With a matter of weeks until Britain goes to the polls to vote on the future of our EU membership, can data insight be used to predict the outcome? At equimedia, we have put our data capabilities to the task of answering the question that is on every news pundit’s lips: which way will Britain vote?
The right analysis of data can unearth all manner of insights into the population - their interests and attitudes, their behaviours and, importantly for marketers, what drives action.
This is not the first time we have used insights from social media and other sources to gauge public opinion and predict the result of a referendum. In 2014 we correctly forecast the result of the Scottish Referendum.
The use of big data and social signals allowed us to identify the underlying voting intentions of the Scottish population and our assertion was proven right.
However, when tracking data, we need to learn from results and re-calibrate the technology at our fingertips to respond to any problems. After applying our model to the 2015 general election, the results did not marry with our predictions. The forecast was skewed by a lack of women sharing their political opinions online - men proved more likely to voice online support for their party. This distorted the results.
Recalibrating is essential for accuracy and we will discover the results when our election tracker and data model are tested against an extremely complex issue.
While it’s interesting and enjoyable to speculate on referendum results, the ultimate goal for marketers is to apply this kind of methodology for businesses and enable them to know what people think.
If you truly understand consumer opinion, you can better tailor your message to your customers. Too many campaigns have all the directional focus of a blunderbuss. What really delivers results is sniper-like precision. Messages that don’t speak to customers’ concerns or interests will not resonate or drive action. Messages that are personalised and relevant produce the best results.
Admittedly, it’s challenging. Opinion is not a static thing - it is constantly in flux. But with the right algorithm, it is possible to track and measure it. However, an automated algorithm alone is not enough. You need intelligent people who are able to combine the technology with analytics to provide meaningful insight.
Making a prediction is risky. What if we are wrong? What if we have misinterpreted the data? There is, after all, a huge amount of data sources to handle and countless factors that could sway Joe Public. Fear of failure is far more risky. Effective use of data can enable us to predict not only voter behaviour, but any other aspect of consumer behaviour.
It is only by making predictions, testing and adapting models that we can learn how to best use available data. Proficient data management is the future of marketing and only through the effective use of data will clients know what to look at and what to do in order to achieve lasting business success.