Toward better supply chain metrics

The idea of using metrics in the distribution center as a tool for understanding performance and to seek the causes of failure is not new, but the effort to find the best metrics never ends.

This search for better metrics was the focus of session titled "Supply Chain Metrics Diagnostics" at the Council of Supply Chain Management Professionals Annual Global Conference. John Caltagirone, vice president of supply chain strategy for The Revere Group, and Robert Novak, associate professor of supply chain management at Penn State University, reported on research by the Penn State Center for Supply Chain Research into the metrics used by 11 food manufacturers. "They felt the measures they were using were not adequate," said Caltagirone. "They were looking for better processes or methodology. They wanted to transform lagging supply chain metrics in leading supply chain indices."

Novak explained that in the parlance of the researchers, a measure looked at a single activity—number of trailers loaded, for example. A metric combines two activities—number of trailers loaded on time. An index combines two or more metrics, such as various ways of gauging the perfect order. "It is very difficult to do," he said. "The big challenge is determining how many indices to include."

The goal ought to be to develop indices that allow managers to drill down into causes and effects, he explained. For example, he said, the food companies wanted to understand all the things that could prevent on-time delivery. The key was to identify which of these causes were "assignable,"—that is, which of those causes were not random but are something that managers could actually take steps to address.

By assessing performance continuously using statistical tools, Novak said, managers can see processes that are beginning to deteriorate even while performance appears to be under control. That allows them to address the causes before a service failure even occurs.