We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!


You can’t plan for what you can’t predict. A huge container ship blocks the Suez Canal, disrupting global trade for several days. A global pandemic wreaks havoc on global supply chains, causing shortages and pushing prices sky-high.

Now, Russia’s invasion of Ukraine has injected chaos into an already fragile supply chain ecosystem. The availability of everything from oil and natural gas to wheat is now in doubt as car manufacturers stop production in Russian factories. Automakers, in fact, are facing their third supply chain crisis in as many years.

Old-fashioned supply chain planning involved charting the journey of a material or product from the raw material stage to the consumer. It also encompassed supply planning, demand planning, production planning, operations, inventory optimization, routing, transportation, logistics, warehouses and more.

But what happens when there’s a weak link — or a complete break — in the chain? Something as minor as a truck breaking down or as major as a global pandemic introduces uncertainty into our planning algorithms. These types of supply chain disruptions are inevitable. While we can’t control for all the variables or predict the unpredictable, we can be better prepared to respond.

Event

Transform 2022

Join us at the leading event on applied AI for enterprise business and technology decision makers in-person July 19 and virtually from July 20-28.

Register Here

Supply chain snags and the bullwhip effect

Let’s take a closer look at how the delicate balance of a supply chain’s complex system can be disrupted at one point, causing chaos further down the line. Today’s mass production environment favors just-in-time manufacturing — an environment that encourages receiving goods only as needed for production. This, ideally, reduces inventory costs and waste.

Meanwhile, factories are tuned to work at full capacity. Considering how expensive it is to build and operate a factory (employees, robots, electricity), the last thing you want to do is stop the production line. In a stable world, with no surprise variables, you can calculate the inventory you need to keep the factory running smoothly. Factor in a few extra “safety stock” inventory units (just in case) and 99% of the time, all is well.

What happens when there’s a supplier snag upstream? Or a consumer change of heart downstream? Exercise bike manufacturer Peloton faced this exact scenario recently. After an uptick in consumer bike purchases at the beginning of the pandemic, consumer demand began to cool over time. The result: thousands of cycles and treadmills sitting jam-packed in warehouses or on cargo ships. Peloton temporarily halted production of its connected fitness products earlier this year, while the company laid off staff and overhauled its management team.

Peloton fell victim to the “bullwhip effect,” a supply chain situation that results from small fluctuations in demand at the retail level causing progressively larger fluctuations in demand at the wholesale, distributor and supplier levels. The phenomenon is named after the physics involved in cracking a whip: a slight flick of the wrist results in increasingly larger motions toward the end of the whip.

A small change upstream, a major change downstream

The Bullwhip Effect describes changing consumer demand patterns, but what happens when there are disruptions upstream? People tend to overreact when there is a small fluctuation at one end of the supply chain, perhaps triggered by a major event (such as Russia’s invasion of Ukraine). You may have 1,000 different parts feeding into your factory. What happens when a part you need is out of stock from the supplier? All the steps you thought you could carefully manage with a small inventory suddenly seize up.

Consider the global chip shortage. In the early stages of the pandemic, early signs of changing demand patterns led to stockpiling and advance ordering of chips by some, which left other companies struggling to obtain needed components. Auto manufacturers cut their orders for semiconductor chips as they predicted demand for new cars would take a nosedive. Those chips were then snatched up by other industries for phones, computers and video games. Meanwhile, worldwide auto production was halted when a missing single part stalled production of the entire vehicle.

It’s not about planning, it’s about responding: Graphing “what if” scenarios

Many claim that poor forecasting and ineffective planning result in these supply chain disruptions. The problem isn’t a failure to plan, it’s a failure to effectively respond. How can you forecast numbers for six months from now when you have no idea what will be happening six months from now? It’s not like you build forecasting engines for the next pandemic. Instead, you should set a baseline number of units for inventory and focus on looking for demand signals in the market — and responding to these signals in a sensible way. Adapting to change is key. All CEOs should ask themselves, “What is our ability to adapt to unforeseeable big changes?”

The key to adapting to change is having data systems in place that showcase your options, as well as quantify the implications of any given option. And you need to do this as quickly within the supply chain as possible. Once that signal goes downstream, it gets more difficult to recalibrate throughout the supply chain (and the result may be overstocked items or an inventory shortage).

Traditional supply chain software is linear, passive and limited by relational databases. Relational databases, which store customer, order and product data in separate tables, were designed for steady data retention rather than dynamic data-intensive use cases.

A graph database, however, can model disparate relationships and dependencies in a way that closely mirrors the real world. Graph, which tracks every individual part from supplier through the manufacturer to the finished product, can load massive amounts of data and uncover real-time relationship patterns. The graph provides a “what-if” engine, allowing companies to create a digital representation of a complex system (such as an automotive supply chain). The graph represents a “digital twin” of your real-life supply chain, allowing you to evaluate alternative plans in response to global changes in supply and demand.

Graph algorithms, which include the shortest path and geographical proximity, can help you manage and mitigate complex dependencies — in real-time. As several internal and external factors (involving parts, people and things) cannot be forecasted, businesses must be ready to respond.

What is the end-to-end impact of a change in supply? If a part is unavailable, what product can you build now with what you do have? Graph empowers you to take an active role in managing your response to demand changes, meaning you move away from a passive view of risk. If demand for a particular car model is suddenly dropping in the U.S. market, what parts will we now have in surplus? How can we best use these parts? What other options do I have?

Graph analytics helps you answer the difficult “what-if” questions — and it even helps you ask and answer questions you had never even imagined. 

Since we can’t predict the unpredictable, our next-best option is to be ready to act at any given moment. If you have a real-time, what-if mastery of the data, relationships and dependencies within your supply chain, you’ll be ready for any snag, shortage, or surplus — minus the sting of the bullwhip.

Harry Powell is the head of industry solutions at TigerGraph

DataDecisionMakers

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.

If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.

You might even consider contributing an article of your own!

Read More From DataDecisionMakers

Author
Topics