Is Your Supply Chain Digital Twin Identical or Fraternal?
A Digital Twin of the supply chain, like any other identical twin, needs to change when the physical version changes or else they could be fraternal! Or not even twins! The whole purpose of a digital twin is to represent your physical world accurately. Creating a digital twin of the supply chain is essential, however maintaining it is critical. A supply chain is constantly changing. Your suppliers and their behavior vary over time, they fluctuate depending on the weather, or changes in their management or policies. A supplier’s behavior is not a fixed parameter; it changes over time depending on many factors. Perhaps because of their improved processes, they can deliver goods in 2 weeks instead of the originally agreed on 3 weeks. If the system does not pick up this information, it keeps scheduling deliveries later than possible. Understanding this trend, whether it is seasonal or permanent, can reduce operations cost as well as the ability to deliver sooner to your end customers.
Capacity of your own resources as well as your contract manufacturers are subject to variations too. Your level of confidence in having the right amount of capacity available in the first two weeks of (say) June for a certain location or a piece of essential equipment is a critical parameter in projecting your financials and customer commitments. Knowing this information helps us to allocate the right amount of load to this location or equipment in that period. Thus, we are increasing the likelihood of making our commitments to the customers using stochastic analysis and machine learning.
Your business processes, policies and priorities are also changing depending on the economy and business conditions. When you build a model of your supply chain, six months later your business will have changed. Has your digital model changed with it? is your current safety stock for different products and locations the same as what it was 6 months ago despite the pandemics, tariffs and other supply chain disruptions? Making decision on your old model, no matter how intelligent it is, leads to poor choices simply because the underlying model is not right. A digital twin needs to look and behave like the physical supply chain.
Much like demand planning and statistical forecasting, we can use stochastic and ML techniques to understand these underlying trends and causes on the supply side; and predict what the model should look like and how it should behave now and in the future. This can be done automatically using self-correcting methods to keep the model always up-to-date. Having accurate models is also the first step to autonomous planning.
