The election result highlights the limits of social media tracking
“Blimey!” was the sanitised exclamation heard in my house just after 10pm last Thursday when the results of the general election exit poll were announced.
Pollsters, pundits and commentators could be heard ripping up their rulebooks. For months, we had been told this was the closest general election in decades with the result the most uncertain, but in the end that was not the case.
On the day of the election, Marketing Week ran a story, which seemed like many other pre-poll predictions to make sense, declaring the Labour party and its now former leader Ed Milliband had won the Twitter battle with more positive mentions than his counterpart David Cameron and the Conservatives.
One very prescient commentator was quick to question whether such engagement would turn into actual votes. It seemed, beyond the Twittersphere anyway, it did not.
Albeit arguably self-serving, research conducted before but released just after the election by TV researchers otherlines.tv found 62% of respondents claiming TV coverage of the general election, including the debates, had been influential in their votes compared with 11% who said the same for social media.
This is not a preamble for a spot of social media bashing. It does, however, demonstrate the limitations of social media sentiment monitoring. As much as sticking Milliband’s face on Jim Morrison’s body might suggest engagement and positive sentiment towards the subject, it doesn’t track intention, particularly when used in isolation.
There is also the glaring but still notable point that sentiment and or intent on Twitter could be a decent barometer of the social media world, but in the case of the general election there were greater influencers – TV news, newspapers, the debates. There is also the question of whether the demographics of those that populate Twitter – received wisdom has them disproportionately left leaning – are representative and therefore robust conclusions are difficult to achieve.
