[vc_button title=”Download Prescriptive Analytics Guide” target=”_blank” color=”default” href=”http://butleranalytics.com/wp-content/uploads/2014/05/Prescriptive-Analytics-Guide.pdf”]
Excerpt:
After fifty years of using information technology to increase the efficiency of business processes we are now firmly in the era where technology is also being used to provide us with information. Business intelligence allows us to establish what has happened and is happening in our business (often called descriptive analytics), and predictive analytics uncover patterns which can be useful in the prediction of future events. This doesn’t complete the picture however.
Descriptive and predictive analytics may tell us what has happened and what may happen, but they do not tell us the best way to deploy our resources to meet the demands of the future. An example will clarify. In a retail environment our descriptive analytics will tell us sales volumes, seasonal fluctuations and so on. Predictive analytics may give us insights into which products tend to be purchased together. Armed with this knowledge we then need to know how shelf space should best be allocated and more generally how resources should be utilised to maximise revenue and/or profitability. This is where prescriptive analytics fits in – think of it as a prescription for action.
The major part of prescriptive analytics is concerned with resource optimisation given a set of business rules (constraints) and predictions relating to demand, customer behaviour, the success of marketing campaigns and so on. In real business problems, optimisation may involve thousands of variables and constraints, and finding the optimal use of resources, given an objective that is to be maximised or minimised, can only be achieved using powerful computerised optimisation software. Examples abound. Airlines use prescriptive analytics to determine the allocation of seats to each particular class. Vehicle rental businesses optimise the positioning of vehicles to maximise revenue and profitability. Energy companies increasingly use prescriptive analytics and especially with the unpredictable nature of renewable energy sources.
Of course this all assumes that business managers buy into the resource utilisation schedules created by prescriptive analytics techniques. As such the analytics initiative needs high level sponsorship and coordinated effort throughout the enterprise. Reporting mechanisms need to be put in place and procedures to deal with the inevitable changes of circumstances all businesses experience. To this end some businesses run some of their prescriptive analytics processes in near real-time to accommodate change, and such is the power of the optimisation algorithms and computer hardware that this has become possible for complex analytics tasks.
Prescriptive analytics is clearly not a trivial undertaking. It needs close liaison between analytics teams and business management, and an integrated analytics environment capable of integrating business rules, predictive models and prescriptive analytics. The integration is important, and particularly in large complex businesses. Without such integration prescriptive analytics may be very difficult to achieve, if not impossible.
Expect to see prescriptive analytics technologies more widely used as the user interfaces become more user friendly, and business managers become empowered to address increasingly complex optimisation problems without recourse to teams of analysts. However for large, complex prescriptive analytics tasks the analytics teams are here to stay.
While most analytics technologies are concerned with what has happened or will happen, prescriptive analytics tells how to best deploy resources to optimise our operational activities – and the benefits are often substantial.