Eras are defined by their significant materials such as bronze or iron, and there’s no doubt that we live in the information age. In fact in 2018, Forbes determined that 90% of the data ever recorded in the world had been created in just the last two years. From our online shopping habits to geological measurements, everyday we produce mountains of information. A large part of the increase is due to the amount we record passively, particularly through the advent of things like IoT. The commonality of these devices can give us a living picture of our world and its various systems and architectures, if only we can interpret it meaningfully. Supply chains represent one of the most complex and multi-faceted infrastructures to our modern world, placing a great significance on sophisticated use of data around them.
Understanding how and when a change will occur with supply chain planning is dependent on a number of factors, each with their own set of dependencies. Anything from a mislabeled delivery to a ship blocking the Suez Canal can affect the flow of goods and services. Knowing exactly the form and time in which these events occur is vital to understanding how it will affect your workload.
But the sheer vastness of our information can be paralysing. It’s all very well having vast databases, but it’s not truly useful until helpful analysis and interpretation can be drawn from it. This is where machine learning and artificial intelligence have become essential tools, as manually deciphering these data sets would be simply impossible. The huge amount of data has then a huge potential, but only with the appropriate read on it. And even then, it can never provide insight into rare events which have no historical precedent. Data analysts everywhere have a number of key responsibilities when it comes to providing useful analysis for supply chain managers.
Identifying key data for building models is essential for both predictive and prescriptive analytics. These data can be used to assist in the creation of simulation models of key processes which underlie their generation, which, upon validation, can be used to explore “what-if” scenarios, including the rare events which widely-used methods would fail to identify. However it’s also important that the conclusions drawn are able to be communicated and visualised in an easily consumable format. Ultimately, real logistical decisions need to be made and if supply chain managers cannot comprehend the relevance of the data on their work, then it is useless.
Analysts need to be able to provide fast results, as well as structure the data and simulation predictions into a single place where it can be observed visually and clearly. Furthermore there has to be a confidence in the data that goes beyond the analysis being mathematically sound. Forecasts must be made with confidence, as well as being shown to realistically aid decision making. One of the biggest problems for supply chain managers is surprising, unexpected circumstances unfolding. Eliminating the lack of readiness by including the possibility of these events in forecasts – insofar as it is possible for humans to envision rare events – will be of extreme importance.
Data-analysis solutions delivered ultimately must be user-driven, as well as going beyond simple uses of historical data. Live simulations that continuously make use of real-time data offer the next step up; holding a greater ability in finding every potential ‘what if scenario’. With this kind of solution, every risk in supply chain planning can be foreseen and mitigated, ensuring a more optimal efficiency and achievement of KPIs.
Or watch our VP of Solutions, Ralf Paschen, discuss how supply chain managers can use live simulations and all available data to best navigate complex workflows and unpredictable circumstances.