In 2013, I spoke about the theory of “social media bubbles”. This concept was unheard of three to five years ago. However, the internet and social media such as Google, Facebook and Twitter have evolved and are now powered by sophisticated machine learning algorithms that work endlessly to predict user behaviour based on a set of online activities, mouse clicks and keywords. Some of the common machine learning algorithms used are Market-Basket Analysis, Naive Bayes Classifier and many other derivatives.
In the last decade or so, most of the decisions by communication strategists, agencies and PR practitioners are based on “folk theories”. It means theories that have been developed, tested and refined over time by practitioners about what works and what doesn’t, without much empirical evidence to validate the effectiveness of the campaign outcome and the rationale behind the campaign design.
In this article, I will share a bit more how we have adopted social data with other data sources to predict an election outcome in a small constituency in Malaysia with 97% accuracy.