1. Sales Tracking

Unlocking the Value of Big Data via Prescriptive Analytics

predictive analytics

Prescriptive analytics has tremendous potential to become a disruptive force on making business decisions more effective and efficient. Done right, big data and analytics can enhance customer engagement and drive big returns.


Overall, raw data, particularly big data, does not offer much value in its unprocessed state. Yet, applying the right set of tools, we can extract powerful insights from a stockpile of bits.

 

Descriptive, Predictive & Prescriptive Analytics

Analysing data is nothing new as organisations have been doing this for ages. Business analytics can include descriptive, predictive and prescriptive phases. Descriptive analytics looks at past performance and its objective is to understand business performance by mining historical data to uncover reasons for past success or failure. The majority of business analytics are descriptive such as social analytics in terms of fans, mentions, page views, likes, check-ins, etc.

On the other hand, predictive analytics answers the question what will happen. Hence, historical performance data is combined with rules, algorithms, and occasionally external data to determine what is about to happen. Thereby, predictive analytics can only forecast what might happen in the future, because all predictive analytics are probabilistic in nature. For instance, sentiment analysis involves predicting data that we do not have, which is called the sentiment label, whether it is a positive or negative sentiment.

Prescriptive analytics is the third and final phase of business analytics and its objective is not only to predict future outcomes, but also to make recommendations based on those pre-specified outcomes. In focusing on the what, when, and why of future events, it attempts to answer the “So what?”-question.

Prescriptive Analytics

 

Prescriptive Analytics & Five Pillars of Success

Currently, a small percentage of companies use prescriptive analytics (Bertolucci, 2013), and even those tend to do it with structured data only. Despite being quite new, prescriptive analytics promises to be extremely powerful and accurate in its predictions through machine learning, artificial intelligence, and mathematical sciences. This will drastically improve decision-making, but it requires a predictive model with actionable data and a feedback system, which Basu (2013) also recognises as key components in his five pillars of prescriptive analytics success:

1. Hybrid data: Using both unstructured (text, image, video and audio) and structured data (numbers and categories), businesses take full advantage of all the available data to make the best decisions possible.

2. Integrated predictions & prescriptions: The functions – predictions and prescriptions – must work synergistically for prescriptive analytics to deliver on its promise.

3. Prescriptions & side effects: Operations research takes into account the objectives, the constraints and the actionable knobs (decision variables) to produce the best course of action that does not lead to undesirable side effects.

4. Adaptive algorithms: As this business process evolves over time, the technology should continually re-predict and re-prescribe so the predictions and prescriptions remain relevant.

5. Feedback mechanism: Companies with highly sophisticated prescriptive analytics software are dependent on humans to act on the prescriptions coming out of the software.

Of course, in order to practice descriptive analytics, massive amounts of data are required. Fortunately, we live in a world that is constantly producing exactly that.

 

Case: Google’s Self-driving Car

Google’s self-driving car is an example of prescriptive automation as it automatically act upon the prescriptions coming from the software side.

During every trip, the car makes multiple decisions about what to do based on predictions of future outcomes. For instance, when approaching an intersection, the car needs to anticipate what might be coming in terms of traffic, pedestrians, etc. and the effect of a possible decision before actually making that decision.

 

References:

Basu, A. T. A. N. U. (2013). Five pillars of prescriptive analytics success. Analytics Magazine.

Bertolucci, J. (April 15, 2013). Prescriptive Analytics and Data: Next Big Thing? InformationWeek.

Evans, J. R. (2012). Business analytics: the next frontier for decision sciences. Decision Line, 43(2), 4-6.

Mohan, D. (2014). Your Data Already Know What You Don’t. Exploration & Production Magazine.

Simon Raun Madsen

I am a marketing, business, and communications academic and practitioner with strong cross-disciplinary skills from courses and work experience within the fields of marketing and business intelligence. In particular, my passion for technology and several years of experience with online marketing have provided me with a flair for digital solutions, data analysis, and front-end development.
Simon Raun Madsen

Latest posts by Simon Raun Madsen (see all)