Generating Supply Chain Plans that are Fully Executable

In a recent paper by Gartner, it was highlighted that planning and execution are converging:

“Organizations that have gone through significant investment in demand and supply planning systems as well as manufacturing execution systems often identify scheduling as a critical missing element in aligning and synchronizing planning with execution.”

Many organizations have adopted S&OP solutions which are helpful in providing some visibility. However, because of lack of model accuracy (accurate digital twin), the system is not capable of predicting potential issues. Nor is it capable of prescribing appropriate actions when a disruption occurs. To this end, the users must spend hours, if not days, trying to figure out what to do. In fact, according to some studies the average time taken to respond to disruptions by companies is around 5 days!

In contrast to S&OP models, S&OE solutions are designed to provide a much more accurate model of the operations and because of that they can predict and respond much faster and more optimally. They are designed such that the more detail and more frequently the data is made available to them, the better they can function and learn to perform in real-time to plan and prescribe.

The traditional wisdom has been to plan and execute in two different processes and systems. They are supposed to be somehow “bridged” by passing data between S&OP and S&OE solutions. We believe that planning and execution are a continuum, not separate. They should be so tightly connected that actions taken by one part can immediately show the impact on the other. This is shown in the figure below. Much more value can be gained by creating more accurate models on a journey to automation.

By passing data and messages back and forth, there is bound to be data and decision latency that can lead to unnecessary additional cost and sub-optimal decisions.

There are three ingredients on the journey to autonomous planning.  These are:

• A system that can support and represent a true digital twin of the supply chain,
• Smart algorithms that can understand the model and create solutions that are optimal in real-time, and
• The ability to constantly learn and get better both in the representation of the model as well as the nature of the algorithms that are deployed.

For a more detailed discussion of these topics including self-improving and self-correcting supply chain planning solutions as well as merging of planning and execution, click Here.