – and the ability to analyse, and understand, these online conversations is surely the next step.
And in four years time, as the next US election comes around? Well, who knows…?
Receive thought pieces from our leadership team, views on the news, tool of the month and light relief for comms folk