I started my working life at Rolls-Royce plc before leaping into FMCG consultancy.
The immediate difference was the systems, processes and software used to provide in-depth understanding of the production and life-cycle (that were critical to the success) didn’t exist in the FMCG space, simply because they were hugely expensive and complex tools.
Four years later, the paradigm has shifted and the digital revolution is upon our sector. Advances in data availability mean that we must be smarter in using it to our benefit, and advances in computing power mean we can do this cost effectively.
“A digital twin is a virtual representation of a product, part, system, process or network that allows you to see how it will perform, sometimes even before it exists.”
Studies have shown that human computing power using more than four variables is only as accurate as chance, yet in many businesses today we leave complex, multi-variable decisions to people, usually because we struggle to replicate real world complexity in computer models.
However, with enhanced computing power the best digital twins use stochastic modelling, which essentially allows performance of all elements of the model to work based on probability, accurately predicting scenarios where specific events vary from actual performance. These digital twins will not only model scenarios but will determine the risk and probability of success based on replicating the future 100,000 times over (or however many times you need to feel confident).
A tool this powerful can unlock significant benefit within your supply chain and can be used in many different forms, depending on the level of adoption. In its most basic form, a digital twin can be used to improve decision making capability within a business to test scenarios, for example, instantly identifying the optimum production sequence following a critical failure to protect customer service or cost.
Using actual performance data it will consider the full complexity of your factory or supply chain to determine the optimum solution to meet desired outcome. This could offer benefits in product costings, both present and future, not only accurately determining the cost of the new product and how it should be scheduled, but mapping the impact to other products commonly missed.
Gone are capital expenditures or investments that don’t deliver, as you use your digital twin to see how changes in assets, suppliers, customers or sites affect the overall network before agreeing to these changes. This allows businesses to gain confidence in security of supply and true costs through ‘what if’ analysis.
To unlock the true capability of a digital twin, a business should integrate its power into their planning cycle, using the twin to schedule and sequence production throughout the supply chain. A digital twin is designed to compliment existing supply chain software, such as ERP and MES systems which are embedded in the infrastructure. This creates a critical link between the two, without the need for vast amounts of Excel sheets.
When effectively employed, a digital twin can align and optimise multiple schedules inclusive of underlying variation, adding optimisation and risk analysis. It can offer real time allocation and rescheduling of shared resources across the entire value chain (e.g. labor, machines, tooling, jigs, forklifts, AGVs, etc.) It can reschedule a complex, multi facility value chain inclusive of multiple distribution networks all with real world variability and constraints, in mere minutes. In essence, a digital twin will exponentially increase responsiveness to a fluid operational environment keeping a business well ahead of its customers and competitors.
So why now? Hasn’t this type of product been around before?
Linear simulation software has been available for over 30 years, the key difference today is the simplicity of the solutions available and their ability to replicate variability or ambiguity. What this essentially means, is you don’t need a team of programmers; software such as Simio uses simple object-based modelling which can be tailored at all levels of complexity to accurately represent the most complex supply chain.
The beauty of using object-based modelling, is once you have built a model, e.g. a factory floor, this can be imported into your network model as an object without the need for a new model to be created. In essence, it is closer to the real world, maintainable and at an investment cost that is affordable space for the sector.
How do I go about this in my business?
It is worth remembering the philosophy in the Tech world of startups; think big, start small and fail fast. In other words, take your vision, complete small trials to prove concept, and accept that identifying something does not work can be just as valuable as the opposite.
In making the model, it is critical to understand the level of complexity needed for the first decisions you want to make, over time the sophistication and variables can be built upon. You could even connect to the weather forecast or likes on Facebook if that drives your business. We have trained our own digital architects, but more importantly we developed a process to align processes and systems, building complexity with capability and ensuring we do not build a White elephant.
Oliver North, Head of Technology at Pollen Consulting Group.