Other parts of this series:
Earlier in this series, we outlined the basics of descriptive analytics and predictive analytics. In this post, we will discuss the basic concepts of prescriptive analytics.
Prescriptive analytics is the highest level of analytics and also the most complex to set up and program.
It examines data from a wide variety of internal and external sources and uses a number of different techniques; from machine learning, image processing, applied statistics and natural language processing. Prescriptive analytics not only tries to predict what will happen but also understand why it will happen and present different options on how to take advantage of future opportunities.
Where descriptive analytics might consider data as a pond of information that can be accessed when needed, prescriptive analytics is a stream that constantly feeds new information into the software to allow for real-time analysis and provide information that enables fast and solid decision-making.
It can sometimes be difficult to appreciate where predictive analytics leaves off and prescriptive begins, but one basic distinction is that it translates a forecast into a feasible plan for moving forward.
Prescriptive analytics often uses optimization tactics such as A/B testing to gain much clearer insight into what works best under different circumstances and scenario planning to better appreciate how changing individual aspects of program or product affects the way that individuals like and use a particular offering. Equally important, it can be used to identify what changes are infeasible and could result in poorer performance.
Now that we have a basic understanding of the three basic type of advanced analytics, in our next post we will provide some real-world examples of insurers who are using analytics to enhance their business.