Supply Chain Planning 2030: Prescriptive Analytics Takes a Lead Role

Figure 1. Role of Prescriptive Analytics in Next Generation Decision Support

There is a new technology revolution in the making with the explosion of Internet of Things (IoT), machine learning, self-driving cars, etc., which conjures an image of large-scale automation of decisions based on real-time information. It is a vision that applies  to the supply chain planning world as well.

However, one important aspect has to be understood by everyone: the brain behind the success of a self-driving car is really the algorithms behind it that determine the best route to follow and if there are hurdles along the way (accidents, detours, etc.). The ability to recalculate the best route becomes crucial. In other words, algorithms are a key enabler of this utopian vision of high levels of automation.

So what is algorithmic decision making and Prescriptive Analytics?

In simple words, algorithmic decision making is utilizing mathematical models of businesses to provide visibility to millions of decisions and their impact on key metrics. Prescriptive analytics is a specific type of algorithmic decision making. It optimizes decision making to show companies what actions to take in order to maximize profitable growth, given their business constraints and key objectives. The most advanced in the spectrum of business analytics, it’s able to make the most impact on large scale business objectives, e.g., increasing profit, decreasing COGS, increasing service levels and improving decision-making agility.

How does Prescriptive Analytics fit within the scope of Machine Learning and IoT?

Figure 1 illustrates the architecture of a next generation decision support system. The following are the key points associated with this architecture:

a) Prescriptive analytics based decision support with a planning horizon of several months to several years.

How to go about it?

    • Process Flow: Build the process flow model for the organization (resources, costs, capabilities)
    • Decision Flow: Build the range of possibilities for decisions made by the organization
    • Data: Adopt historical data to feed the model and validate the process flow

Constraints: Identify the upper constraint and lower constraint limits for all decision possibilities. Identify the objective function for the organization

  • Prescription: Optimization identifies the best way to utilize the resources, costs, and capabilities of the organization

The output is a prioritized list of roadmap improvement opportunities. This can also be thought of like the picture you see in jigsaw puzzles that guides a child on how to put the various pieces together. Another analogy is an air traffic schedule or plan that is generated before a plane trip.

b) Track Performance

In this area, the organization can identify the impact of implementing decisions that were made in the previous step and how they perform in real-life. It is key to being able to take decisions in near real-time without manual intervention (or minimal). Machine learning and IoT provide ideal opportunities for enabling this automated scenario generation and decision making. The IoT captures data associated with millions of sensors related to how a company is interacting with its customers/partners/suppliers etc. All of these are stored in a data repository. Machine learning algorithms will read this data repository constantly and identify if certain threshold conditions are reached.

For example, let’s assume sales performance is poor for a particular day or if sales improved significantly due to a social media event. Machine learning will then trigger the running of scenarios through the Tracking/Monitoring component of Prescriptive Analytics — new decisions may or may not be made due to these changes.

An analogy for this component is the need for course correction while flying a plane to reach the original destination after facing different wind conditions, weather issues, etc.


Do not assume that there is going to be a reduced role for planning and analytics in the future. On the other hand, planning and analytics are going to become more important because have to be made more rapidly. Prescriptive analytics and machine learning enable supply chains to be more agile than ever before, and also more efficient.

Wow Your C-Suite with 8 S&OP Strategies

You May Also Enjoy Reading
Supply Chain Brief