Data analysis and prescriptive analysis are not something new, in the past known only as Business Intelligence or descriptive analysis had the purpose of answering the question: “What happened within our organization?”.

With competition on the rise and customers increasingly difficult to retain, it is necessary to create competitiveness and grow sustainably through responses that go beyond the descriptive analysis that we explain in this post. Companies need to obtain predictions for future actions where the application of Big Data and Machine Learning has become indispensable in recent years.

Prescriptive Analysis

Prescriptive analysis has been hailed as the last piece of the Big Data puzzle. The whole process begins with descriptive analysis – describing what happened and why, along with predictive analysis – predicting what might happen.

The objective of the prescriptive analysis is not only to predict results but to offer an insight into what actions to take according to them, suggesting the best course of action among the available options. This can completely change a current scenario of a company and may have physical, marketing or management changes for example.

Prescriptive analysis promises to be extremely powerful using historical dados, business rule algorithms, computational modeling techniques, constraints and variables, and machine learning algorithms, where these are combined to get the best predictions and decision making.

One of the best-known examples of using prescriptive analysis is google’s car, the one that drives alone. During each trip he makes, he makes several decisions about what to do based on predictions of future results, for example: when approaching an intersection the car needs to stop and determine whether it goes left or right, and based on several future possibilities makes a decision. So the car needs to anticipate what can come in terms of traffic, obstacles, pedestrians, etc., finally, what will be the result of that decision before actually making that decision.

Another example we can see is in the oil and gas industry, prescriptive analysis allows us to analyze a variety of structured and unstructured data sets (including video, image, and sound data) to predict whether or not new oil wells open.

In other more conventional business types, typical examples of prescriptive analysis applications include inventory management, production planning, operational resource allocation, supply chain optimization, utility management, transportation and distribution planning, marketing mix optimization – sales, pricing, and financial planning.

Anyway, it can be used in any segment.

Where can it be applied?

It applies in situations where decisions need to be based on many data collections and variables, where the human mind is unable to evaluate everything without the use of technology.

They are ideal for situations where experimentation in business operations would be excessively risky, expensive, or time-consuming. Analytical models and simulations are performed using complex, randomized variables to discover and optimize the range of potential results.

Prescriptive analysis can also identify decision options that take advantage of future opportunities or limit future risks by illustrating the implications of each decision.

How does it work?

Briefly, we can say that in this analysis the algorithms are programmed in such a way that they can assume and adapt based on information and parameter changes established by data analysts or data scientists.

Synergistically it combines data, business rules, and mathematical models. Data entries can come from various sources, internal (within the organization) and external (social media, website, surveys, etc.). The data can also be structured where it includes numerical and categorical data, and also unstructured, such as texts, images, video and etc.

Most prescriptive analytics is concerned with resource optimization given a set of business rules (restrictions) and demand forecasts, for example customer behavior, the success of a marketing campaign, and so on. It can continuously collect new data to review forecasts and prescriptions automatically, thereby improving the accuracy of forecasts, and prescribing continuously more insightful and impactful decision options.

It is important to remember that it does not work alone, it needs a close link between the data analyst and business management, and the data collected for full integration.