Strengthening the Weakest Link

Brian Tomlin explores how companies can manage risks lurking in their supply chains and discovers that the nature of the risk matters.

Airline executives weren’t the only ones anxiously awaiting the reopening of European airspace during April’s volcanic eruption in Iceland. Businesses relying on airfreight to import or export product—from fresh fruit to critical pharmaceuticals—had to scramble for alternative suppliers or transportation to avoid losing current and future customers.

While the volcanic-ash incident had a global impact and garnered worldwide attention, less obvious risks lurking in a company’s supply chain can lead to major supply shortages that don’t necessarily affect competitors. Case in point: Genzyme’s costly shortage of two key drugs due to a supplier-introduced virus in its Allston, Mass. production facility.


Tuck associate professor Brian Tomlin studies how companies can build robust supply chains that mitigate and absorb significant supply shocks. Recently, he and his colleagues explored the value of two strategies: dual sourcing and supplier improvement. The latter involves collaborating with a supplier to improve its reliability. Studies suggest that U.S. companies have historically lagged behind the Japanese somewhat in supplier development. “Academics have devoted attention to dual sourcing as a risk-mitigation tactic,” says Tomlin, “but supplier improvement had been largely overlooked.”

Setting out to remedy this important gap, Tomlin and his colleagues developed a sophisticated mathematical model that allowed them to investigate the merits of each strategy. A critical part of the model was its ability to capture the important characteristics that distinguish companies and industries. Particularly, it allowed for two crucial types of supply risk: process and capacity risk.

Process risk is involved when the underlying production process is inherently uncertain. The result is that the quantity of good product produced is a random fraction—including zero—of what the supplier had hoped to produce. Semiconductors and biotechnology offer prime examples or process risk. Capacity risk comes about when production resources (capital equipment, labor, energy) may fail or be less productive than anticipated, with the result that good products are made but fewer than planned.

Tomlin says that while both risks are present to some extent in all operations, different industries will feel more pressure on one or the other dimension. Industries relying on production technologies at the scientific cutting edge are more prone to process risk, whereas industries with well-understood and easier-to-control processes face capacity risk if the operating infrastructure within the plant or country is unreliable.

Businesses need to tailor their supply-risk management strategy to the nature of the risk. Consider two companies, each with two suppliers that differ greatly in their reliabilities. Dual sourcing is likely to be the preferred strategy in the case of capacity risk whereas supplier improvement is likely to be preferred in the case of process risk. “By hedging your bets and ordering a large amount from the reliable supplier and a smaller amount from the unreliable supplier,” says Tomlin, “you can manage capacity risk. But the same is not true for process risk: small orders are still subject to large risks.”

When supply risk is very high—regardless of its nature—a company might want to simultaneously develop dual suppliers and work with them to improve production reliability. Tomlin notes that this strategy is pursued by Xilinx, a leading semiconductor company that outsources production to two contract manufacturers. Their competitor Altera pursues a fundamentally different strategy of working exclusively with one contract manufacturer.

Does Tomlin’s research mean that Altera is wrong? “Not necessarily,” he says. “Altera would argue that the exclusive relationship allows their supplier to improve yields more rapidly than if they dual sourced, a factor not considered in our research.” So as usual, in the act of answering one important question, researchers often uncover another. “It’s what keeps our work exciting.”

Y. Wang, W. Gilland, B. Tomlin, “Mitigating Supply Risk: Dual Sourcing or Process Improvement?”, Manufacturing & Service Operations Management, forthcoming 2010

Yimin Wang, assistant professor of supply chain management at W.P. Carey School of Business, Arizona State University, was previously a Ph.D. student of Professor Tomlin. Wendell Gilland is associate professor of operations, technology, and innovation management at Kenan-Flagler Business School, The University of North Carolina, Chapel Hill.