Proving the Business Case for the Internet of Things

DHL uses machine learning to detect supply chain risks

Steve Rogerson
June 6, 2017

DHL is adding machine learning to its supply chain risk management platform. The extension of the logistic company's early warning system uses machine learning and natural language processing to detect disruptions in a company's supply base before they cause financial losses or long lasting reputational damage.
Called DHL Supply Watch, it will become an integral part of its Resilience 360 supply chain risk management platform. With Supply Watch, Resilience 360 is adding a broad range of risk categories to the system's existing portfolio to monitor supplier risks on a company level, including financial indicators, mergers and acquisition, environmental damages, supply shortages, quality issues, and labour disputes, using publically available data found by monitoring online and social media sources.
"We provide our customers with a solution that detects and mitigates potential supplier failures before they happen, allowing them to focus on early risk mitigation and auditing activities of their most relevant suppliers and third parties," said Tobias Larsson, head of Resilience 360. "The insights and transparency customers gain through Supply Watch are another example of how digitalisation can benefit end-to-end supply chain operations, through building resilient supply chains and enabling businesses to be more competitive."
The addition monitors 140 risk categories including financial, environmental and social factors among risks resulting from crime, labour breaches, quality defects and supply chain perils such as shortages, capacity constraints and delays. Using advanced machine learning and natural language processing technologies, the adopted system analyses data based on the monitoring of up to 30 million posts from more than 300,000 online and social media sources to detect potential supply chain disruptions.
Instances such as the bankruptcy of one of the world's top ten container shipping lines, which led to capacity shortages and supply chain disruptions worldwide, was for many businesses unexpected, despite indications that were made public before. Supply Watch closes this gap and flags potential risks at an early stage.
The recent global WannaCry ransomware attack is another example for a situation in which the system could help identify which suppliers may have reportedly been affected, and therefore allowing companies working with them to take appropriate precautions in their supply chain.
The development of the intelligent Supply Watch system was supported by linguistics experts and data scientists to establish a reliable analysis of content and context of online discussions. Supply Watch is capable of understanding human language and evaluates how people talk about risk-relevant events and situations around the world. Among different risk types that may affect supply chains, many are particularly hard to detect, such as quality issues.
Monitoring and analysing discussions and articles in online and social media about such concerns, information about recalls, protest or delays, helps identify early indicators of supplier and partner distress. This feature distinguishes the system significantly from conventional search approaches as it operates nearly in real time. Users can address emerging issues quickly, preventing reputational and financial losses.
Supply Watch is available independently but can be fully integrated into Resilience 360.