Are You Ready To Take The Leap From Descriptive Analytics To Predictive?

Ashish Morzaria

untitled-300x188Predictive analytics and the whole data science realm has really become front and center in the news these days. Somewhere between the hype of Big Data solving the world’s problems and data science proving to be the missing link in every analyst’s toolbox, there’s a little matter of actually making it work in your business. Some would argue that “data science” is just the new term for “statistics,” but statistics never claimed to directly improve your business intelligence systems. If you aren’t considering adding predictive analytics to your BI tool chest, you definitely are missing the bigger picture.

Predictive analytics has a perception problem

First, let’s address the elephant in the room: Most business people think that predictive analytics is the exclusive domain of data scientists. While that might make a few data scientists feel very special, the cold hard truth is that predictive analytics isn’t black and white – it’s a spectrum ranging from light statistics to full-on coding of statistical algorithms by hand.  Saying that business users cannot use predictive analytics is like saying that you need a university degree to drive a car with a manual transmission. One could argue that those with a higher education can understand the mechanics of gear shifting better, but is this knowledge really required to drive a car to the grocery store?

To follow on this driving analogy, predictive analytics is perceived as the “manual transmission of the analytics world” – it requires training, a certain amount of forethought, and if you use it incorrectly, it can make some incredibly nasty sounds. Manual transmissions used to be preferred for those who wanted more versatility, a better feeling of control, and superior gas mileage, but automatic transmissions have improved so much lately that these statements are no longer true.

In the predictive analytics world, there are more manually-oriented tools such as SAS, SPPS, and even the Expert Analytics interface, but there are also more automatic predictive analytics tools. And just like in the automobile world, the benefits of choosing the manual approach over an automatic one are diminishing all the time. Read more about this manual versus automatic analogy here.

Predictive analytics makes business intelligence better

Contrary to how it sounds, predictive analytics is not really about predicting the future – it’s about detecting and understanding the relationships in historical data so that we can make better decisions in the future. If you really think about it, this is the whole premise behind business intelligence itself: understand your business so that you can make it better. Business intelligence is a visual method of representing historical data so humans can identify patterns, trends, and outliers in the data. Predictive analytics uses algorithms to programmatically analyze data to find insights that cannot be found through visual analysis alone.

However, the real value of predictive analytics can better realized when you combine it with descriptive analytics – or what we know today as business intelligence. While it is true that predictive analytics can handle new problems that were either difficult or impossible to solve before, it is excellent at helping you solve or at least reduce some of the problems you are dealing with already. Predictive models generated by analyzing historical data are exceptionally good at “scoring” or highlighting key elements in the data that are worth a further look.

For example, a major Canadian bank analyzed the effectiveness of their previous marketing campaigns and used predictive analytics to find which of their customers were more likely to take advantage of new offers. This bank increased its response rates by 600% while reducing costs by 50% simply by not contacting customers who were likely not interested or would get annoyed by a marketing offer. Not only was the job of the marketing department made easier, but the business intelligence system has far less data to process since a large number of customers were able to be filtered out.

The time for predictive analytics is now

For many customers, predictive analytics is in the very distant future. Part of the reason is the “perception problem” identified above, but also because predictive analytics is new to some – and “new” takes time to learn and implement. Predictive analytics is not be a total cost of ownership (TCO) discussion: It is really about the return on investment (ROI). For example, the bank from the example above realized a 100% increase in their typical ROI for a marketing campaign because it was more effective (by generating more revenue) and ended up costing less.

Once discussion turns to ROI, the real question is: What are your real-world use cases where predictive analytics could make a tangible and calculable impact? How would increasing your average revenue per customer by 10% improve your profitability? What if you could run a customer retention program to reduce churn by 5%? What if you could more effectively forecast demand and realize cost savings by being better prepared for orders?

When you really think about it, these are problems that we typically try to solve by slicing, dicing, and drilling with our traditional business intelligence tools in hopes we find something useful. By adding predictive analytics to your arsenal, you have an algorithmic flashlight to ensure your data cannot hide its insights. Charts can lie, but algorithms are not so easily fooled.

Your move from descriptive to prescriptive analytics is inevitable: If you are not already moving in this direction right now, you can be sure your major competitors are. In today’s highly competitive economy, even the slightest advantage can be the difference between winning and losing. Predictive analytics doesn’t just give a slight advantage – it helps you understand your own data and make better decisions to improve your business. After all, isn’t that what analytics are about?

Follow me on Twitter @ashishmorzaria

Learn more about Predictive Analytics in our SAP Community Network.


Ashish Morzaria

About Ashish Morzaria

Ashish C. Morzaria, Director of Advanced Analytics for SAP, is responsible for SAP’s Predictive Analytics portfolio strategy and driving its go-to-market efforts. He has been with SAP (and BusinessObjects before it) since 2005 in a variety of Product Management roles across the Analytics portfolio including the SAP BI Suite and Platform, SAP Lumira, SAP’s BI Cloud strategy, and numerous acquisitions. Prior to his career at SAP, Ashish spent almost a decade as a product manager, development team manager, and software engineer.