Does Predictive Analytics Have the Higher Value than Descriptive Analytics

Predictive analytics relies on the analysis of the past as well.

Doing Business Analytics is an evolutionary journey, from descriptive analytics, to predictive and prescriptive analytics, every step provides organizations different perspective of their business: Descriptive analytics focuses on analyzing what has happened based on the historical data; the predictive analytics tries to figure out what will happen, the future trend, and then the prescriptive analytics suggests the necessary actions businesses can take. Though predictive analytics is more forward-looking and advanced, does it mean that the predictive analytics always have the higher value than descriptive analytics?

Organizations need both, predictive analytics and descriptive analytics. You may not compare the reliability of Predictive vs. Descriptive that directly… one is understanding what has happened and why, the other is predicting the likelihood of a specific outcome happening. They are used for different purposes, provide different results, and answer different questions. Organizations need them both, and like most things, with the right patience and testing, predictive can deliver massive results. With smart meters, power companies now have a lot more data to include, and it's a massive enough industry that the "Power of a 1% improvement" could be significant.

Predictive analytics relies on the analysis of the past as well. You develop a predicting model based on the information you have from the past. The quality of your past data defines the quality of your predictions. If your past data lack any trends then you may design a predicting model based on white noise resulting in nonsense. The trend involves retrospective and prospective data. And, predictive analyses are immensely less reliable than descriptive analyses. So it is really important to choose carefully which past information you have to use in order to train the model. Descriptive analytics helps in this procedure called 'features selection'

The predictive analytics is a step that businesses usually take after descriptive analytics. Descriptive analytics is essential in order to identify the scope of predictive analytics and tune them up. In this fashion, Although predictive analytics is the emerging trend, businesses shouldn’t play down the importance of "old school" analytics and the insights that may be populated - given that they are performed correctly! Know your data, then see what it predicts. But it's all in what you do with it that ultimately matters. The true power of Predictive Analytics comes when you integrate the insights into your business processes (applying the model in real-time to predict the likelihood a specific person will take a specific action).

From noble business purpose perspective, both descriptive and predictive analytics help optimize various business management, directly or indirectly related to long-term revenue, descriptive analytics is applied more often to the traditional optimization such as operations research, root cause analysis or marketing efficiency; while the predictive analytics can be used to predict the customers’ need, the future product/service trend, talent management, etc. Organizations can track the behavior of the past decisions to support future business directions by using them both effectively.

Posted in: Big Data,Business Intelligence,Connecting Dots