How tech can build more resilient supply chains

Over the past two years, the global supply chain has been hit with two major upheavals: the United States-China trade war and, more cataclysmically, COVID-19.

When Reefknot Investments launched its $50 million fund for logistics and supply chain startups last September, the industry was already dealing with the effects of the tariff war, says managing director Marc Dragon. Then a few months later, the COVID-19 crisis began in China before spreading to the rest of the world, disrupting the supply chain on an unprecedented scale.

Almost all industries have been impacted, from food, consumer goods and medical supplies to hardware.

Reefknot, a joint venture between Temasek, an investment firm headquartered in Singapore, and global logistics company Kuehne + Nagel, focuses on early-stage tech companies solving some of the supply chain’s most pressing issues, including risk forecasting, financing and tracking goods around the world.

In March, around the time the World Health Organization declared the COVID-19 crisis a pandemic, Reefknot surveyed nine shippers about the challenges they face. While there are other macroeconomic factors at play, including Brexit and the oil price war, the survey’s main focus was on the combined effect of COVID-19 and the U.S.-China trade war on the supply chain and logistics industry.

According to the study, the main things shippers want is the ability to dynamically manage supply chain risks and operations and optimize cash flow between corporate buyers and their suppliers, who often struggle with working capital.

Many of the current solutions used in the supply chain involve a lot of manual tasks, including spreadsheets to predict demand, phone calls to confirm capacity on planes and ships and checking goods to make sure orders were fulfilled properly.

“The big topic was around digitization, because when you talk about cash flow, risk management and dynamic operations, it all alludes to having visibility. Having visibility of your stuff, what capacities are available, what is your demand, how long will it take to fulfill an order,” says Dragon. “Supply chain visibility and data and application integration have become very foundational and central to this, and we think that is one of the things that will continue to stay beyond this crisis.”

“The other thing that will stay is the sudden realization that cash flow is very important, not only for buyers themselves, but especially for SMEs who are the suppliers,” Dragon adds. “Besides the physical and data portion of it, the financial part of the supply chain needs to be just as resilient, and right now it is not.”

An unprecedented crisis

Reefknot helps run a think tank called Next Supply Chain (NEXST) that looks at emerging supply chain technology and trends. One of its co-founders is Wolfgang Lehmacher, the former head of supply chain and transport industries at the World Economic Forum.

The unpredictability caused by the pandemic “is a phenomena we haven’t seen before,” Lehmacher says. When Chinese cities were reopening, businesses in Europe and the U.S. were closing. Some factory furloughs have become permanent closures. Cargo planes have been hit with massive flight cancellations, disrupting shipments because passenger travel accounted for 40% of global air capacity. On top of this, some countries have implemented on-and-off lockdowns in response to new waves of the coronavirus, and e-commerce companies are coping with a massive surge in demand as people stay home.

“An important part of the story to understand is that from the beginning, it was a simultaneous supply and demand shock,” Lehmacher says. Supply chains are usually able to withstand disruptions because most of them are vertically oriented (for example, food, furniture and medical equipment supply chains are all separate from one another), but the pandemic has affected almost all industries.

Helping suppliers survive

One of the biggest vulnerabilities in the supply chain is the financial health of suppliers, who are typically small- to medium-sized businesses. One of Reefknot’s portfolio company is Previse, which enables corporate buyers to make instant payments to their suppliers, instead of the usual lengthy invoicing cycles. Co-founder and chief executive officer Paul Christensen says Previse has interviewed thousands of suppliers, who report that their number-one issue is cash flow.

The average payment cycle is about 63 days, and many also have difficulty accessing financing from banks, limiting their ability to increase production. “It’s a fundamental macroeconomic inefficiency,” Christensen says.

He adds that the industry needs to start thinking about what will happen once the world starts to emerge from the COVID-19 crisis. “The issue right now is that businesses are shut down, but when demand surges, particularly for small businesses, cash flow is going to be just as big a problem, if not bigger.”

This is critical because the fulfillment cycle varies widely between different industries and often stretches over months. For example, fashion retailers begin putting in orders for seasonal clothing several months before they appear in stores. For suppliers who are currently weathering the impact of the pandemic, this means they may not have the cash available to fulfill orders when they come in.

Previse, whose investors also include MasterCard and Bessemer Venture Partners, offers a tool called InstantPay. To enable quick payments while protecting buyers, Previse’s solution looks at all the data in their enterprise resource planning (ERP) systems and predicts future interactions. This means Previse can green-light suppliers who usually fulfill orders without any issues for immediate payment. The machine-learning tech also tags high-risk suppliers who need to go through a validation process before they are paid.

But Christensen says less than 2% of invoices processed are flagged for manual review because many corporate buyers establish long-term relationships with suppliers, who usually provide the same type of goods in each order.

“It’s such a small number of invoices that are likely to cause problems, so why is everyone waiting months going through manual processes to do the validation? It should be done on an exception basis,” he says.

Some buyers set long invoicing terms to protect their own working capital, but the reason for slow payments is often simply because corporations handle huge numbers of invoices, and their accounting processes are time-consuming and often paper-based.

Before the pandemic, it was challenging to convince corporate buyers to implement InstantPay, Christensen says. Around April, however, corporations became more willing to adopt the tool. “The resilience of supply chains and SME cash flow are now C-suite level topics, so we’ve seen a huge change in people’s willingness to engage with us and prioritize deploying our solution.”

Christensen says that the economic crisis caused by the pandemic makes it more critical than ever for corporate buyers to pay attention to their suppliers’ cash flow, especially with the increase in SMEs going out of business. “That raises questions like ‘Is my supplier still going to be there or not,’ and it also makes it hard for corporate buyers to find new suppliers, because they need cash immediately to stay in business.”

Forecasting amid uncertainty

Prowler, another company in Reefknot’s portfolio, offers a product called the Decision Engine (formerly called VUKU) used by clients in finance and logistics to create prediction models.

“What makes supply chains unique from everything else is just the scale of it,” says co-founder and CEO Vishal Chatrath. For example, raw materials might be mined in Africa, then refined in China before being shipped to a semiconductor fab. From there, chips are sent to electronics factories, and finished hardware is then shipped again to warehouses for distribution to retailers.

While factories can be equipped with cameras and sensors, visibility in the rest of the supply chain is very low. Chatrath believes the Internet of Things is not a viable solution for tracking shipments, because of the huge amount of devices that would be required. Prowler’s objective is to provide more visibility into the supply chain without using sensors, while also requiring very little data.

The Decision Engine launched last year and many of its supply chain clients began implementing the system in January, just as COVID-19 was beginning to disrupt life in China. This presented a unique challenge for Prowler, whose investors also include Tencent and SGInnovate, to show how the Decision Engine can make predictions with sparse data.

The foundation of the Decision Engine is the Gaussian process, named after German mathematician Carl Friedrich Gauss, who died in 1855. Chatrath explains that Gaussian process models are unique because they can work with very little data. Though the Gaussian process has been around for almost two centuries, they were previously used mainly by academics for “toy problems.” Prowler applies it to real world problems (its team includes chief scientist Carl Edward Rasmussen, who wrote one of the leading textbooks on Gaussian processes in machine learning).

“Forecasting is done by taking data from the past if similar things are expected to happen in the future. Then you forecast based on that with some form of probability,” Chatrath says. “Of course, all these companies had data from the past, but a pandemic on this scale had never happened before.”

To make predictions during disruptions, trade managers rely on data from past black swan events, including financial crises, terrorist attacks and policies like Brexit or tariff wars that impact global trade. The absence of data for a pandemic like COVID-19 meant that Prowler’s Decision Engine had to generate synthetic data to model various scenarios.

“We were able to give [clients] a tool that was able to do ‘what if’ scenarios. As the situation changed, week by week or even on a daily basis, as new data points were coming in, they were able to react very quickly and make decisions with those trade offs,” says Chatrath.

Examples of how Prowler’s clients used the Decision Engine include fast-moving consumer goods (FMCG) companies forecasting demand for items like hygiene products, or deciding how much capacity to book on cargo vessels for international shipments. Despite the chaos caused by the pandemic, Chatrath says some of Prowler’s clients saw a 25% improvement in supply chain efficiency after they started using the Decision Engine.

Making change permanent

Because the supply chain and logistics industry is so vast, fragmented and dependent on manual tasks for operations, buyers and suppliers are often slow to adopt new technologies. When asked if the COVID-19 crisis will make them more open to implementing new tech, Dragon said, “It’s a good question, and the answer is a mixed bag.”

The “double whammy” of the U.S.-China trade war and COVID-19 has increased interest in big data and AI solutions, and many corporate buyers have been more willing to spend money on new tech. But Dragon added it’s still unclear if this openness will continue long term.

When Prowler was launched, Chatrath says “one of the things we said would kill a lot of companies is that their supply chains are more fragile than they think. In 2018, 2019, nobody wanted to hear that.”

More companies have now started to move away from focusing solely on supply chain efficiency, however, and begun to think about resiliency. Chatrath believes that over the next few years, every Fortune 500 company, and even governments, will have a “chief resilience officer,” or someone whose role is to quantify and give demonstrable evidence that supply chains are resilient.

Lehmacher says that there needs to be cooperation among countries to develop global standards so planes and cargo vessels can continue international shipments without disruption even during major crises.

“There needs to be more communication between suppliers and companies, so they better understand their raw material suppliers’ production capacity and financial situation. That means pooling together data and giving a more holistic picture, for example, if there is a strike or another type of disruption,” he adds.

Chatrath echoes that sentiment. He wants to see international standards developed for the supply chain and logistics industry, similar to the ISO 9000 family, which outlines quality management systems for manufacturers and other organizations, or the ISO 27000 series on best practices for information security management. Many other industries are also subject to international standards.

The supply chain and logistics industry should have a similar certification process, Chatrath says, with a common international framework that requires organizations to model black swan events and show how their supply chain would hold up.

“In aircraft manufacturing, there is certification, so you just can’t say ‘my aircraft is safe and allowed to fly.’ Similarly, if you are a bank, you can’t just say ‘my bank is secure and it’s allowed to trade,'” Chatrath says. “In supply chains, somehow people are allowed to say ‘my supply chain is resilient’ without any proof or need to demonstrate. I think what we need, at the very least, is an ISO certification for supply chain resilience.”

This article has been updated with a more accurate description of Temasek.