McKinsey’s Operations Extranet has a new article that looks at the impact of variability in “volatile systems” (Achieving operational excellence in volatile systems — you will need to register but it is free). As the article notes, assembly lines can run with little variability but variability is an inherent part of complex systems. The article then claims that
[W]e have found it possible to smooth out month-to-month performance using a blend of new tools, tested operational levers, and a new “operating curve” methodology. The combination allows companies to measure and manage variability and, for the first time, to quantify the relationship between variability, lead times, and utilization.
The approach supposedly has delivered impressive results.
One container port operator, for example, was able to reduce variability in the arrival times of its cargo ships while boosting overall port throughput by 10 percent. A company in the steel business was able to cut the changeover time required to shift its production line from one product to another, improving it from 20 minutes to 5 minutes per instance, and to reduce lead times by 8 percent, freeing significant additional capacity. And a semiconductor manufacturer was able to minimize the impact of production-line disruptions by maintenance and R&D events—such as the ad hoc testing of new chips-significantly cutting variability and pruning manufacturing costs by 20 percent.
This, of course, begs the question of what the “operating curve” is. So here you go:
The point of this graph is that waits and queues grow with utilization and how badly they will grow depends on the utilization of the system. The article also includes a bit of animation that demonstrates how this can occur.
OK, now I will be snarky.
This hardly seems special. Consider, for example, the following graph I lifted from Managing Business Process Flows (a textbook written by several of my colleagues):This relationship between variability, utilization and performance is the bedrock of queuing theory. The article has a legitimate point that many businesses would benefit from thinking of creative ways to reduce variability and that techniques from lean ops as well as Six Sigma can be helpful. It is hard to see it as a novel methodology.