Published on the 21/08/2018 | Written by Pat Pilcher

Who needs a crystal ball when you’ve got a better mouse-trap?…

Business analytics has been around for some time, but a newer branch, known as predictive analytics is helping businesses predict the unpredictable.

It sounds a lot like some weird form of black magic, but the reality is that predictive analytics makes use of established techniques such as machine learning, data mining, statistical modelling and AI. These combine to make increasingly accurate predictions about future events that would be all but impossible for a human to predict with any degree of confidence.

“Sophisticated analytical techniques are embedded into more and more applications.”

Several factors are driving the uptake of predictive analytics, the most significant being the rise of big data. It has gone from being something used by multinationals with deep pockets and plentiful resources to be a viable tool for smaller businesses. Much of this is thanks to increasingly cost-effective tools and methodologies that allow enterprises to manage the volume, velocity and variability of big data. Just like gold mining, there are invaluable nuggets of information hidden amongst a sea of big data. The rise of affordable hardware, cloud architectures and open source software have helped to bring big data within reach of smaller businesses.

Even startups can purchase server time in the cloud. Additionally, there is a growing awareness of the benefits associated with business analytics. This awareness has been given added momentum thanks to easier-to-use analytics software that is bringing analytics into the mainstream. It is no longer the exclusive domain of statisticians and mathematicians, but the tool of choice for business analysts. Perhaps the single most crucial driver of all is prevailing economic conditions which are pushing many businesses to turn to analytics to develop competitive differentiation and to work smarter as well as harder.

IDCs global research manager Chandana Gopal also sees the sheer volume of business data generated as driving the uptake of predictive analytics: “Sophisticated analytical techniques are embedded into more and more applications. Forward-looking analytics is going to become much more mainstream, as enterprises harness more and more data from a variety of sources.”

IDC is bullish in their outlook saying that predictive analytics software market reached $3.1 billion globally last year and is projected to grow over the next five years by 9.4 percent.

Like most technologies, predictive analytics holds with the GIGO (garbage in/garbage out) principle. Its accuracy tends to be reflected by the quality of data used as well as the modelling principles applied. When it comes together, predictive analytics can be very powerful says Scott Bailey, VP of strategy for Target Data: “We have built a model that can predict within 75 percent accuracy the likelihood that a home will sell in the next 30, 60 or 90 days.”

One of the first questions asked by those curious about predictive analytics typically revolves around the dangers of overestimating predictive analytics capabilities. Bailey says that this involves having realistic expectations: “There will always be risk or ‘error’ involved. For example, we have built a model that can predict with 75 percent accuracy the likelihood that a home will sell in the next 30, 60 or 90 days. That means we are wrong 25 percent of the time. But, only about 15 percent of the homes on the market sell in a given 60 day time period. So, if you just guessed whether a group of homes would sell in the next 60 days, you would only be right 15 percent of the time. Suddenly, 75 percent looks really good!”

The uses for predictive analytics are many. Fraud detection is popular. It works by combining different analytic methods and datasets, allowing enterprises to detect, and potentially prevent criminal behaviour. Predictive analytics can also play a role by examining network activity in real time to identify activity that may indicate zero-day threats, vulnerabilities and other cyber threats.

Optimising marketing campaigns is another task to which businesses are increasingly applying predictive analytics. Doing so helps remove the sorcery from marketing, making it less of an arcane art and more of a measurable and repeatable science. The upshot of this is that businesses are better placed to predict customer responses, purchase patterns, and to spot other less obvious promotional opportunities.

In addition to tweaking marketing campaigns, predictive analytics is also helping enterprises improve operations by providing more accurate forecasts of inventory, unforeseen spikes demand and other business resources. A typical example of this is airlines who use predictive analytics to forecast demand for specific routes to set airfares. Hotels have also been known to use predictive analytics to predict guests numbers and maximise occupancy rates and revenues.

Perhaps the most critical consideration of all when it comes to predictive analytics is that while it is a powerful tool, it isn’t a universal cure all for business. The last word goes to IBM’s Christer Johnson, head of advanced analytics: “First, decide what problem you want to solve.” In other words, while big data and predictive analytics benefits from an enterprising spirit, the best results come from having concrete and realistic goals to start with.