Industry Talk
Data Analysis and the Smart Supply Chain of the Future
Data Analysis and the Smart Supply Chain of the Future
Fred Hartung,
Vice President, Supply Chain Solutions & Logistics at Jabil

Over the last few years, I’ve read with interest articles imagining the supply chain of the future. Forward-thinking writers have cast their gaze over the industry and made bold predictions about the technologies, both emerging and yet to be developed, that the logistics industry might benefit from. Yet in many ways, I think the supply chain of the future is here already. This belief isn’t based on a gut instinct or an educated guess. In fact, it’s driven by data. Data with the power to enhance logistical operations lies everywhere. But data by itself is useless - it must be analysed in order to generate insights that can improve performance. To achieve this, we use a proprietary intelligent supply chain digital platform called ‘InControl’, which oversees all of Jabil’s data capture, quality and analysis. InControl works across four main types of analytics:

  • Descriptive: Metric-type analytics that report the status of the business 
  • Predictive: Analytics that uses data, past and present, to predict a future state 
  • Prescriptive: Actionable analytics that show what actions need to be taken to achieve the desired result 
  • Cognitive: Ability to analyse and find patterns in huge data sets 
Through the application of complex event processing and cognitive analytics, it’s possible to possess a deep understanding of the entire supply chain ecosystem and to orchestrate it, identifying opportunities for improvement and anticipating issues well in advance of them becoming serious problems.

By applying both predictive and prescriptive analytics to large sets of operational and supply chain data, we can gain insights that were not previously possible and decisions can be made at a faster pace and higher level of accuracy than just a few years ago.

Increasingly, we’re also deploying cognitive computing to augment human supply chain managers and analysts. These systems are capable of interrogating vast pools of data and seeing patterns and trends that a supply chain manager could never do.

For instance, if there’s a strike in a port in Amsterdam and simultaneously a change in a key customer’s demands that relies on that supply route, from historical strike data a smart system can determine there will be a two day delay and that current inventory levels means the customer request cannot currently be met. The system can then suggest a solution that takes into account a variety of factors, for instance air freight options and current manufacturing timescales This process takes place within hours of the initial strike being announced and occurs in real-time, meaning that preparations can be made almost immediately and all knock-on effects calculated. The customer would never even know there had been an issue.

Through the application of such analytics, we’ve seen a 10-15% save on spend on inventory. Traditionally procurement only tended to focus on big ticket items, but analysis showed that there was not necessarily a correlation between these and where the greatest areas of risk or opportunities were. The beauty of analytics is that it is totally agnostic and not dependent on individual knowledge or informed guesswork, but concrete and actionable insight. We have been able to remove hundreds of millions in inventory based on demand and lead times, while decreasing supply chain risk from potential material shortages, non-conformance, poor quality and the like for customers by 10%.

So if the future of logistics is already here, what is next for the smart supply chain?

The ultimate goal for data capture and analysis is to have structured information flow from suppliers, manufacturers, retailers and partners, and unstructured data from various sources such as end user feedback about a particular product that’s been shared online, from forums, review websites, even social media platforms. Imagine a scenario where data not only allowed for analysis about demand patterns, but also component production. Combining structured with unstructured data would allow for the anticipation of features and elements that might be added or removed. Inventory management would go from the current ‘just in time’ model to one that actually anticipated the needs and requirements of manufacturers, reducing risk and overheads even further.

This is what we’re working towards. And, as history has taught us, the future is closer than many might think.

by Fred Hartung, Vice President, Supply Chain Solutions & Logistics at Jabil