Predictive Analytics Tools are Only as Good as the Data They Rely On
Businesses can use predictive analytics to gain a genuine competitive edge. But only with strong data backed up by the right expertise.
Everyone has heard of the old saying “garbage in, garbage out” and nowhere is this truer than in the world of data analytics. It might sound obvious, but today’s business leaders can become so obsessed with finding the latest forecasting and modeling tools that they forget to pay enough attention to the very data that is being fed into those tools.
An equally important ingredient to the perfect recipe is the knowledge and experience to interpret the analysis and thereby to make the right decisions. Get those two factors right, and predictive analytics can provide your business with a genuine competitive edge.
What is predictive analytics?
The field of predictive analytics is a branch of analytics that uses both historical and contemporary data to forecast future trends and behavior. The sheer quantity of big data that is now available, in combination with advances in machine learning and data analytical tools, mean that it is an area at the very cutting edge of technology.
Early adopters have included financial service institutions, marketing organizations, insurance companies and online services, providers. The applications are numerous – for example, it can bring real benefits in such activities as fraud detection, predicting mechanical failures, targeting online advertisements and identifying patients at risk of developing particular medical conditions.
Get the data right
The are obvious, but unfortunately, it is just as clear that even with the tools at our disposal, it is only too easy to get things wrong. We need only look at the results of the EU Referendum or the US Presidential Elections to see some of the most high-profile instances imaginable where the forecasters got it very wrong.
Throughout the election campaign, Hillary Clinton relied heavily on data analysis to inform strategic decisions regarding where to focus her campaign. The idea was that these insights would provide her with a competitive advantage over the opposition.
Experts cite this as an example of the results being thrown off track by the integrity of the data. In this case, it was information from opinion polls, data which traditional analysis would have spent far more time scrutinizing to understand where it came from and whether it was leaning towards particular demographics.
The downside to machine learning is that it does not necessarily care how the data is created. It simply uses what it has been given without question.
Back it up with the right knowledge
This brings us to the second ingredient. Statistician George Box has generally credited with the aphorism: “All models are wrong.”
His point is, of course, that every model is an inaccurate reflection of reality from which we can derive useful information – but only if we know where to look and what we are looking at.
Kenneth Sanford is a Professor at Boston College. He remarked in a recent interview: “There is never going to be a silver bullet with some single software panacea; it will always take expertise in data and in the business.”
The implications are clear. 21st-century advances in technology are making predictive analytics an ever-stronger tool for businesses. But it will never bring any benefits without the right data and the industry expertise to understand it.