Supply chain data analytics may be a significant area of investigation, but there is a distinct lack of case studies that showcase its practical application.
Supply Chain Analytics (SCA) isn’t a new revelation, manufacturers have always been enthusiastic adopters of data-driven technologies. So, apart from the fancy name, what is new?
Alexandra Brintrup, lecturer in Digital Manufacturing at the Institute for Manufacturing at University of Cambridge took to the stage at Industrial Data Summit 2019 to explore data-driven approaches to supply chain disruption management and to highlight the opportunities and potential pitfalls.
According to Brintrup, supply chain analytics is the ability to obtain data from a combination of:
- Structured, traditional enterprise sources
- Your internal and external supply chain
- Unstructured, diverse sources
- And other intelligent supply chain services
Three developments are making real-time gathering and analysis possible: computational power, powerful algorithms and AI computational paradigms, and – crucially – new, previously untapped sources of information.
Click the links below to read overviews of the Summit’s panel discussion, keynotes and roundtables:
Supply Chain Analytics
Brintrup posited that SCA is an umbrella term that encompasses a multitude of capabilities. Not all the capabilities will be suitable for every organisation, nor should this be viewed as a to-do list of objectives; rather, businesses should be mixing and choosing capabilities to suit a particular supply chain function.
Step 1: How much data and from what source?
Traditionally, supply chains leveraged ERP information, a structured source data that was predominantly manually populated with some low-level automation (i.e. RFID tracking). Today, smart products monitor and track their own use, location and health, relaying the information in real-time back to a central hub.
Social media has also become an important source of consumer information and market insight, but the data it contains is unstructured, vast and challenging to work with.
Step 2: Awareness
“Are you using this data to create an understanding of what is going on?” asked Brintrup. “At the very least, you might integrate these disparate streams to provide a dynamic overview of what’s happening. This enables you to start exploring current state analytics and predictive analytics.”
Step 3: Decision capability
“Your data gives you a ladder to step up and see the different systems you are connected to. What do you do with that new found awareness? Can you find optimal solutions for what you do now, or even what might happen in the future? can you learn and adopt that learning to changes?
“Where does your decision reach – is it just day-to-day operations or could you take a more strategic, medium to long-term view? Could you come together collectively as an entire supply chain and optimise across these different layers – vertically and horizontally?
Step 4: Autonomy and control
Brintrup described the automation or semi-automation of mundane supply chain tasks as the ‘Holy Grail’, studying data to uncover, interpret and act upon hidden patterns and trends to improve supply chain operations.
The Institute of Manufacturing (IfM) recently conducted a study on the global automotive industry, funded by a British OEM. This study was followed by others in the FMCG and aerospace.
Their goal was to the map the supply chain structure, understand how disruptions cascade and impact this structure, and then use this knowledge to inject resilience.
The studies generated some key learnings, which Brintrup shared with the audience:
Lesson 1: Don’t underestimate the power of descriptive analytics because it will lead to more realistic solutions
Lesson 2: Don’t go crazy! A minimalist approach may yield the best results
Lesson 3: One solution doesn’t fit all
Lesson 4: Domain knowledge is golden
Lesson 5: Strive to create traceability, accountability and buy-in