Having it all
There is no tradeoff between quality and quantity. The laws of physics dictate only operations that successfully reduce variation and improve quality will achieve high output and low work in progress. The good news is that we can have it all.
It is ironic that to this day we find people engaged in a quality versus quantity debate. Some folks are concerned the production of improved quality will result in slower production and higher unit costs. However, when the approach to improve quality based on reducing variation is successfully used, it almost certainly leads to higher throughput volume capacity, lower work in progress and cycle time, reduced unit costs, and improved customer service levels.
Many companies are directing great effort into work in progress, cycle time, inventory, adherence to schedule, throughput volume, and time in queue. While there is nothing wrong with wanting to improve these characteristics, most are suffering the frustration of seeing one measure improve after much work, only to note other measures worsen. The problem is widespread on both sides of the Pacific.
The first principle
The first relationship we need to understand is called “Little’s Law” and it is expressed in the formula: Throughput Volume = Work In Progress/Cycle Time
Bear in mind this law, while surprisingly effective, is not as exact a law as E=mc^sup 2^. It cannot be since every factory is unique with its own constraints, limitations, and capabilities. Nonetheless, the relationship expressed in Little’s Law is remarkably robust and useful in the long term.
The major problem is that many people try to isolate and improve throughput volume, work in progress, or cycle time as separate issues. Inventory is another aspect people struggle with, and it is closely associated with these facets.
Any approach that attempts to isolate and improve one of these characteristics is probably doomed. If the business is in poor shape to begin with, it might be possible to make some initial advances. But very soon we note that attempts to reduce work in progress usually result in lowering the throughput volume and perhaps late deliveries or out of stock situations; or that when we tried to increase throughput volume, all we achieved was a glut of work in progress. A brief look at the formula shows why. These characteristics are interconnected; inexorably enmeshed. To understand and improve one of them, it will be necessary to understand the relationships that exist between all of them.
Recently, much has been written claiming the answer lies in reducing cycle time. Themes such as timebased manufacturing are the latest craze. Again, the formula shows why. If work in progress is held constant and cycle time is reduced, throughput volume must increase. Great. Alternatively, if cycle time is reduced, but there is no need to increase throughput volume, work in progress can be reduced. Again, very good, but there are catches.
Another way to express cycle time is residence time, or the length of time the material or components reside in the process. Clearly, if this is reduced, work is flowing through the process faster. Also, Little’s Law seems to indicate if we can significandy reduce cycle time, we get a lot of control over throughput volume and work in progress, and this is correct (within the physical limitations of each operation).
Sadly, a common outcome from this revelation is that corporations, blissful in their ignorance, lowered targets for cycle time. In many cases, the people in the factory have few ideas on how to reduce cycle time without simply driving machines harder. When this is done, the most common outcomes are a rise in downtime and other process disturbances, which increase delivery delays and degrade quality, sometimes to the point where the customer rejects the product.
Cases abound in both Australia and the United States where the new target value for cycle time (or work in progress, throughput volume, or inventory) was plugged into computer models such as MRP 2 before the process was improved to this level. The disaster that follows is entirely predictable. So, attempts by many well-meaning people often result in a financial disaster. There is a better way.
The second principle
If you are familiar with the dice experiment, this principle will come as no surprise. The lessons from the experiment are that both cycle time and work in progress are a function of variation in volume, or if you prefer, variation in flow rate through the process. This is a fundamental principle of both statistics and operations management.
Unlike computer models, factories have physical constraints. Because of this, when variation is successfully reduced, throughput volume capacity increases in nearly every case.
In particular, variation in the inputs to a process, or in the first few events, has the greatest impact on cycle time, work in progress, and throughput volume. And yet it is quite common to find our most talented people clustered around the transformational event or that part of the process that does the core work that the process was created to do. Usually, they are too late.
This understanding gives us a way out, regardless of whether our problems are throughput volume, work in progress, or cycle time. Reduced variation always lowers both cycle time and work in progress for any given throughput volume. In factories, this reduced variation usually increases throughput volume capacity. This, in turn, provides an opportunity to reduce unit costs.
Knowing that reduced cycle time provides a competitive advantage is one thing. Knowing what to do to reduce cycle time without pushing the process past its breaking point is another. In this case we are talking about reducing variation in volume, as opposed to reducing variation in product characteristics (or quality). So, how do we do that?
Two proven approaches
You might think reducing variation is imperative. Maybe you can also convince a few executives, but this is not enough. Somehow it is necessary to create an approach that disentangles the middle managers and technical folk from the knot of day-to-day issues so they can focus on reducing variation. After studying many approaches, only two enjoy a predictably high level of success.
But let us commence with an approach that, by itself, does not work. Sadly, it is also very common It is a training-based approach where mass training is conducted and a plea is made by senior executives at the end of each course for participants to go forth and do battle with variation.
There is nothing intrinsically wrong with large scale training. It can be invaluable in creating common understandings, a shared experience, a common language, and new skills. But by itself, it seldom brings about significant change. Something more is needed
Approach one – Making reduced variation the aimof the operation.
In this approach, senior executives make reducing variation the number one business imperative. Everything else is, at least temporarily, subordinate. This is what Geoff Ward did at Sola Optical Australia. His managers and team leaders had no doubt that conquering variation took precedence over all else. This is easy to say, but much more difficult to do. It severely tests the courage of whoever issues the instruction.
Because Ward never wavered, his people stayed focused on reducing variation. The approach worked. Cycle time fell, as did work in progress. Throughput volume capacity rose. Yields increased and costs fell. Ward took the business through a metamorphosis-and a profitable one at that.
Approach two – Steady state trial.
This approach is most successful in continuous and semi-continuous processes, although it has been applied successfully in all types of manufacturing operations. A plant trial, with its aim to hold every conceivable variable constant for any given product, is conducted. Raw materials, machine set-ups, operating procedures, temperatures, pressures, flow rates, and all other variables are kept as constant as possible for any given product.
This approach is spectacularly successful when the process under examination is unstable from the start. Sometimes the trial never ceases.
Once low levels of variation are achieved, the technicians find that isolating and understanding causal relationships is simplified, and further improvements flow. A key characteristic of this approach is that because it is a formal plant trial, complete with deadlines, a strong focus on the job at hand is achieved. This is the type of approach used by Peter Smith at Leinster Nickel Operations. Again, cycle time fell and throughput volume rose. Recoveries improved and unit costs fell. Smith took a struggling operation and transformed it.
ICI had a similar experience with a major petrochemical process. A steady state trial resulted in better quality and higher output. Again, the financial improvement was measured in the several millions of dollars.
Quality and quantity
Chief among causes of variability in volume are quality and reliability problems. The latter are obvious. If breakdowns and other disturbances occur, variability in cycle time is inevitable. Product quality also heavily affects variability in volume. If people and machines must struggle to obtain a satisfactory result, variation in work rates and therefore throughput volume is almost inevitable.
Rework and scrap have a similar effect. Quite apart from the obvious cost of poor quality, we also suffer an increase in variation in volume. This additional cost almost never appears in the cost of quality analyses, and yet it is nearly always significant.
Interestingly, the first successful applications of this approach I saw were in Metropolitan Permanent Building Society before it became Metway Bank. Howard Manning led several projects that significantly reduced cycle time for aspects such as production of monthly accounts and loan application turnaround time. It works just as well in the service sector as it does in manufacturing.
So we note poor quality and reliability are chief among the causes of variation in volume and therefore of high work in progress, poor throughput volumes, and higher unit costs. Little’s Law is a law, similar to the laws of physics. Understanding this law and the effect of variation can lead to large scale cost reduction and impressive increases in productivity.
Improved quality, reduced work in progress and inventories, higher throughput volume, lower unit costs, and enhanced customer service levels. We can have it all, if only we know how to reduce variability in the flow of work. References:
1. W. Hopp and M. Spearman, Factory Physics, (mcGraw-Hill, 1996).
2. J. McConnell, Metamorphosis, (Wysowl Pty Ltd, 1997).
Editor’s note: This article was previously published in Quality Magazine (Australia) April 1999, and is reprinted with the author’s permission.
John McConnell is a Brisbane, Australia-based consultant with a client roster that includes Fortune 100 corporations. He has written four books. The latest, Meta
morphosis, expands on the concepts presented here and takes a fresh look at the change process.
Copyright Association for Quality and Participation Winter 2001
Provided by ProQuest Information and Learning Company. All rights Reserved