Why the Shanghai Supply Chain Collapse Is an Opportunity for AI

Businesses around the world are bracing for yet another shock to the supply chain due to Covid lockdowns in China.

Since March, container dwell times have skyrocketed and cargo deliveries to and from Shanghai Port have been delayed or cancelled. The number of container ships waiting outside Chinese ports today is 195% higher than in February.

Shanghai’s port system handles about a fifth of China’s export containers. The volume of shipments to and from the port has fallen by no less than 85%. The bottleneck means businesses around the world are experiencing significant delays in the delivery of goods. Cargo waiting times at Shanghai’s maritime terminals have increased by nearly 75% since the start of the lockdowns. Delays at the Shanghai terminal have sent ships to neighboring ports in Ningbo and Yangshan, but they are also becoming overloaded.

The disruption will have a significant impact on global shipping schedules this summer and fall. Companies that rely on large freight volumes are under pressure to accelerate supply chain job bookings before congestion worsens in the coming weeks. Companies are also bracing for inflationary conditions due to product shortages at a time when inflation is on the rise in the US.

It is clear that disruptions such as the closure of the port of Shanghai will flare up again and again. Unfortunately, companies such as retailers and CPG companies are ill-equipped to handle disruptions on a global scale. Continued global supply chain disruptions, inflation and the emergence of COVID-19 variants continue to devastate essential functions such as demand forecasting.

This type of disruptive market does not seem to be going away anytime soon. It is therefore up to companies to effectively plan for these disruptions by combining artificial intelligence with third-party and first-party data to monitor rapidly changing conditions in real time and adapt processes such as demand forecasting.

Third-party data, such as weather forecasts and satellite maps of port traffic, give companies a real-time snapshot of conditions that could impact supply chain operations. For example, third-party data on shipping routes (available from aeronautical intelligence services) vividly illustrate the magnitude of the crisis in Shanghai:

Vasu 1.png

Vasu 2.pngFor example, third-party data gives a US retailer more detailed insight into the likely impact of how the congestion will slow cargo ships that take a few weeks to reach their destinations in US ports. From there, the retailer can more accurately estimate the impact on supply over a period of weeks and months and adjust its forecasts accordingly. Merchandisers can weigh the impact on costs and pricing strategies more effectively.

Even better, retailers can combine third-party shipping and weather data with consumer-generated data, such as Google search trends, to more closely match supply with demand at a regional level (always a moving target). They can weigh this information against their own first-party data on inventory levels and customer buying patterns. A supply chain crisis does not affect all regions of the US equally. A shortage of rain-resistant clothing will affect retailers in Seattle more in the summer than retailers in Phoenix.

No human can monitor, assimilate and analyze this data on a large scale. To do this, companies must adopt machine learning, a form of AI. Machine learning allows CPGs and retailers to browse third-party data and find patterns and associations that would go unnoticed by manual means.

Machine learning is especially adept at finding non-linear relationships that are crucial for demand forecasting, such as search behavior, where the intention to buy is not always clear. Even an automated platform would struggle to expose those non-linear associations without machine learning.

Machine learning and real-time data can be a powerful one-two punch together. Machine learning combined with real-time data from third parties and first parties can help businesses in a number of ways, including:

  • Prepare for the next disruption with effective scenario planning. CPGs and retailers can do what-if analysis with computer simulations. For example, they can analyze the likely impact of a port closure long before it happens, and be ready with corrective action. They can also run scenarios about the ripple effects of a disturbance. How could a product shortage, coupled with rising gasoline prices in a city, affect a planned promotion for a non-essential CPG product, compared to a commodity in rural areas versus cities? This kind of planning can be done with little investment.

Research has shown that by leveraging machine learning and third-party data such as search trends and real-time data to measure demand during the pandemic, CPG companies have reduced forecast error by more than a third, the volume exposed to extreme error in half and increased the realized value of investments in people, processes and technology related to planning sixfold.”

  • Get a better view in real timeReal-time data allows a company to identify the status of inventory throughout its supply chain. It can know exactly which trucks are no longer delivering goods to which locations in a crucial port. A retailer can find out which models of flat screen TVs are affected, how many, and for how long. With that level of visibility, they can more effectively tailor their in-store sales plans to major seasonal events. Companies need to know where their goods are at all times if they are to successfully detect and respond to changes in supply and demand. Machine learning and third-party data can provide that.

Given the global conflict, an ongoing pandemic, inflation and a gas shortage, we need to define a new ‘business-as-usual’ model. By combining AI with machine learning, we have a few tools that will help businesses achieve more predictable outcomes, no matter what market chaos throws our way.

Vasudevan Sundarababu is senior vice president and head of digital engineering at Pactera EDGE.

Leave a Comment