Why you can't trust social media


Social listening can be considered a window into our thoughts and used as a basis for research into trends, reaction and commentary.

Entire campaigns, creative platforms and even brands are built on a lot less. It’s become a common source of data because it’s readily, publicly available and there are many inexpensive tools that can mine it for us. But in my head, there’s a fundamental flaw with social listening. Social media doesn’t reflect what we really want to say, how we would naturally say it and it doesn’t represent that which is never said at all. The thought has been passed through so many filters before it’s posted (both figurative i.e. “does that make me sound cool/clever/funny enough?” and literal i.e. “Valencia or Sierra?”). We may be presenting a “hyper-idealistic” version of ourselves by self-editing. What about all the thoughts you had that you didn’t think were worth posting? Even not posting those is a filter of sorts. So already we’re working with a suboptimal data set.

The reasons you’re on social will also affect how and what you post, making yourself look as favourable as possible. Do you post online to build your personal brand as an extension of your work, for example? Chances are that you’ll be filtering out specific thoughts and opinions, talking more about industry trends, opinion pieces and relevant topics, less about your opinions on things like current affairs or your personal life. You can safely assume I will have done at least a few revisions of the intro as I post this piece to promote it on social… The irony isn’t lost on me. Some say certain social media platforms (for example Twitter) have become such toxic landscapes that trolls reign supreme when opposing opinions are expressed. This is enough of a deterrent for many to stop voicing their opinions at all, and may even force them to leave altogether, like Guardian journalist Lindy West. This doesn’t mean they don’t have opinions, it just means they won’t show up in a social listening report anymore (and to skew this data even more, most of the social listening tools I’ve used in the past seem to massively over-index on Twitter content).  

We all know that social media doesn’t reflect reality, so why do we fall into the trap of thinking it does? There is also the 90/9/1 theory which poses the idea that the vast majority of chatter on social, forums and comments sections is coming from 1% of users, a bit is coming from 9%, and radio silence coming from the other 90%, also known as “lurkers”. Any trends we glean from this type of data is clearly not going to be representative of a whole group of people (unless you specifically want people who are vocal on social… in which case, jackpot). If you base your campaign on any “insight” found here, you could plough ahead in completely the wrong direction... and this should make us insight people sad (and believe me, it does). This is why social media and polling haven’t accurately reflected what we’re actually thinking, which has led to shock outcomes for Trump, Brexit and Theresa May. We took those “temperature checks” as a representative sample of public opinion, when really they weren’t reliable at all. We’re still filtering what we’re saying to make ourselves sound better, withholding opinions due to concerns over privacy or fear of judgement about our real opinions, or we’re simply not being asked (and not actively volunteering that information into the public sphere where it can be heard). We may not even be conscious that we’re doing it. It’s been said that the reason why the polls got it so wrong with Trump is that methods of polling and collecting this data were flawed and people didn’t say (visibly, out loud, to polls, on social) what they really thought, as outlined in this Vanity Fair piece.

This isn’t to say there’s anything wrong with our self-preservation on social media, it’s our prerogative to share only what we want - or to *shock* not be on social media at all if that’s our choice. What I am saying is that analysis collected in this way is like looking at just one piece of the puzzle, an incomplete data set; albeit a fascinating one that does tell us about the type of people we want to be seen as. Once we start to accept the limitations of social media listening (and polling) as a self-selecting/editing sample, then we know what we need to do to improve it or supplement it. It can’t be the sole source of insight, no matter how cost-effective it is. We must complement and validate it with other types of research, such as conversations with real people, qualitative research by specialist agencies that can see intent and unconscious intent behind the words that people say if we’re to understand people and connect with them in an authentic way.

BlogSarah Evans