December 16, 2021
Predictive analytics, data standards and visibility software could be critical to avoiding supply-chain logjams, members of the Braintrust say.
Good afternoon! Global supply chains have been put through the wringer over the course of the last couple of years, with an increasingly tricky situation unfolding over the course of the past couple months. While much of the focus is on alleviating logjams currently happening, we asked the experts to look a little further into the future and tell us about the tech that could ultimately help avoid supply chain crises in the years to come. Questions or comments? Send us a note at firstname.lastname@example.org
Dr. Michael Feindt
Co-founder and Strategic Adviser at Blue Yonder
Supply chains are usually optimized to achieve the highest profitability and efficiency, lowest cost and just-in-time delivery. When the world behaves as expected, uncertainties/risks are mostly ignored, and this delivers good profits.
Complex production and supply-chain planning are handled by linear programming optimization. Plans are defined by countless parameters, thus as one point in a very high-dimensional space. The optimum is found by minimizing a (linear) objective function of many parameters given many constraints. It always lies at an outer edge of the high-dimensional subspace allowed by the constraints. Such a solution is never robust. Small parameter changes can cause large plan changes.
Advanced technology considers uncertainties, assumed relatively harmless and normally distributed. This already is much more complicated and needs more input data for the uncertainties.
Black swan events like COVID-19 are rare but massively impact parameters. Afterwards, everybody calls for resilience, robustness and regionality, despite cost. This requires robust planning algorithms performing well even if parameters are very different from normal.
Future planning will handle uncertainties everywhere, even allowing non-normal extreme cases. Note that any such modeling always will bear some subjectivity. Being non-linear, probably the technology will apply to a Monte Carlo simulation. But one thing is clear: From a certain point on, more robustness means less efficiency in the normal case. There always will be a risk aversion parameter to be tuned, and my experience is that humans forget about past catastrophes quite fast.
CIO at XPO Logistics
Technology and innovation are the way of the future in supply-chain transportation: being able to automate how transactions happen between shippers and carriers and having access to massive capacity. We’re seeing an increase in shippers outsourcing freight transportation to brokers, particularly those with digital capabilities that can provide more reliable capacity. Digital freight brokerages enable customers to plan, manage and analyze their shipments and business in real time using machine learning, predictive analytics and location tracking.
An industry that has typically conducted business through calls, emails and texts will need to evolve to one where business processes — making bids and offers, tracking shipments, billing and monitoring customer service — are automated. XPO Connect, our digital freight marketplace, leverages proprietary algorithms to turn massive amounts of data into relevant information in split seconds so that our customers can buy, sell and manage transportation efficiently and automatically.
As more carriers and shippers continue to automate their processes, more data can be gathered. Analyzing data from a broad range of sources will also increase the ability to predict with a higher degree of accuracy. Predictive analytics will need to go beyond historical and transactional data across the supply chain and enhance the algorithms by adding external factors such as weather and even geopolitical issues such as tariffs that can potentially impact the supply chain.
Director of Policy at The Coalition for Reimagined Mobility
With supply-chain disruptions taking over the headlines, the fragility of the system has become the fodder of cocktail parties. Yet, stop-gap solutions designed to deliver holiday gifts in time will not build critically needed resiliency into the supply chain. To ensure operational visibility and create resiliency, we need to invest in a digital freight system, powered by broadly adopted open-data standards.
Today, digital solutions are unevenly distributed across the freight value chain. While some large players have deployed digital solutions to facilitate operations, smaller stakeholders typically still rely on manual processes. Not to mention, the digital solutions used by larger players are often built on proprietary standards that are difficult to scale.
The result is fragmented communications across the freight value chain, leaving efficiencies and system flexibility on the table. Without leveraging open data standards to enable scalable communications across the value chain, our efforts to mitigate crises and promote resiliency will be thwarted.
National governments around the world are standing at the ready to alleviate the acute supply-chain issues of today, and we can build the digital infrastructure needed for long-lasting solutions. Through investing in open data standards to enable digital freight, we can seek to ensure that gifts for future holiday seasons are wrapped and ready with time to spare.
Global GM, New IoT Markets, Smart Cities and Intelligent Transportation for IoT Solutions at Intel
Technology that provides end-to-end visibility into supply chain, including real-time insights on delays at the edge, will be crucial for managing similar crises. The current supply-chain crisis isn’t just about using the right technology — at the end of the day, it requires deeper coordination across cities and transportation assets, such as fleets, railways, marine and airports. For example, transporting a piece of cargo is dependent on tracking multiple types of assets to ensure optimal performance, including the driver and vehicle. Most fleet owners have a fragmented fleet, with trucks and vehicles from different manufacturers. Some of them may be older or newer, while some may be diesel-based rather than gas or electric. This presents an enormous challenge. With so many types of data being collected in different ways, how do you create a single window pane into fleet performance while taking into account issues like driver safety? This is where the right networked technology tools at the edge can help — tracking real-time insights on driver behavior, theft prevention and predictive vehicle maintenance — to provide visibility into the origin of issues so they can be addressed before larger supply-chain disruptions occur.
To be clear, there will never be a magic technology bullet to solve all issues. If we look to smart cities for key learnings, we’ll find that identifying the right technology is just as important as the ability to cultivate ecosystems of partners who can work collaboratively together on solving or preventing the next big crisis.
VP, Solution Management for Digital Supply Chain at SAP
The pandemic has exposed the risk in our global supply chains, and for the past 20 months, the word everybody is using to describe what is needed is "resiliency."
To reduce risk and increase resiliency, companies have been looking to re-balance their strategies for on-shore, near-shore and off-shore manufacturing, and improve collaboration across the business network to identify alternate sourcing strategies of key resources and materials. They are also looking for ways to leverage integrated business planning and logistics processes to optimize their inventory across the supply chain. This starts by improving visibility across the supply chain, both within your organization and your network of suppliers, contract manufacturers, logistics service providers and other partners.
The adoption of technologies such as IoT, big data analytics, AI, machine learning, blockchain and cloud is accelerating initiatives in real-time visibility, transparency, risk monitoring and security. This helps to ensure organizations can anticipate and manage challenges, and continuously optimize their supply chain to successfully navigate any economic conditions.
With Industry 4.0 technologies providing a roadmap for the future, companies are beginning to see the full potential of a digital supply chain. To make this a reality, knowing when and how to use and combine these technologies to solve specific business problems is key. IoT, for example, generates huge amounts of data. Predictive analytics and machine learning process this data to make it smarter. And by leveraging artificial intelligence, the data can be used to automate processes, with blockchain ensuring the processes remain traceable and secure.
Chief Technology Officer at FourKites
Deep supply-chain data has recently become widely available and enabled technology that improves operations. With a data network of shippers, carriers and 3PLs on platforms like FourKites, companies easily answer the question, “Where are my goods and materials?” But this is only the beginning of how tech can solve supply-chain problems.
In addition to saving time-tracking shipments, there is an enormous opportunity to move from descriptive solutions — “Your shipment left LA” — to predictive solutions — “Your shipment left LA and will arrive three hours late” — and, ultimately, to prescriptive solutions — “Your shipment is late but we found an alternative lane that will save time.”
To realize the full potential of technology, the supply-chain industry must do three things.
First, we must improve our data standards and continue creating interconnected networks. This is critical because highly accurate real-time ETAs and prescriptive solutions depend on large, quality datasets. For example, FourKites uses artificial intelligence and machine learning to provide accurate ETAs on 97% of loads that previously couldn’t be tracked.
Quality data will continue to improve models, which can be made even better through continued collaboration using appropriate security controls and anonymization. With technology's help, together we can identify overabundant supply or demand and proactively prevent problems before they occur.
From there, we can test the waters and begin placing our trust in AI and machine-learning models — we can move away from triaging crises to letting the machines flag any potential issues and provide potential solutions.
EVP and Global Head of Manufacturing at Infosys
If there’s one thing that the current crisis has taught us, it is the need for an adaptive, responsive, high-visibility supply chain. This is a concept that we like to call the Live Supply Chain which — like a living organism — is able to sense, feel and respond in real time. The technology solution that we have envisaged has a layer each to cater to the three imperatives.
Sensing is all about capturing supply-chain events across geographies, logistics, warehouses and factories. Examples include adverse weather, derailments, supplier bankruptcies and labor actions. From a technology perspective, this requires the capability to collate and organize data from both internal and external sources.
Feeling is about understanding the impact of the event. This impact needs to be derived by connecting the dots digitally between the disruptive event and its implication on orders, inventory positions and downstream manufacturing processes.
Responding is where the solution needs to arrive at next best action to overcome the impact of the adverse supply-chain event. The response could range from an alternate mode of transportation to diverting the order to a different supplier. The response needs to be determined based on criteria like capacity, availability, on-time delivery commitments, costs and risks. All of this requires artificial intelligence and machine-learning solutions that take in real-time supply-chain data stream as well as historical information on costs, lead times and reliability, to then recommend next best actions.
President and COO at ResMed
To minimize or prevent future logjams across the global supply chain, I’d advise leveraging the distinct talents of today’s emerging technologies: gathering data and learning from it.
At ResMed, we mine data from over 10 billion nights of clinical sleep and respiratory data to learn how various populations use and thrive on cloud-connected medical devices and digital apps — all to help millions more people in these populations get the treatment they need.
Similarly, companies can deploy machine learning and advanced analytics to track demand and production of critical components that go into all the products we rely on — and forecast how many might be needed next quarter, month or day. And when something as unforeseen and daunting to model as a pandemic occurs, people can collaborate on updating these forecasts to show how many semiconductor chips device makers will need, how many circuit boards the chip makers will need, the wafers board makers will require, etc.
Freight demands also can be forecasted this way. Today, ResMed models risks in the variability of shipping via air versus sea to determine the most cost- and time-efficient freight mix with which to deliver finished goods to our customers.
One non-technical element remains key: people. Relationships between colleagues, companies and public-private partnerships are what it takes to put ML-generated recommendations into action. We’ve seen it during the initial global response to COVID-19 and the current global chip shortage. No matter the power of our technology, we’ll always be as successful in times of ease and crisis as the relationships we forge.
CEO of Symbia Logistics
See who's who in the Protocol Braintrust and browse every previous edition by category here (Updated Dec. 16, 2021).
There is a lot of opportunity in the future state of supply-chain technology. The industry is very segregated and still depends upon many pen-to-paper strategies. Because of the multitude of players in the supply-chain space, and the many different platforms being used, consolidation of these will continue to shape the future. Technologies that create real-time visibility and transparency will become more critical and the accuracy of this information will play a major role in forecasting for the future. An accurate forecast will ultimately set up the success of any future supply chain, and the analytical technology connecting the forecast to the operations will be critical to success.
Kevin McAllister ( @k__mcallister) is a Research Editor at Protocol, leading the development of Braintrust. Prior to joining the team, he was a rankings data reporter at The Wall Street Journal, where he oversaw structured data projects for the Journal's strategy team.
More from Braintrust