How Social Listening can Reveal Public Opinion
2016 was not a good year for the experts. From Trump to Brexit to The Apprentice, it seems that no one saw the big events coming. In final weeks, the experts tumbled over each other to find explanations for the inexplicable, citing such factors as Fake News and Post-Truth Politics – almost as if the last year heralded an unprecedented era of uncertainty.
It has been seen as enough of a problem for a formal investigation has been set up to determine where polling is going so wrong. But there’s no point in this enquiry – the reason is already perfectly well-known.
Those of a certain vintage will remember the 1992 UK general election, when Labour was expected to sweep to power after 13 years in the wilderness. In the event, John Major’s Conservatives confounded the psephologists by winning the largest number of votes in British history.
This was attributed to the “Shy Tory Factor”: the phenomenon that, when polled, Conservative voters were reluctant to disclose their true voting intentions. Those with slightly longer memories will recall the Bradley Effect in 1982, when California failed to elect an African-American candidate who had led the opinion polls right up to the vote.
So there’s nothing new about 2016’s “surprises”: the limitations of polling have been apparent for years—in fact, since at least the 1950s. The Hawthorne Effect was coined following efficiency studies at an Illinois factory. These studies discovered people modify their behaviour or their opinions in response to their awareness of being observed.
The limitations of traditional polling have been apparent for years.
Bradley and Hawthorne: it sounds like a respectable firm of family solicitors, or an old-fashioned building society. But these names actually show the historical limitations of traditional polling. It seems that when faced by a pollster many people are, to use Alan Clark’s phrase, “economical with the actualité”. They lie.
To Know What People Think, Observe Them
What lesson does this teach those of us who earn our living by gauging and influencing public opinion – people such as politicians, pollsters, marketers, and more? Simple: if you want to know what people think, don’t ask them.
That’s how we used Spredfast to accurately predict the outcome of Brexit. We didn’t conduct a phone poll, nor stand in the street with a clipboard—we spotted that social conversation was more pro-Leave than polling suggested was actually the case.
Referendums are national, not local affairs, and social conversation is not an opinion poll. Our analysis of social chatter did show building momentum on the Remain side, with Remain conversation up 69% in the last seven days before the results. In the same timeframe, Leave only saw a 39% uptick.
It’s easy to mock social media as a source of psephological enquiry. Sites such as Facebook and Twitter are perceived as “echo chambers”, full of fake news, false flags, and sock puppets. They are full of millennials—and very little else.
This a gross misunderstanding of how universal social media has become, but it’s a common reason why these sites are neglected by pollsters. This is a shame, because people tend to be much more honest about their opinions on social media than they would being door-stepped by a pollster, for example. Unlike employees at the Hawthorne Works, social media users don’t realise they are being watched by anyone outside their own intimate circle of friends. Even if Facebook (and other social sites) can be echo-chambers at times, it at least creates a safe space where people can express what they truly feel.
This opens up huge opportunities for the people and brands that are listening.
Listening to what people want on social media is the perfect R&D playground for brands. Pushing out a product, ad or new link and listening to the conversational response is incredibly valuable. Rather than throwing money at polls and trusting it enough to invest all your marketing budget in it, social allows real-time feedback. Given some brands handle more than 10,000 social engagements a day, they can glean insights at scale and each new interaction should be seen as an opportunity to build a relationship.
Social media isn’t perfect, but innovations in social listening are developing as fast as the platforms themselves. The latest technology allows brands to mine unlimited real-time social data more accurately than ever before.
Innovations in social listening are developing as fast as the platforms themselves.
So as we start 2017 the lessons from history are there to be learned—as, indeed, they have been for decades. Now we have the technology to drop in on people’s thoughts (expressed, it must be stressed on public forums) and monitor sentiment on a massive scale and in extreme detail. We need to employ this technology to find out what people really think—and not just what they think it’s safe to tell us.