Once upon a time, manufacturing and logistics were independent entities, or so the story goes. The manufacturing people produced the goods, and then the transportation and distribution people took over and dealt with whatever came their way.
We use the classic fairy tale opener "once upon a time," because that was never entirely true. What was the case—and far too often, remains the case—was that business functions were walled off from one another, which impeded communication and created the kind of inefficiencies we no longer tolerate. Manufacturing is no longer an independent variable in your world, at least if that world is one in which supply chains are integrated.
The argument for integrating manufacturing with supply chain functions is compelling, whether the manufacturing source is across the street, across the country, or across the ocean. But whatever the situation, we cannot afford to simply let manufacturing "happen," figuring we'll deal with the consequences later.
So, in the spirit of this series, let's look at some fundamental issues in manufacturing.
For the past three decades, the business world has been deluged with programs designed to transform manufacturing—and with all the attendant acronyms. All of these programs were promoted as transformative ideas that would elevate manufacturing performance to stratospheric levels. We've had just-in-time (JIT), total quality management (TQM), and kaizen; statistical process control (SPC) and single minute exchange of die (SMED); efficient consumer response (ECR) and quick response (QR); time-based manufacturing (TBM); six sigma, and more.
The concept du jour is "lean"—lean manufacturing, lean transportation, lean warehousing, lean logistics. You can't go anywhere without reading or hearing about "lean." But to be honest, we've peeked inside some lean programs and have found a remarkable resemblance to what we were doing 15 years ago, which itself wasn't all that different from programs dating back to the '70s.
Does that mean that these efforts have all been frauds? Not at all. The point is that the concepts behind organized manufacturing improvement have been around for a long time. What makes things different today—and improves the likelihood of a program's success—is the richness and robustness of modern information systems. We knew what to do, back in the day, but we were frustrated by shortfalls in data analysis capacity, by communication gaps, and by supply chains that were still inwardly focused. At the heart of things, it's all pretty simple. Today's manufacturing needs to be agile—nimble, flexible, waste-free, and in sync with ultimate demand. What it takes to make this happen is similarly straightforward. Manufacturers must drive up process reliability, build demand-based run strategies, synchronize with demand and respond to demand variation, and manage and communicate demand.
What does that mean? Let's talk about the component pieces, keeping in mind that in this limited space, this is merely an introduction to some key concepts.
From JIT to lean, nearly all of the process-improvement concepts aim at asset utilization—human assets, facility assets, material assets—and the elimination of waste, whether it's wasted time, effort, or products and materials.
In manufacturing, process reliability, for instance, has three components—uptime, dependability and first-run yield. Mastering performance in all three is crucial to achieving reliability. Reliability is expressed as a composite percentage; e.g., 90 percent uptime x 90 percent dependability x 90 percent first-run yield = 72.9 percent reliability. Looking deeper, uptime is the ratio of scheduled operation to what's available—16 hours out of 24, five days out of 7, or 50 weeks out of 52. Adding shifts or days raises human resource and facility wear-and-tear issues, or course. But it's important to note how it fits into understanding productivity; an operation with 95-percent dependability and 99-percent first-run yield that only runs two shifts, five days a week has an overall reliability of 44.8 percent (47.6 x .95 x .99)—not a figure to impress the CEO with.
First-run yield is the ratio of good output to input, subtracting waste, spoilage, trimmings and rework. Sometimes the opportunity to improve yield is trivial; sometimes it is enormous. Most often, the process improvement initiatives are aimed at boosting capacity or improving quality. Quality improvement, generally seen as actions taken to prevent waste, almost by definition improves first-run yield, reducing such things as spoilage and rework, for instance.
Dependability is a measure of actual versus scheduled operations, the ratio of the actual hourly run rate to the capable hourly run rate. The factors influencing the ratio include breakdowns, changeovers, time spent waiting for material, and off-speed operations.
Finally, there's run speed. It may be manufacturing's dirty little secret, but run speed can deliver big-time payoffs. In an operation that was designed, engineered and installed with a nominal rate of, say, 2,400 units/hour, performance can easily deteriorate over time to three-fourths of that rate or less. Reducing setup time, by whatever name, is key to short runs and flexibility.
Manufacturing managers address those issues and others with an eye to chipping away at waste, reducing setup times, establishing consistent run rates, optimizing facility utilization, and eliminating extraneous activity.
The complete solution requires many tools and techniques. And you may find there's some value to borrowing from a number ofprograms—lean, JIT, whatever—tailoring the overall approach to the organization's specific needs and priorities (and culture).
But wait: that's just the foundation. As we suggested at the outset, manufacturing efficiency is just part of the business equation, not a free-standing one. Once the manufacturing house is in order, or at least well on its way, the enterprise is positioned to better synchronize production—and inventory—with customer demand. That's easier said than done, because: 1) it's not always easy to know demand; 2) demand can be skewed by unnatural factors that are nonetheless common business practices (e.g., promotions, diversions, minimums); 3) multiple supply chain touch points can filter or distort ultimate demand; and 4) events can overlay baseline demand.
Manufacturing must have decent knowledge of real demand and good visibility of events that can affect it for good or ill. With that groundwork in place, you can develop run strategies to better align manufacturing output with demand patterns. To give an admittedly oversimplified example, that might mean items in high demand are run every demand cycle and those in lesser demand every few cycles. (A cycle is the smallest capable time frame—daily is often ideal.) Adjusting the quantities of each item class based on actual consumption tightens the synchronization, and largely confines low-volume goods to small inventories. These principles apply, again, whether manufacturing is in Pekin, Ill., or in Taipei.
Demand communication is key to making all this happen. It's essential to adjust production based on timely notice of variations in baseline demand, advance notification of events and promotions, seasonality, and event and season tracking. This requires collaborative planning, forecasting and replenishment (CPFR) tools, or something akin to them, plus point-of-sale current demand data.
Even with the best systems, demand management is an imperfect science. Our marketers and salespeople are attuned to selling, not to the supply chain. Can we ever force them to behave? Maybe someday, but not anytime soon. So, it behooves us to get the manufacturing act as together as it possibly can be. That will allow us to handle the normal crises with some grace and style, conserving our energies for the extraordinary ones.