We use cookies on this site to enhance your user experience.

By clicking any link on this page you are giving your consent for us to set cookies.

Yes, I agree

No, Please tell me more



Do not look to Twitter to predict the next government

When the exit poll results were finally revealed at exactly 10pm last Thursday it’s fair to say that even the Conservatives, who were rightly predicted to become the majority party, were a little taken aback.

The figures that flashed on our screens were a world away from the predicted hung parliament, where Labour and the Conservatives were fighting to be the largest overall party.

What, for me, is particularly interesting is that the exit poll was also significantly different to the conversations on social media throughout the election campaign.


Brandwatch analysed ‘social buzz’  (the volume of Tweets mentioning political parties, leaders and topics) on Twitter and, in the majority of cases, it found that Labour were the most talked about party with the most positive mentions.

Even on the day of the election, Labour had a higher percentage of positive mentions on social (39%) compared to the Conservatives’ (36%) and Ed Milliband was ‘winning on social’ with more positive mentions (63%) than David Cameron (41%).


Firstly, let’s look at the demographics on Twitter. According to the British Polling Council, the average Conservative voter is 65+ whereas Labour is represented almost equally in all the different age brackets. Therefore, with the majority of Twitter users in the UK in the 18-34 age bracket (50%), the answer could be that your average Conservative voter is much less likely to be on social media.

It’s also important to think about why people share updates on social media, in particular Twitter. According to an infographic posted by AdWeek 26% of people like to share their opinions on social media and 84% of people want to share updates about issues they care about.

Given these statistics why weren’t Conservative voters sharing their views as much as Labour supporters on social? This is where the ‘Shy Tory Factor’ comes into play.

This phrase, coined in the 1990’s, refers to a predicted one in eight Tories that won’t admit to polls that they voted Conservative. Looking at the Brandwatch statistics it appears that the ‘Shy Tory Factor’ now also impacts results on social media.

Finally, and rather fittingly for this election, you can never underestimate the power of an undecided voter. Investigations are already underway into why the polls got it so wrong this year, so it’s not too surprising that social media also failed to predict the right result too.


To use social as a tool to predict elections, we can’t just rely on measuring the mentions for each party, as people will discuss parties they don’t intend to vote for. We can’t even rely on the positive/negative split, as with the volume of posts around elections this becomes harder to track (and we’re a nation that loves sarcasm).

What could work is a platform that does not analyse what people say about the election on social media to determine the result but instead looks at people’s social profiles to suggest the outcome.

This platform could analyse the key words, locations and previous social posts of people to determine which box they will tick on decision day.

Fingers crossed that somebody somewhere is working on a platform similar to this right now, as it would be fascinating to see the potentially predictive powers of social media play more of a role in the 2020 election campaign.