The Roots of Continuous Improvement

by Dr. Gordon D. Booth

Perhaps the most clearly distinguishing feature of the modern quality movement is that of Continuous Improvement. Both Deming and Juran have made this a core concept of their approaches to quality. Taguchi's contributions are somewhat controversial in many respects, but the single most important contribution that Taguchi made is to emphasize continual improvement, and for most quality workers, this is not controversial.

If continuous improvement is something we should strive for, it would be useful to know where its roots are planted. We don't have to search for long to find that continuous improvement can only go forward when its roots are well-established. Furthermore, those roots must lie in the predictability of the process being improved. Unless a process is predictable, any adjustments we make to the process will amount to nothing more than what Deming calls "tampering," and that usually leads to increased variation. Only when a process is predictable can we determine which core components must be "tweaked" in order to bring about improvement.

Therefore, the first step in improving any process is to make the process predictable, and the only processes that are predictable are those processes that are stable. To attempt improvement of an unstable process is likely to result in changes that make the process worse rather than better.

Unfortunately, far too many companies are involved in trying to improve unstable processes. The initial step of stabilizing the process is skipped in order to save time. The result is a process that resists almost all efforts at improvement. Companies, that have tried to improve their processes but have failed due to lack of process predictability, often blame the failure on the quality staff and respond by reducing quality efforts. The best protection for the quality professional is to establish the need for stabilizing the process as a first step to improvement.

Continuous Improvement is not a pie-in-the-sky wish, but rather an achievable objective--if the improvment effort is applied to a stable, predictable process.