Eric Buatois
This article was recently published by INSEAD. To view the article, please click here.
As the Coronavirus wreaks economic turmoil around the world, our modern supply chains are facing unprecedented stress. For months prior to the COVID-19 crisis, trade tensions have been mounting due to the escalating tariff war between Washington and Beijing, and a broader populist streak running through several other capitals. This rise in protectionism, coupled with concrete costs and new financial barriers, has fueled broader challenges and concerns for logistics networks operating on the global level. Against this backdrop, our modern supply chain infrastructure is well overdue for a significant rethink.
Today’s globalized supply chain network has been optimized to identify minimum lead times at the lowest possible costs. We want our electronics to be made in China, so we can buy them on the cheap. However, rapid political developments, extreme climate events and, now, global pandemics have all revealed the hidden costs of single source dependencies and poor flexibility in adapting to real time shocks, with fast changes to supply and demand. Today, we might be willing to tolerate higher prices for certain goods, if it means we can get them delivered faster and more reliably. As a result, we can anticipate a shift to multi-level sourcing and an increased focus on flexibility. Over the next several years, as we undertake a broader overhaul of our logistics infrastructure, I believe, a new order will emerge based on three key dimensions:
From Globalization to Regionalization
Similar to the 1980s, logistics hubs will re-emerge at the regional level. To eliminate single source dependencies, and to establish a flexible, and adaptable, supply chain, product integrators, sub system suppliers and component suppliers will all have to source, assemble and deliver from their own backyards.
A shift to multi-level sourcing, on the regional level, with new suppliers, is no easy feat, however. A mid-size car company, for example, may have three or more brands, each with 200,000 employees, and can have parts sourced from 6,000 suppliers. If each car is made up of 4,000 components, then delivering these parts, in real time, generates a very real optimization challenge. Today, large electronic equipment manufacturers are sourcing about 40% of their part from China including sub assembly. Given the incredibly high number of parts required — each with different lead times — a return to multi-level sourcing generates an incredibly complex optimization challenge.
That being said, this shift is already underway in the medical industry where equipment manufacturers and pharmaceutical supply chains are moving in high gear to respond to government orders. Dyson, the British technology company best known for its vacuum cleaners, hair dryers and fans, received an order from the UK government to produce 10,000 ventilators for COVID-19 patients. Although it awaits regulatory approval, the company developed its design in just 10 days based on its existing motor technology, and has committed to developing 15,000 ventilators, so it can donate the surplus to the international effort.
In the US, General Motors, Ford and Tesla have stepped in to address the ventilator shortage. Last Friday, the White House employed the Defense Production Act (DPA) to order GM to put production lines at the hands of the Health and Human Services Department and prioritize contracts from the Federal government. As reports suggest that New York may reach its anticipated “apex” of coronavirus care patients in the coming weeks, the state is building stockpiles of ventilators, N95 masks and personal protective equipment for hospital workers. But as Detroit, New Orleans, and another set of cities each approach their own apex, the COVID-19 pandemic assures that governments, at several levels, will all want a regional supply of medical equipment and ventilators, as well as specific antiviral drugs. Today, the pharmaceutical industry in Europe imports 80% of the active components for its drug supply from China and India. It’s entirely predictable, and expected, that these governments want to ensure that they can draw on these supplies from their own region. In addition, no single component of the supply chain can be single source, or subject to vulnerability.
Beyond the medical field, the electronics supply chain is quickly following suit. The overnight shift to a global “work from home” work force has sharply increased demand for laptops and notebooks. As businesses struggle to outfit their employees with the tools they need to be productive, the inability to secure laptops is creating a major revenue loss. Under the circumstances, companies may well be willing to pay a 10% or 20% premium to have laptops delivered to their workforce today, rather than wait until back orders are filled due to stalls at production plants in China.
The Supply Chain Stress Test — A New Norm
After the 2008 financial meltdown, regulated financial institutions around the world have been forced to stress test their balance sheets to assure preparedness for an economic shock. Governments wouldn’t have it any other way. Similarly, a series of large scale cyber-attacks in the past 10 years have forced technology companies to institute penetration tests to identify any vulnerabilities and to check the robustness of their cyber-security mechanisms. Boards of directors won’t have it any other way. In a post-COVID-19 world, supply chain stress tests will become a new norm. The distributed global business model, optimized for minimum lead time delivery, is finished. Tomorrow’s model demands new priorities in optimization.
Large suppliers and logistics operators in the supply chain industry will have to prepare themselves for major catastrophic events such as weather events (fires, flood, tsunami), lethal pandemic outbreaks, strikes, social unrest and associated disruptions. In many respects, we’re still likely at the beginning of the Coronavirus crisis, which certain models predict may last another 12-18 months. International commercial air travel has effectively ground to a halt, and one can easily imagine harbor blockades, shipping disruptions and a further tightening if death tolls mount in certain hot zones. Independent of the COVID-19 crisis, more and more externalities are causing shocks to the supply chain that require operators to prepare to operate in crisis environments.
To maneuver choppy waters, navigators need visibility. In certain industries, like microprocessor development or consumer technologies, advanced electronics manufacturers have already produced comprehensive dashboards that lay out the full status of production and shipment, down to the last detail. These control centers are similar to the network operation center of communications service providers or electricity utility companies. They can map, for example, the delivery of microprocessors from the Intel manufacturing plants in Israel to the assembly lines in China, then back to the distribution centers in Europe, where they are sold to retail outlets. Yet, these dashboards are refreshed every 20 minutes to provide full visibility, in real time, on the entire supply chain. Dashboards like these, once reserved for a few industries, will soon emerge across many other industries as a norm. In the pharmaceutical industry, for example, there is no single database, either centralized or distributed, from which to map the critical components for drug manufacturing. To the detriment of suppliers, and ultimately, end-users, the visibility on sourcing is a prime imperative.
Simulation tools will also be required for scenario mapping. Given the complexity, crisis forecasting, and scenario modeling, is crucial. The amount of hops required for the production of consumer electronics, for example, is truly staggering. Even the slightest disruptions, at any step, can create compounding challenges to assembly lines. Software simulators, therefore, must be able to identify shortages of a component, a strike at an airport, a fire, flood or natural disaster, or a political blockade on a crucial shipment port. By identifying these data-points, placing them against a knowledge graph, and feeding them real time updates pertaining to all of these externalities, simulation tools can identify risks relating to each scenario. Moreover, predictive tools can introduce a prescriptive element to mitigate disruptions. Ultimately, supply chain stress tests will help us to better manage risks, and find load rebalancing options.
The Human Dimension Is Back
As unemployment rates rise across the US, and around the world, certain key industries actually face labor shortages. Businesses operating in the healthcare industry, in agriculture, in grocery stores, and other key “essential work” fields are ramping up their hiring. Amazon announced 100,000 new roles in fulfillment centers and delivery networks to support the surge in service demand during the pandemic. Meanwhile, in China, the impending return of quarantined workers to the production plants and factories has generated relief in the West, as those production lines can return to efficient operational levels. The human dimension is back, and it will play a prime role in rebalancing the global supply chain, during this crisis, and well beyond.
The very same humans that were doomed to be replaced by artificial intelligence and automation have now become the most critical asset to assure ongoing service and supply chain efficiency. When Elon Musk opened the first fully automated Tesla car factory in Silicon Valley — fully equipped with a robotic assembly line, computer vision and machine learning manufacturing — things didn’t go as planned. The hyper-automated robots couldn’t deal with unexpected orientations of certain objects, like nuts and bolts, or maneuvering in the car frame. To meet the targets, Musk famously spent nights sleeping on the factory floor, and he had to quickly retain assembly workers on site to optimize the workflow and streamline production processes.
Certainly, over time, improvements in AI and machine learning will automate more and more of these functions. However, what’s limiting e-commerce traffic today is a shortage of commercial drivers and shipment mechanisms. Logistics distribution centers may automate with increasing speed and frequency, but they will still require more and more drivers, drone flight operators or operators of autonomous truck operations centers. The “last mile” to deliver the final product from a growing number of distribution centers to the customers, is critical, and remains human.
Rather than viewing labor as a commodity, it should be managed as a key asset of flexibility and adaptation that plays a fundamental part of a crisis response program. The shortage of N95 masks, for example, requires producers to open new manufacturing plants to keep up with demand. Given the dramatic price increases, at up to 5x markups in online retail outlets, increasing producer’s labor costs is entirely feasible and manageable. The opening of a new production line, or the moving of an existing one, requires a surge in human capital. To put it in perspective, each smartphone touches 80 human beings during the manufacturing process. So moving an assembly line from China to Mexico, for example, requires training 80 individuals to produce the same quality output as the original operation. A post-COVID-19 view of the workforce will have to accommodate this outlook.
The Investment Opportunity
These three trends demand a brand-new generation of software systems. To bring these solutions, and norms, into being, entrepreneurs in AI, mathematical modeling, machine learning algorithms and cloud services will have to find ways to extract critical data from the enterprise supply chain and correlate it with external “major events” data to deliver real time business insights. These insights must be directly relevant and actionable to workers on the factory floor. At the same time, management must be able to use them to identify and consolidate high level trends.
The extended logistics ecosystem, including delivery, flow, capacity, and transportation must all be integrated. On a granular level, this can mean optimizing a dynamic blue-collar workforce for staffing and management needs. It means tracing each component, and mapping the work of every line operator, in real time, that is assembling that component — not to replace the worker with robots, but to assist them, by identifying the specific gestures and tasks that assure high quality labor norms. Integrated learning solutions are key when reviewing manufacturing issues that may have led to a product recall. It’s also helpful for businesses that need to transfer a plant site from China to Mexico, for example, and retrain a new assembly line of workers. The sheer complexity of these tasks demand the creation and implementation of new algorithms to deliver predictable, trusted and repeatable results.
To deliver insights in days, hours or minutes, rather than quarters, these solutions will have to be integrated with, and embedded in, existing workflow processes. As such, they need to be industry-specific, leveraging deep domain knowledge, and customized to specific verticals. At BGV, we’ve been at the forefront of the Enterprise 4.0 space, and are very bullish on the future of this space. The investment thesis we’ve been working on, and the trends we’ve identified, are now occurring much faster than we’d imagined. If anything, the COVID-19 pandemic has accelerated the oncoming wave, and this future will emerge sooner than we’d imagined.
The enterprise solutions for the supply chain space have been reliant on ERP, and an infrastructure that was designed over 25 years ago, with an eye towards frictionless globalization and increased automation. The future, however, must be optimized for a regional level of integration, with multi-level sourcing, flexible work forces, and shock-resilient supply chains. While this new reality has been thrust upon the healthcare field, the trends have been readily visible for some time. And in the coming years, it will touch all industries, whether we anticipate the change, or not.
