The odd industrial barometer of capacity utilization

Today the Federal Reserve released its capacity utilization stats for February. The number stands at 76.4 percent—up a few tenths from February 2013, way up from a low of 64.0 in 2009, but not quite as high as experienced during the best years of the 1990s, when utilization hovered in the mid-80s.

That’s interesting and all, but I’ve always felt that capacity utilization was one of those odd barometers, because it becomes more complex at what happens to be the most common type of manufacturing business: the high-mix, low-volume manufacturer, including job shops and contract manufacturers. Many in metal fabrication of course fall into this category.

Plants measure utilization by the constraint process: However many parts can be shoved through its constraint process over a certain time determines maximum capacity. But what if that constraint process becomes more efficient? Capacity utilization falls, yet the business is producing more products. Moreover, most manufacturers in this country produce dozens, sometimes thousands of different parts or products. Each has its own demand cycles, lead-time requirements, and constraint process. The constraint process for one part may be in bending, for another part it’s in welding, and for yet another part it’s in hardware insertion.

Consider Atlanta Custom Fabrication, a Douglasville, Ga.-based fabricator for commercial kitchens. When I visited the plant last week, its new fiber laser sat idle. Workers weren’t scurrying to find material or set up the machine, nor was the machine down for unexpected (reactive) maintenance. The machine simply had finished its work for the day.

The fabricator purchased the Ermaksan laser to make sure it had available excess capacity in laser cutting. After all, if lasers can’t produce, they starve most operations downstream. Did the new system increase overall capacity? In one sense, yes, because it can now deliver a cut part to its constraint processes (be it forming, welding, or assembly) much faster, making those constraint processes a little more productive. If there is some last-minute change or hot order, workers downstream no longer wait for parts. The laser can cut those parts in extremely short order. But potential capacity didn’t jump dramatically, because cutting isn’t the constraint process.

For most of ACF’s products, the constraint is in final assembly, where workers put together components—from upstream processes and from external suppliers—for complex kitchen systems. But if a customer were to need flat cut components, ACF’s available capacity just went through the roof. The company’s average utilization, though, doesn’t really capture this potential.

Economists and other industry watchers of course use capacity utilization to judge changes in demand. In the short term, capacity utilization goes up, so does demand. But comparing the stats with those from past decades may not be that valuable, considering all the technological and philosophical (lean manufacturing, Six Sigma, et. al) changes manufacturers have undergone in recent years.

As just one example, those practicing quick-response manufacturing aim to maintain low average capacity utilization to ensure the shop can handle spikes in demand. In the QRM world, low capacity utilization doesn’t necessarily mean low demand or a struggling business; it may instead imply an extremely efficient shop that can respond to unpredictable customers.