Wednesday, October 23, 2013

Predictive Analytics vs. Prescriptive Analytics

The Analytics is the Link between Information and Knowledge; but What's the Link between Knowledge and Wisdom?

The evolutionary journey of data analytics reflects the trend moving from analyzing historical perspective into capturing business foresight, from operational perception to customer insight; from predicting what will happen to step further: What should you do upon it; or put simply, from traditional BI into advanced analytics; and from predictive analytics to prescriptive analytics.

Predictive Analytics is to predict future outcomes; Prescriptive Analytics goes one step beyond  and tell us not only what might happen (prediction) but also prescribes an action (Prescription) in order to convert the information into actionable knowledge.

Predictive Analytics is to win foresight: What will happen? Which customers are most likely to respond to your next offer? What do customers want for the next cycle of products/services? Prescriptive Analytics is to lead decision making: What should you do? Which offer should you make to customers? Though prescriptive analytics need not be confined to actions taken based on predictive analyses. One could prescribe many things analytically without predictive response analyses. 

Logic scenario from decision to action: You have your past data -> Model building -> Prediction based on your model -> Action based on prediction and its implications (Prescriptive Analytics). Prescriptive analytics is suggesting actions to benefit from the predictions and showing the implications of each decision option. Prescriptive Analytics defines a set of analytical capabilities that include:
- Define a preferred course of action
- By calculating expected outcomes of alternative decision options

Additionally, statistics combines analytics with a decision engine, that enables direct integration of domain knowledge into modeling, enables conclusions from models in terms of prescriptions, usually referenced as prescriptive analytics. You may create a model to predict when assets or machinery may fail and need preventive maintenance. Prescriptive analytics will provide you with the optimal order in which the work orders need to be serviced within the budget, time and other constraints. 



In any advanced analytics scenario, thinking and intuition will always be required. In fact thinking is inevitable! To do both predictive and prescriptive analytics, you have to ask questions whose answers are potentially actionable. And keep in mind, analytics is not the big ANSWER looking for the business problems, instead, how to frame the right question is the key.




0 comments:

Post a Comment

Twitter Delicious Facebook Digg Stumbleupon Favorites More