«Operator Errors – Time for a New Look J. M. Juran There are good grounds for asserting that our managerial convictions about “operator error” ...»
Operator Errors – Time for a New Look
J. M. Juran
There are good grounds for asserting that our managerial convictions about “operator error”
are derived mainly from unsupported beliefs which, over the years, have acquired an aura
of mystery, superstition and dogma. It further appears that these beliefs have become
deeply rooted in the managerial culture to a point that they are stubbornly resistant to
challenge, whether by competitive theory or by contradictory facts.
My grounds for these assertions are broader than just one man’s experience. In recent years, in my courses on Management of Quality Control, I have been taking soundings from the attending managers through assignment of group projects for discussion and solution. Some of these projects deal with the subject of operator error. Tallying up the results of the polls, discussions and conclusions of these and related soundings taken in these courses, it is my conviction that despite all the programs, posters, pledge cards, pep talks and the rest, many companies have been fighting a war without knowing clearly who is the enemy. This is no way to fight a war.
What is lacking is clear, quantitative knowledge about the nature of errors. Only through such knowledge can the superstition and dogma be challenged successfully. However, since we lack this knowledge, there persists a widespread confusion about the nature and extent of operator errors. Amid such confusion, it should not be surprising if managers appear to grasp at straws, at panaceas, at the lure of the demagogue. To many of these managers, such programs are squarely in line with their axiomatic beliefs, and hence, completely logical.
It is high time for our theoreticians and practitioners to wade into this subject, to think it through, and to put together the factual data needed to identify and quantify the elements which make up what we so glibly call “operator error.” What follows here is one man’s contribution.
Concept Of Self Control One most useful step we could take is to talk about first things first. There has been too much talk about operator motivation as if it were the starting point in dealing with errors. In modern industry the starting point is not operator motivation, but rather the concept of self control. Under this concept, we set up a job so as to make the job holder self-sufficient. We provide him with the means needed to carry out his assigned job, and we then hold him Selected Papers nº 11, 1968 1 Copyrights 1994 TPOK/Juran Institute responsible for getting results. We have been rather clear about insisting that the job holder produce results, but we have been less than clear about defining the "means needed" to
carry out the job, i.e., the criteria for self control. These consist of the following essentials:
1. Means for knowing what he is supposed to do.
2. Means for knowing whether he is doing what he is supposed to do.
3. Means for changing what he is doing if it does not conform with what he is supposed to do.
If we have failed to provide any of these essentials, the resulting errors should be classed as management controllable. If we have provided every one of these essentials, the resulting errors should be classed as operator-controllable, and it then becomes timely to talk about motivation.
It is most important that we understand clearly, in any specific situation, whether we are dealing with management-controllable or operator-controllable errors. These two species differ remarkably as to methods of diagnosis and especially, as to methods for remedy. The one spices requires that we look mainly to the designs, processes, methods, instruments and other features of the "situation," using the tools of science, engineering and management.
The other spices requires that we look mainly to the operator, using the tools of the behavioral sciences.
In the majority of companies there has been little study to determine what proportion of errors is operator-controllable and what proportion is management-controllable. Such is the feedback from my courses. However, enough of such studies have been conducted to tell us a great deal about these proportions. By and large, about 20% of all errors are operatorcontrollable, though this varies greatly among companies, and especially, from process to process. Accordingly, each company should know its own situation, in toto, and as to the important subdivisions. Lacking this knowledge, the grand strategy of error reduction is blind – the managers don't really know what they are fighting.
Managerial Theory Of Operator Error
The feedback from the courses on Management of Quality Control has been rather illuminating in discovering managers' beliefs about operator error. The polls (originally by a show of hands and more recently by secret ballot) have been in response to two questions
relating to man-agers' beliefs about worker motivation:
1. On a by-and-large basis, do your fellow managers back at home subscribe to the indifference theory* of industrial behavior (i.e., workers exhibit no enthusiasm for good work because people are no damn good) or do they subscribe to the craftsmanship theory* (i.e., workers exhibit no enthusiasm because the managers haven't designed industrial jobs in a way which makes them very interesting). In 12 such polls conducted during courses held in the United States (1964-1967), out of 522 attending managers polled, 209, or 40% felt that the managers back home adhere to the indifference theory, and 313, or 60% felt that the managers back home adhere to the craftsmanship theory.
Selected Papers nº 11, 1968 2 Copyrights 1994 TPOK/Juran Institute (The polls taken in the courses I have conducted abroad reveal some very interesting differences in managerial attitude internationally, but that is another story.)
2. If you ( the managers attending the courses ) were given charge of a brand new plant, would you run it under the indifference theory or the craftsmanship theory?
Out of 129 attendees polled:
121, or 94% would choose to operate under the craftsmanship theory;
8, or 6% would choose to operate under the indifference theory.
As to this question, there seem to be no differences internationally. In no country has such a poll shown less than 80% of the attendees favoring the craftsmanship theory. From these
polls it is my conclusion that:
1. Company management’s differ widely in their beliefs as to what causes industrial behavior adverse to company interests.
2. To an astonishing degree, managers attending the courses do not concur with these beliefs as presently held by their company management’s.
Under such a state of affairs, it seems evident (to me) that an essential prerequisite of a motivational program is an understanding of (a) what is the prevailing company attitude toward worker motivation, and (b) to what extent is this attitude really supported by the managers. This is important, since the managers do not merely theorize. They act on the theories. The decisions on whether to trust operators or not, whether to use many or few inspectors, whether to rely on system vs. people, are all related to what are the managers' beliefs about motivation. The character of the motivational program is likewise affected by these beliefs.
Subspecies Of Operator Error
Beyond the need for understanding the state of managerial theory about operator motivation, there is a need to look deeper than ever before into the nature of operator error itself. Right now there exists a widespread belief among managers that once we have met the criteria for operator self-control, the manager's job is done, and the rest is up to the operator. In terms of organization theory, this belief is well-founded. However, in terms of getting results, the belief is a gross over-simplification, since it ignores the wide differences in the kinds of operator error.
Some years ago this editor undertook to categorize the subspecies of inspector error. It turned out that these subspecies were so different from one another that knowledge of these differences was an essential prerequisite to reduction of inspector error**.
In the light of this experience, and with the clear need for a deeper understanding of the nature of operator error, it is timely to propose that we classify operator error into its
subspecies. There are at least three of these:
a. Willful errors. The operator is deliberately failing to comply. He could comply, but he has no intention of doing so, for reasons which are good enough for him.
Selected Papers nº 11, 1968 3 Copyrights 1994 TPOK/Juran Institute b. Lack of skill. The operator is unintentionally failing to comply. He is aware of the errors as he makes them, but he is unable to eliminate the errors – he isn't skillful enough.
c. Inadvertence. Not only is the error unintentional; the operator is even unaware he has made the error.
We have very little data on what is the industry-wide breakdown of operator errors among these subspecies. (This editor would welcome such data.) Since, as we shall see shortly, each of these subspecies has its own problems of diagnosis and remedy, the lack of this breakdown is again a handicap to grand strategy. Lacking data on the breakdown of operator errors by these subspecies, it is all too easy for managers to conclude that the errors are all willful, and all too many managers actually believe this.
Willful errors result from a wide variety of personal reasons: a short cut to make it easier to meet some other standard (cost, schedule, etc.); a belief that the quality standard is nonsense; a way of getting even with the boss; and so on. Some, perhaps most, of the reasons are the result of "good" intentions rather than malice. However, the logic behind all of these reasons is based on the premises of the operator. The manager who tries to understand these willful errors by logical reasoning from his own premises ends up in a state of frustration or in a state of unwavering belief in the "X" (indifference) theory of operator motivation.
The missing ingredient is the understanding of what is on the worker's mind when he makes and continues to make willful errors. Seldom can this understanding be arrived at solely by discussion among managers. They are trapped by their own premises within their own self-sealing thought system, including the fact that they (the managers) are also a part of the problem. To get objective answers the managers would need to sample employee thinking. In turn, since most managers are but amateur psychologists, such a study would require that they enlist the aid of the personnel specialists and behavior scientists. To bring these people into the act has complications of another sort. Such a project is difficult to organize and it carries the risk that at the end of it the manager will get answers not in the form of meaningful data, but in the form of the gibberish of the world of psychology.
Lack of Skill In studying processes which are highly operator-controllable, it is common to find that some operators make less errors than others. This is a consistent phenomenon – it goes on month after month.
What causes this consistent difference? There is a wide variety of theories: the better operators come from an ancestry with a tradition of good work; they are naturally better motivated; they have a knack; they have a natural aptitude (how do you explain Bill Tilden?); and so on.
Selected Papers nº 11, 1968 4 Copyrights 1994 TPOK/Juran Institute This editor takes a dim view of these armchair explanations. In my experience, when there is a consistent difference in operator performance, the place to discover the reason for the difference is the factory floor, not the office. The technique is to study, in detail the operations as performed by both the “good” operators and the “bad” operators. Most of the time these studies are rewarding; they disclose what it is that the one does which the other does not. In this way the “secret” is made known to the managers who then possess the key for bringing all operators up to the level of the best.
The foregoing technique is not limited to quality improvement; it has universal application in other industrial functions and outside of industry as well. In its generalized form, the
technique consists of:
1. Measuring the performance of multiple performers to discover who are consistently the best and who are consistently the worst.
2. Observation of the actual practice of the performers, both best and worst, to discover the differences in their practice.
3. Experimental verification to identify which of these differences of practice bring about the differences in results.
4. Extension of the best practices to all performers through retraining. redesign of the process, or other appropriate means.
The main significance of such studies lies in the fact that a solution is already in the house, i.e., the "best" operators are in fact solving the problem somehow. However, the nature of this "somehow" is not known to the management or to the "worst" operators. ( I have seen plenty of instances where it wasn't known to the "best" operators either) The problem in lack of skill is more usually one of finding this secret solution than one of inventing a new solution.
The successful end of such studies is to bring everyone up to the level of the best, to make a champion out of every duffer.
Inadvertent Errors The distinguishing feature of this subspecies is that at the time of the error the operator is not even aware of its existence until it is called to his attention, either by someone else or by some evident consequence of the error. Such errors are inherent in the fallible nature of the human animal (or at least such is the experience and hence the belief of some of us).
Recently we have seen an attempt to solve this problem by denying its existence, i.e., by asserting that human fallibility doesn’t exist once the right motivational lever is thrown. In due course the experimental evidence will have the last word as to this new theory.