Being able to predict the future has been an ambition of mankind’s for centuries. Whereas ancient seers used to base their prophecies on celestial movements, our present day approach to predicting the future is much more scientific. In fact, our desire to control and make sense of the disorder that we see around us in everyday life has led to a branch of statistics concerned chiefly with predicting what is yet to occur: predictive analytics.

[easy-tweet tweet=”Our desire to control the disorder that we see in everyday life has led to predictive analytics” user=”bsquared90″ usehashtags=”no”]

Since the birth of computer modelling, predictions have developed from those involving relatively simple mathematics to the kind that can only be made by the most advanced supercomputers. Early examples of predictive technologies included automated anti-aircraft weapons that used sensors to measure flight speeds and angles to help hit targets. Other early developments included the use of computers to aid weather forecasts, air travel and credit risk decisions.

However, the sea change that enabled predictive analytics to enter the mainstream was the enormous increase in the amount of data being produced on a daily basis. The rise of Internet and smartphone technologies has meant that the amount of data being created is growing exponentially. In fact, the digital universe is doubling in size every two years and is expected to multiply 10-fold between 2013 and 2020. This big data explosion has given predictive analytics software much more information to work with, making predictions relevant to a broader range of businesses.

the digital universe is doubling in size every two years

Predictive analytics has already been implemented across a wide variety of industries and its applications are continuing to grow. In healthcare, Stanford University used predictive analytics to train computers to analyse microscopic images of breast cancer tumours, while Google Flu Trends research was able to analyse search patterns to forecast an increase in influenza cases at a specific hospital. Similarly, urban planners have been able to use data to predict potential traffic bottlenecks and financial institutions have looked to statistics to foresee possible risk levels in their work.

However, perhaps the most commonly used application of predictive analytics is in marketing and the wealth of data now available about existing and potential customers represents a huge financial incentive for businesses. In a recent speech on the potential benefits of predictive analytics, Eric Siegel, an expert in the field, highlighted a less obvious example of the technology being used for marketing purposes.

During the 2012 United States presidential elections, Barack Obama’s campaign used predictive analytics not to determine which citizens would vote Democrat, but which ones would be receptive to campaign activity. The campaign team gave each individual voter a predictive score on how likely they would be to be influenced by certain types of activity – a tactic known as persuasion modelling. In this sense, Siegel claims, predictive analytics “empowers an organisation not just to predict the future, but to influence it.”

[easy-tweet tweet=”predictive analytics empowers an organisation not just to predict the future, but to influence it” user=”bsquared90″ usehashtags=”no”]

Many other marketing campaigns have also used predictive analytics to good effect by identifying customers that are more likely to respond to activity. Using data from previous campaigns, in conjunction with any other information you may have about customers, businesses can produce more targeted materials. Ultimately, predictive analytics doesn’t need to tell a business with certainty that an individual will purchase a product after viewing a campaign. Even a prediction that provides an increase of just a few percentage points in terms of marketing conversion rate could be worth thousands to a business.

While most current examples of predictive analytics can’t tell us exactly what will happen next, the future of the technology is promising. In five or 10 years down the line predictive analytics may be able to improve road safety by foreseeing collisions, predict the likelihood of a household appliance breaking down or provide a number of services that are personalised to your individual taste. In fact, much of this is already technologically possible and it only remains to be seen to what extent consumers will embrace it. As more data becomes available and processing power increases, it’s likely that more and more of the future will become easier to predict.

… it’s likely that more and more of the future will become easier to predict