With very few exceptions, existing S&OP solutions rely on a centralized application in order to provide visibility and plan operations of the supply chain. There are a couple of major flaws with this approach: they are too big and slow to handle data granularity and frequency. Both of which are a must to create a digital twin. To this end, they cannot plan fast enough and respond to real-time events. Instead, they rely on users to create executable plans and run what-if scenarios resulting in sub-optimal decisions. S&OP solutions rely on yet another system, with a disjointed logic and model, to address the execution of the plans. This leads to a disjointed process of planning and execution that are incompatible. Inability to model granular data with fast enough data frequency leads to rough models and outdated models of the supply chain resulting in inaccurate financial projections and unreliable commit dates. It is simple, accurate models deliver accurate results with the right level of intelligence built in to guide the system for planning and responsiveness. S&OP and S&OE should not be separate systems but a continuum.
Supply chains are large and complex as well as multi-dimensional. S&OP’s horizontal planning is not enough. They need to plan vertically and be re-entrant. Moreover, take into account a plethora of other factors that are constantly changing. Every plan needs to accommodate for regional regulations, accurate serialization, environmental factors such as plan’s carbon footprint, geopolitical issues and trade disputes and many others. A big centralized system cannot possibly deal with all this data and be able to produce a plan that can respond to supplier/customer issues in real-time, monitor weather and other geophysical disruptions and respond fast enough to keep the supply chain resilient.
In order to be able to create solutions that are scalable and perform on-going and increasing complexities of the supply chain, it is imperative to have a solution that can sense, act, and learn on an on-going basis. In other words, a distributed environment with “intelligent sensors” that collaborate and even negotiate with each other when events occur to respond properly and in a timely manner. This is possible using Edge Computing and Swarm technology. With Edge computing, business processes are positioned as intelligent and distributed agents that can sense, act, learn and negotiate with each other. Negotiation is carried out using Swarm technology, much the same as what bees and ants do. They do not make decisions by consensus but by signaling to each other the best options and come to an agreement for the greater good of the whole community. Consider a sales person asking for expedited delivery at a higher cost because not doing so may cause loss of a customer. If the CFO and COO decide otherwise, they “win” and the customer is lost. So, decisions based on majority voting are not always the best decisions. Same kind of argument applies when product managers fight over dedicated capacity to their products.
For a truly autonomous supply chain planning environment that is constantly changing a multi-dimensional and distributed approach offers a scalable, flexible and extensible solution to provide optimal operations in real-time. The future belongs to companies that can make decisions faster and better. Stock market is a perfect example. There are tens of millions of investors having access to the same data. The winners are the ones with better and faster algorithms to react!