• What was new or surprised you?
• What you agree or disagree?

The Efective Decision systematic process, with clearly defined elements, that is handled in a distinct sequence of steps. By Peter F. Drucker tives know that decision making has its own Effective exectives systematic process and its own clearly defined decisions. They cones do not make a great many elements. tant. They. They concentrate on what is imporsions on they try to make the few important deci Sequential Steps standing. They try to find the constants in a The elements do not by themselves "make" situation, to think through what is strategic and the decisions. Indeed, every decision is a riskgeneric rather than to "solve problems." They taking judgment. But unless these elements are are, therefore, not overly impressed by speed in the stepping-stones of the executive's decision decision making; rather, they consider virtuosity process, he will not arrive at a right, and cerin manipulating a great many variables a symp- tainly not at an effective, decision. Therefore, tom of sloppy thinking. They want to know in this article I shall describe the sequence of what the decision is all about and what the un- steps involved in the decision-making process. derlying realities are which it has to satisfy. There are six such steps: They want impact rather than technique. And 1. The classification of the problem. Is it ge they want to be sound rather than clever. Effective neric? Is it exceptional and unique? Or is it the Effective cxecutives know when a decision first manifestation of a new genus for which a as to be based on principle and when it should rule has yet to be developed? be made pragmatically, on the merits of the 2. The definition of the problem. What are case. They know the trickiest decision is that we dealing with? between the right and the wrong compromise, we 3. The specifications which the answer to the and they have learned to tell one from the other. problem must satisfy. What are the "boundary They know that the most time-consuming step problem witions"? in the process is not making the decision but 4. The decision as to what is "right," rather putting it into effect. Unless a decision has than what is acceptable, in order to meet the "degenerated into work," it is not a decision; boundary conditions. What will fully satisfy the it is at best a good intention. This means that, specifications before attention is given to the comwhile the effective decision itself is based on promises, adaptations, and concessions needed to the highest level of conceptual understand- make the decision acceptable? ing, the action commitment should be as close AUraor's xorE: This article is derived from 2 chapter as possible to the capacities of the people who in my forthcoming book. The Effective Exrcutire, to be have to carry it out. Above all, effective execu- publisbed by Harper \& How, Pablisbers, Inc. 5. The building into the decision of the action The compuny Effective Decision 93 to carry it out. What does the action commitment from another, that receives an offer to merge have to be? Who has to know about it? 6. The feedback which tests the nalidity atd situation as fat as the individual company, its effectiveness of the deciston against the actual board of directors, and its management are conconurse of erents. How is the decision being car- cerned. But it is, of course, a generic situation ried out? Are the assumptions on which it is based which occurs all the time. Thinking through appropriate or obsolete? Let us take a look at each of these individual some gencral rules. For these, however, the execuelements. tive has to look to the experience of others. The Classification Next there is the truly exceptional event that The effective docision maker asks: Is this a The huge power failure that plusged into darksymptom of a fundamental disorder or a stray news the whole of Northeastern North America event? The generic always has to be answered from the St. Iawrence to Washington in Novemthrough a rule, a principle. But the truly ex- ber 1965 was, according to first explanations, a ceptional event can only be handled as such and truly excepeional situation. So was the thalidomide as it comes. tragedy which led to the birth of so many deformed babies in the early rog 90 . The probability of Strictly speaking, the executive might dis- either of these events occurring, we were told, was tinguish among four, rather than between two, one in tem million or one in a hundred million, and different types of occurrences. First, there is the truly gencric event, of ever to rocur again as it is unlikely, for instance, which the individual occurrence is only a symp- for the chair on which I sit to disintegrate into its tom. Most of the "problems" that come up in constituent atoms. the course of the executive's work are of this Truly unique events are rare, however. Whennature. Inventory decisions in a business, for ever one appears, the decision maker has to ask: instance, are not "decisions." They are adapta- Is this a true exception or only the first man!tions. The problem is generic. This is even festation of a new genus? And this - the early. more likely to be true of occurrences within manifestation of a new generic problent - is manufacturing organizations. For example: the fourth and last category of cvents with A product control and engineering group will which the decision process deals. Thus: typically handle many hundreds of problems in We lnow now that both the Northeastern power the course of a month. Yet, whenever these are failure and the thalidomide tragedy were only the analyzed, the great majority prove to be just symp- first occurrnces of what, under conditions of modtoms +- and manifestations - of underlying basic ern power technology or of modern pharmacology, situations. The individual process control engl- are likely to become fairly frequent occurrences neer or production engineer who works in one part unless generic solutions are found. of the plant usually cannot see this. He might All events but the truly unique require a have a fewy problems each montha with the coup- lings in the pipes that carry steam or hot liquids, generic solution. They require a rule, a policy, and that's all. Only when the total workload of the group orer developed, all manifestations of the same geseveral months is analyzed does the generic prob- neric situation can be handled pragmatically lem appear. Then it is seen that temperatures or that is, by adaptation of the rule to the concrete pressures have become too great for the existing circumstances of the case. Truly unique events, equipment and that the couplings holding the however, must be treated individually. The various lines together need to be redesigned for executive cannot develop rules for the excepgreater loads. Until this analysis is done, process evecution tional. control will spend a tremendious amount of time to the control will spend a tremendous amount of time tional. situation. determining with which of the four different The second type of occurrence is the prob- types of the above situations he is dealing. He lem which while a occurrence is the prob- knows that he will make the wrong decision if vidual institwile a unique event for the indi- be classifies the situation incorrectly. vidual institution, is actually generic. Consider: By far the most common mistake of the deci- 94 HBK Jan. -Feb. 1967 sion maker is to treat a gencric situation as if it lem of automotive safety. It was this lack of awarewere a series of unique events - that is, to be ness - far more than any reluctance to spend pragmatic when lacking the generic understand- money on safety engineering - that eventually, ing and principle. The inevitable result is frus- In 1966 , brought the industry under sudden and tration and futility. This was elearly shown, I sharp Congressional attack for its unsafe cars think, by the failure of most of the policies, both and then left the industry wotally bewildered by domestic and foreign, of the Kennedy Adminis- the attack. It simply is not true that the industry tration. Consider: has paid scant attention to safety. For all the brilliance of its members, the Ad. highway engineering and at driver training, believministration achieved fundamentally only one suc- ing these to be the major areas for concem. That cess, and that was in the Cuban missile crisis. accldents are caused by unsafe roads and unsafe Otherwise, it achieved practically nothing. The drivers is plausible enough. Indeed, all other main reason was surely what its members called agencies concerned with automotive safety, from "pragmatism" - namely, the Administration's re- the highway police to the high schools, picked the fusal to develop rules and principles, and its in- same targets for their campaigns. These camsistence on treating everything "on its merits." Yet paigns have produced results. The number of it was clear to everyone, including the members accidents on highways built for safety has been of the Administration, that the basic assumptions greatly lessened. Similarly, safety-trained drivers on which its policies rested - the valid assump. have been involved in far fetrer accidents. tions of the lmmediate postwar years - had be- But although the ratio of accidents per thoucome increasingly unrealistic in international, as sand cars or per thousand miles driven has been well as in domestic, affairs in the 1960 s. going down, the total number of accidents and the Equally common is the mistake of treating a severity of them have kept creeping up. It should new event as if it were just another example of therefore have become clear long ago that somethe old prob if it were just another example of thing would have to be done about the small but the old problem to which, therefore, the old significant probability that accidents will occur rules should be applied: despite safety laws and safety training. This was the error that snowballed the local This means that future safety campaigns will power failure on the New York-Ontario border have to be supplemented by engineering to make into the great Northeastern blackout. The power aceidents theniselves less dangerous. Whereas cars engineers, especially in New York City, applied have been engineered to be safe when used corthe right rule for a normal overload. Yet their rectly, they will also have to be engineered for own instruments had signaled that something quite safety when used incorrectly. extraordinary was going on which called for exceptional, rather than standard, countermeasures. There is only one safeguard against becoming By contrast, the one great triumph of Presi- the prisoner of an incomplete definition: check dent Kennedy in the Cuban missile crisis rested facts, and throw out a definition the moment it on acceptance of the challenge to think through fails to encompass any of them. an extraordinary, exceptional occurrence. As The effective decision maker always tests for soon as he accepted this, his own tremendous signs that something is atypical of something resources of intelligence and courage effective- unusual is happening. He always asks: Does ly came into play. the definition explain the observed events, and The Definition does it explain all of them? He always writes out what the definition is expected to make hapOnce a problem has been classified as generic pen - for instance, make automobile accidents or unique, it is usually fairly easy to define. disappear - and then tests regularly to see if "What is this all about?" "What is pertinent this really happens Finally, he goes back and here?" "What is the key to this situation?" thinks the problem through again whenever he Questions such as these are familiar. But only sees something atypical, when he finds phethe truly effective decision makers are aware nomena his explanation does not really explain, that the danger in this step is not the wrong or when the course of events deviates, even in definition, it is the plausible but incomplete one. details, from his expectations. The American automobile industry beld to a laid down for medical diagnosis well over 2,0o0 plausible but incomplete definition of the prob-. years ago. They are the rules for scientific ob- The Effective Decision 95 servation first formulated by Aristotle and then velt who changed. The sudden economic collapse reaffirmed by Galileo 300 years ago. These, in which occurred between the summer of 1932 and other words, are old, well-known, time-tested the spring of 1933 changed the specificutions. A rules, which an executive can learn and apply policy appropriate to the goal of national economic systematically. recovery - which a conservative economic policy might have been - was no longer appropriate The Specifications when, with the Bank Holiday, the goal had to be- The next major element in the decision pro- come political and social cohesion. When the cess is defining clear specifications as to what the boundary conditions changed, Noosevclt inmedidecision has to accomplish. What are the ob- his former economic one (recovery). jectives the decision has to reach? What are Above all, clear thinking about the boundary the minimum goals it has to attain? What conditions is needed to identify the most danare the conditions it has to satisfy? In science conditions is needed to identify the most dan- these are known as "boundary conditions." A gerous of all possible decisions: the one in which decision, to be effective, needs to satisfy the the specifications that have to be satisfied are boundary conditions. Consider: "Can our neods be satisficd," Alfred P. Sloan. is the decision that might - just might - work Jr. presumably asked himself when he took corm- is President Kennedy's Bay of Pigs decision: mand of Gencral Motors in 1922, by removing the autonomy of our division beads? His answer One specification was clearly Castro's overthrow. On the the autonomy of our division heads?" His answer One specification was clearly Castro's overthrow: was clearly in the negative. The boundary condi- The other was to make it appear that the invasion tions of his problem demanded strength and re- Was a "spontaneous" uprising of the Cubans. But sponsibility in the chicf operating positions. This these two specifications would have been comwas needed as much as unity and control at the patible with each other only if an immediate center. Everyone before Sloan had seen the prob- Island-wide uprising against Castro would have lem as one of personalities - to be solved through completely paralyzed the Cuban army. And while a struggle for power from which one man would this was not impossible, it clearly was not probable emerge victorious. The boundary conditions, Sloan in such a tightly controlled police state. realized, demanded a solution to a constitutional Decisions of this sort are usually called "gamproblem - to be solved through a new structure: Decisions of this sort are usually calied "gam- decentralization which balanced local autonomy bles." But actually they arise from something decentralization which balanced local autonomy bles." But actually they arise from something of operations with central control of direction and much less rational than a gamble - namely, a of operations with central control of direction and bope against hope that two (or more) elearly policy. A decision that does not satisfy the boundary multancously. This is hoping for a miracle; conditions is worse than one which wrongly de- and the trouble with miracles is not that they fines the problem. It is all but impossible to happen so rarely, but that they are, alas, singusalvage the decision that starts with the right larly unreliable. premises but stops short of the right conclu- Everyone can make the wrong decision. In sions. Furthermore, clear thinking about the fact, everyone will sometimes make a wrong boundary conditions is needed to know when decision. But no executive needs to make a dea decision has to be abandoned. The most com- cision which, on the face of it, seems to make mon cause of failure in a decision lies not in its sense but, in reality, falls short of satisfying the being wrong initially. Rather, it is a subse- boundary conditions. quent shift in the goals - the specifications which makes the prior right decision suddenly The Decision inappropriate. And unless the decision maker The effective executive has to start out with has kept the boundary conditions clear, so as to what is "right" rather than what is acceptable make possible the immediate replacement of the preciscly because he always has to compromise outflanked decision with a new and appropriate in the end. But if he does not know what will policy, be may not even notice that things have in the cnd. Butisfy the boundary conditions, the decision changed. For example: maker cannot distinguish between the right Franklin D. Roosevelt was bitterly attacked for compromise and the wrong compromise - and his switch from conservative candidate in 1932 may end up by making the wrong compromise. to radical President in 1933. But it wasn't Roose- Consider: 96 HBR Jan,-Feb. 1967 I was taught this when I started in t944 on my In fact, no decision has been made unless carryfirst big consulting assignment. It was a study of ing it out in specific steps has become someone's the management structure and policies of General work assignment and responsibility. Until then, Motors Corporation. Alfred P. Sloan, Jr., who was it is only a good intention. then chairman and chief executive office of the it is only a good intention. company, called me to his office at the start of my cially those of business, is that they contain no assignment and said: "I shall not tell you what to action commitment - to carry them out is study, what to write, or what conclusions to come no one's specific work and responsibility. Small is to put down what you think is right as you see it. wonder then that the people in the organization Don't you worry about our reaction. Don't you tend to view such statements cynically, if not as worry about whether we will like this or dislike declarations of what top management is really that. And don't you, above all, concern yourself not going to do. with the compromises that might be needed to Converting a decision into action requires make your conclusions acceptable. There is not answering several distinct questions: Who has one executive in this company who does not know to know of this decision? What action has to how to make erery single conceivable compromise be taken? Who is to take it? What does the without any help from you. But he can' make action have to be so that the people who hare to the right compromise unless you first tell him what action have to be so that the people who hare to do it right is." questions are too often overlooked - with dire The effective executive knows that there are results. A story that has become a legend among two different kinds of compromise. One is ex- operations researchers illustrates the importance pressed in the old proverb: "Half a loaf is bet- of the question, "Who has to know?": ter than no bread." The other, in the story A major manufacturer of industrial equipment of the Judgment of Solomon, is clearly based on decided scteral vears ago to discontinue one of its the realization that "half a baby is worse than models that hat for years been standard equipno baby at all." In the first instance, the bound- ment on a line of machine tools, many of which ary conditions are still being satisfied. The pur- were still in use. It was, therefore, desided to sell pose of bread is to provide food, and half a loaf the model to present owners of the old equipment is still food. Half a baby, howerer, does not for another three ycars as a replacement, and then satisfy the boundary conditions. For half a to stop making and selling it. Orders for this parbaby is not half of a living and growing child. ticular model had beea going down for a good It is a waste of time to worry about what will many years. But they shot up immediately as cusbe acceptable and what the decision maker tomers reondered against the clay when the modshould or should not say so as not to evoke re- el would no longer be available. No one had, sistance. (The howerer, asked, "Tho needs to know of this happen. while objections and difficulties no one decision?" Consequently, nobody informed the purchasing thought about may suddenly turn out to be al- clerk who was in charge of buying the parts from most insurmountable obstacles.) in other words, which the model itself was being assemblod. His the decision maker gains nothing by starting instructions were to buy parts in a given ratio to out with the question: "What is acceptable? "current sales - and the instructions remained For in the process of answering it, he usually unchanged. gives away the important things and loses any Thus, when the time came to discontinue furchance to come up with an effective - let alone ther production of the model, the company had in the right - answer. years of production, parts that had to be written The dction off at a considerable lons. Converting the decision into action is the fifth The action must also be appropriate to the major element in the decision process. While capacities of the people who have to carry it thinking through the boundary conditions is the out. Thus: most difficult step in decision making, convert- A large U.S, chemical company found itself, in ing the decision into effective action is usually recent years, wich fairly large amounts of blocked the most time-consuming one. Yet a decision currency in two West African countries. To prowill not become effective unless the action com- tect this money, top management decided to invest mitments have been built into it from the start. it locally in businesses which (a) would contribute The Effective Decision 97 to the loeal economy, (b) would not require im- rather than a nationalired, telephone system. Yet ports from abroad, and (c) would if successful be this policy statement might have remained a dead the kind that could be sold to local investors if and letter if Vail had not at the same time designed when currency remittances became possible again yardsticks of service performance and introduced To establish these businesses, the company devel- these as a means to measure, and ultimately to reoped a simple chemical process to preserve a trop- ward, managerial performance. The Bell manaical fruit - a staple crop in both countries - gers of that time were used to being measured by which, up until then, had suffered serious spoilage the profitability (or at least by the cost) of their in transit to its Western markets units. The new yardsticks resulted in the rapid The business was a success in both countries. acceptance of the new objectives But in one country the local manager set the business up in such a manner that it required highly C In sharp contrast is the recent failure of a skilled and technically trained management of a brilliant chairman and chicf executive to make kind not easily available in West Africa. In the effective a new organization structure and new other country the local manager thought through objectives in an old, large, and proud U.S. comthe capacitics of the people who would eventually pany. Everyone agreed that the changes were have to run the business. Consequently, he worked needed. The comprany, after many years as leader hard at making both the process and the business of its industry, showed definite signs of aging. In simple, and at staffing his operation from the start many markets newer, smaller, and more aggreswith local nationals right up to the top manage- sive competitors were outflanking it. But contrary ment kevel. A few years later it became possible again to new ideas, the chairman - in order to placate transfer currency from these two countries. But, the opposition - promoted prominent spokesmen though the business flourished, no bayer could be of the old school into the most visible and highert found for it in the first country. No one available salaried positions - in particular into three new locally had the necessary managerial and technical executire vice presidencies. This meant only one skills to run it, and so the business had to be liqui- thing to the people in the company: They don't dated at a loss. In the other country so many local really mean it." If the greatest rewards are given entrepreneurs were cager to buy the business that for behavior contrary to that which the new course the company repatriated its original investment of action requires, then everyone will conclude with a substantial profit. The chemical process and the business built on and are going to reward. the first country no one had asked: "What kind Only the most effective executive can do what it were esentially the same in both places. But in of people do we have available to make this deci- Vail did - build the execution of his decision sion effective? And what can they do?" As a into the decision itself. But every executive can result, the decision itself became frustrated. This action commitment becomes doubly im- follow from it, and what people are available to portant when pcople have to change their be- carry it out. havior, habits, or attitudes if a decision is to become effective. Here, the executive must The Feedback make sure not only that the responsibility for Finally, information monitoring and reportthe action is clearly assigned, but that the people ing have to be built into the decision to provide assigned are capable of carrying it out. 'Thus continuous testing, against actual events, of the the decision maker has to make sure that the expectations that underlic the decisions. Decimeasurements, the standards for accomplish- sions are made by men. Men are fallible; at ment, and the incentives of those charged with best, their works do not last long. Even the best the action responsibility are changed simultane- decision has a high probability of being wrong. ously. Otherwise, the organization people will Even the most effective one eventually becomes get caught in a paralyzing internal emotional obsolcte. conflict. Consider these two examples: This surely needs no documentation. And - When Theodore Vail was president of the every exccutive always builds onganized feedBell Telephone System 60 vears ago, he decided back - reports, figures, studies - into his dethat the business of the Befl System was scevice. cision to monitor and report on it. Yet far too This decision explains in large part why the United many decisions fall to achieve their anticipated States (and Cinada) has today an investor-owned, results, or indeed ever to become effective, de- spite all these feedback reports. Just as the view it has ceased to be appropriate or even rational. from the Matterhorn cannot be visualized by This is true for business decisions as well as for studying a map of Sivitzerland (one abstraction), governmental policies. It explains in large meaa decision cannot be fully and accurately evalu- sure the failure of Stalin's cold war policy in ated by studying a report. That is because re- Europe, but also the inability of the United ports are of necessity abstractions. States to adjust its policies to the realities of a Effective decision makers know this and fol- Europe restored to prosperity and economic low a rule which the military developed long growth, and the failure of the British to accept, ago. The commander who makes a decision until too late, the reality of the European Comdoes not depend on reports to see how it is be mon Market. Moreover, in any business I know, ing carried out. He - or one of his aides - failure to go out and look at customers and margoes and looks. The reason is not that effective kets, at competitors and their products, is also decision makers (or effective commanders) dis- a major reason for poor, ineffectual, and wrong trust their subordinates. Rather, the reason is decisions. that they learned the hard way to distrust ab- The decision maker needs organized informastract "communications." tion for feedback. He needs reports and figures. With the coming of the computer this feed- But unless he builds his feedback around direct back clement will become even more important, exposure to reality - unless he disciplines himfor the decision maker will in all likelihood be self to go out and look - he condemns himself even further removed from the scene of action. to a sterile dogmatism. Unless he accepts, as a matter of course, that he had better go out and look at the scene of action, he will be increasingly divorced from reality. All a computer can handle is abstractions. And abstractions can be relied on only Decision making is only one of the tasks of if they are constantly checked against concrete an executive. It usually takes but a small fracresults. Othervise, they are certain to mislead. tion of his time. But to make the important To go and look is also the best, if not the decisions is the specific executive task. Only an only ivay, for an executive to test whether executive makes such decisions. the assumptions on which his decision has been An effective executive makes these decisions made are still valid or whether they are becom- as a systematic process with clearly defined eleing obsolete and need to be thought through ments and in a distinct sequence of steps. Inagain. And the executive always has to expect deed, to be expected (by virtue of position or the assumptions to become obsolete sooner or knowledge) to make decisions that have signifilater. Reality never stands still very long. cant and positive impact on the entire organizaFailure to go out and look is the typical rea- tion, its performance, and its results characterson for persisting in a course of action long after izes the effective executive