At the turn of the modern century, companies began using predictive analytics to gain more insight into how their supply chains were performing, and how they might perform in the future. The impact was transformative, helping to blaze the trail for the next two decades of innovation that supercharged global commerce.

Now, with the ubiquity of powerful new artificial intelligence and machine learning technology, supply chain industry leaders can go deeper, not only analyzing the past and present to plan for the likelihood of outcomes, but also looking to the future to create steps to manage those outcomes — a practice known as prescriptive analytics.

According to McKinsey, prescriptive analytics is “far more scalable” than predictive analytics alone. By using machine learning and tapping into a wider array of internal and external data, it “reveals previously unknown patterns” and can make “real-time, easy-to-understand recommendations.” And while supply chain leaders can utilize prescriptive analytics as a standalone capability, the technology becomes much more powerful when used in tandem with descriptive and predictive analytics, combining historical performance with new pattern recognition and the real-world experience and instincts of industry veterans. 

Consider the Suez Canal incident of 2021, when a stuck containership blocked shipping lanes for six days and held up more than $60 billion worth of trade. Could supply chain leaders have predicted that a ship might get stuck somewhere? Perhaps, but nobody was prepared to deal with the multi-faceted fallout across the globe, including delayed shipments of food, textiles and construction materials; container shortages, and shipping congestion that affected nearly every sector of the global economy. While predictive analytics alone might have forecast the possibility of this type of event, combining prescriptive analytics could have helped supply chain leaders react, respond and rebound from this crisis much faster.

As our world becomes more interconnected, the risks and complexities of supply chains escalate. When the inevitable moments of sudden upheaval occur (like the COVID-19 pandemic or the ongoing war in Ukraine), companies tend to wind down operations, adapt procedures, and eventually ramp back up again with revised plans. All of this turmoil significantly impacts the supply chain. A recent study from the Economist estimated that disruptions have incurred “substantial financial costs (averaging 6%-10% of annual revenues), as well as reputational costs — in terms of customer complaints and damage to brand reputation — as companies have struggled to maintain supplies of their goods.”

Following are three core pillars that, when used as the basis of a connected supply chain, can make prescriptive analytics effective at mitigating risks, and delivering flexible, dynamic and resilient supply chains:

Supply chain optimization. Complex supply chains are composed of a variety of facility locations, distribution channels and transportation routes, each of which is susceptible to its own unique risks. In order to consistently account for potential interruptions and opportunities for improved efficiency, companies need access to comprehensive data that enable flexibility and responsiveness. For example, an assessment of an existing supplier can uncover risk factors that require a timely and strategic pivot, but companies must also maintain the integrity of the supply chain so that customers aren’t left hanging.

Prescriptive analytics can assess the risk associated with different suppliers by analyzing factors such as financial stability, geographic location and past performance. The technology can then recommend strategies to de-risk those suppliers, such as diversifying the supplier base or renegotiating contracts. UPS was a pioneer in prescriptive analytics, using the technology as early as 2015 to help ensure that packages were delivered on time and in the most efficient way. The company has since doubled down on its investment, partnering with Google Cloud and using more than 1 billion data points to provide more precise forecasting, and better control how packages move through the UPS network.

Demand forecasting and pricing. Unstable external factors like shifting consumer behaviors, dynamic competitor pricing and lagging economic indicators can often require an immediate reconsideration of inventory. For example, the rise of short-form popular content on social media like TikTok has turbocharged what’s known as a “trend cycle,” leaving companies with limited time to capitalize on the opportunity. One viral video has the potential to create unprecedented demand for a single product, and prescriptive analytics can offer recommendations on how to respond — not in general terms or historical best practices, but concrete steps based on a company’s current supply chain capabilities.

Business leaders can also use prescriptive analytics modeling to design pricing strategies that maximize revenue and profitability while meeting customer demand. This requires procurement teams to work closely with marketing, finance and operations to ensure that enterprise-wide data is included in prescriptive algorithms. Remember: The most valuable data is un-siloed data.

Inventory and production management. Following demand forecasts, data flows into physical product management, where it’s utilized for production planning, scheduling and stocking. The process must account for a wide array of variables, including supplier lead times, machine capacity, labor availability, storage capacity and order priorities. Supply chain leaders can rely on prescriptive analytics to analyze their historical data and advise on best practices, helping to head off errors such as stockouts or oversupply.

A good example of prescriptive analytics in production management comes from Harley-Davidson, which opted for just-in-time manufacturing and reduced its inventory levels by a whopping 75%. By reducing its waste and optimizing its process, Harley was able to centralize parts supplies, improve logistics, free-up factory floor space and create a more efficient supply chain and manufacturing operation.

The data used for supply chain optimization, demand forecasting and inventory management serves as the foundation for a larger, all-encompassing connected ecosystem. It can be helpful to imagine an efficient supply chain as an actual chain, taut and fastened from one end to another. If that chain sags or disconnects, there’s a noticeable decrease in strength that causes all sorts of problems. In the case of a retailer, for example, that might be a drop in on-shelf availability on one side, or an increase in inventory costs on the other. The root cause of this issue is sub-optimal processes and a lack of end-to-end visibility. Therefore, to truly deliver on the benefits of digital supply chains, companies must synchronize every aspect of their supply chains with connected data, creating a single ecosystem that ties everything together and allows all analytics to run unimpeded.

By switching to a connected supply chain, companies can widen the aperture on their entire process, bringing together disparate internal departments and disciplines, and external vendors and technology, to improve operational oversight. When data flows freely, analytics are juiced to their highest levels and can help companies optimize on-shelf availability, reduce inventory and logistics costs, and make faster decisions when problems arise.

Innovation is accelerating, and global supply chains will only get riskier and more complex. To compete, companies must invest in their own resilience. By utilizing prescriptive analytics, supply chain leaders will be better equipped to anticipate and manage change, mitigate risk, and create value across the organization.

Sunder Balakrishnan is supply chain analytics leader at LatentView Analytics.