Opinion: We need more than data to better predict the future

One of the most interesting things to emerge last year was the lack of predictability in both world and local events.

The supposed experts and pundits seemed to get it consistently wrong. Even the bookmakers, who generally get things right, gave odds of 4,000 to 1 against Leicester City winning the Premiership and got hit badly by some local supporters whose heart overrode their brains and wallets at the start of the season.

The pollsters got it wrong in both the EU referendum and the US presidential election. All the conventional methods of assessing opinion seemed to propose the wrong outcome.

However, two of my colleagues at Loughborough, Prof Tom Jackson and Dr Martin Sykora, correctly predicted the outcome of both votes. In the case of the US election, they did this by analysing the emotions expressed on Twitter about the two candidates.

The measurements were not based on a simple head count of support. If the candidates’ support showed an emotional ‘choppiness’, they lost, while if there was a consistent emotional response, they won.  

This was a new way of looking at the strength of the support for the candidates, which turned out to be the key to predicting the results. More information can be found at am-online.com/prediction.

So what has this got to do with our sector? Well, it again throws down the challenge of whether we are measuring the right things. I have argued in previous columns that satisfaction is a weak emotion and is therefore a poor predictor of repurchase intention.

One of the major challenges is trying to understand whether we are looking at the right data. The rise of social and digital media has made it more difficult to identify what we should be measuring.

If someone ‘likes’ a communication about a car or brand on Facebook, does that have influence on a potential purchase?  Are views expressed on social media relevant?

From the ongoing research we are involved with, it appears people do not wear their hearts on their digital sleeves. Their real opinions are expressed in more controlled environments, such as WhatsApp groups.

Maybe we are looking at social and digital media from the wrong perspective. I have come across WhatsApp dealer/customer groups in Pakistan, India and Nigeria, but could find nothing in the UK. Maybe we should stop trying to search for data, but actually give platforms where we can both create and collect real opinions in an environment in which the customer feels safe.

In the meantime, I am keeping a close eye on what my colleagues in the Business School predict next – it may just be financially beneficial.