Many of those in supply chain management learned early on the principles of just-in-time (JIT) manufacturing. Inspired by the work of Taiichi Ohno at Toyota Motor Co., the JIT revolution tightly coupled production and demand, resulting in little or no work-in-process inventory. Subsequently, it went on to change how supply chains ran. JIT fostered demand-driven, agile and lean processes, and contributed to breaking down functional silos inside an organization.

The adoption of JIT principles resulted in more than four decades of economic prosperity for global manufacturing and distribution. Yet much of that success relied on enabling factors that were existing at the time. To anticipate customer demand, companies relied on past sales data. They also assumed that customers would act rationally, with trade-offs that could be easily evaluated, and would respond logically to incentives.

To reduce inventory, companies depended on predictable transportation, with fast loading and unloading at ports around the world, and high availability of road, rail, barge and final-mile transportation. They were able to synchronize supply and demand by relying on abundant raw materials and components flowing through integrated supply chains, dedicated suppliers and manufacturing capacity.

Many of those assumptions no longer hold. COVID-19 has changed sales patterns and created holes in historical data. The new generation of consumers — digitally native, socially and eco-conscious — is changing demand patterns, making them more unpredictable than ever before.

This new world brings fresh opportunities to rethink how we plan and run supply chains. Succeeding under these conditions begins with three simple changes.

Adopt a probabilistic mindset. Probabilistic modeling is a technique that factors possible events or actions and their probabilities. Probabilistic planning involves predicting future outcomes and making plans that lead to desired cost and benefits. The most common example is a probabilistic demand forecast that outlines expected demand, while including a confidence interval or a probability model around it.

In fact, many parameters in planning should be modeled probabilistically. Lead times in master data are often inflated to represent worst-case figures. Planning leaders should demand that their systems model all the variability around lead times to factor the impact of shipments arriving early or late. Think beyond a future described by “one number,” and take in the wide range of potential outcomes.  

Explore options. The scenario analyses of yesterday no longer cut it. It’s not acceptable to spend days building scenarios and running what-if examples on stale data in spreadsheets, or use planning software to look at simple upside or downside scenarios. Planning leaders should have the capability to build comprehensive sets of scenarios that explore a wide range of outcomes. Some of these scenarios will be purpose-built, specifically modeling business events. However, planning software should also offer the ability to perform repeated sampling of the potential outcomes to evaluate the expected likelihood and rewards. Most of us would prefer plans with lower risk, even if the expected outcomes are lower. Look at all options and their potential outcomes.

Sense supply disruptions. Just as demand sensing significantly helps companies keep the pulse on rapidly changing trends and adapt to demand shifts, the same concepts should be applied to supply. Planning leaders should build processes that enable rapid detection of supply issues that impact the plans. Not all late orders result in fulfillment issues, so being able to narrow the alerts to what truly impacts the business is key. It's important to capture early signs of disruption, such as late supplier shipments, rolled cargo and port delays at arrival, but only if that’s paired with intelligence that translates these signs into new predicted arrival dates. Sense early and intelligently to expand the time and set of potential resolutions.

Together, these three approaches all have one theme in common: the need to anticipate quickly what will happen and plan for it, just in case.

This isn’t to say that just-in-time principles aren’t as relevant as in years past. By anticipating customer needs with the help of an outside-in perspective on demand, companies have achieved higher product availability and customer satisfaction. They’ve also realized significant working-capital reduction by understanding the bullwhip effect, reducing lead times and variability, and right-sizing safety stocks. And they’ve increased order-fulfillment metrics by driving toward small-batch processing, setup time reduction, optimized sequencing and better production planning.  

To succeed today, however, we need more. Supply chain leaders must embrace new principles, focusing on just-in-case:

• Think probabilistically. Don’t focus on the one number, but on the range of possibilities. 
• Build options. Develop more than one plan, and evaluate each for its likelihood and reward.
• Sense early to respond better. Know when disruptions will impact goals, to buy yourself plenty of time to respond.

Valerie Tardif is vice president of product management at Infor.