You are on page 1of 2598

Making Management Decisions:

the Role of Intuition and Emotion

By Herbert A. Simon Camegie- M
ellon U niversity

The work of a manager includes making decisions (of

participating in their making), communicating them lntuition and Judgment
to others, and monitoring bow they are carried out.
Managers must know a great deal about the industrial As an appendix to the Functions of the Executive
and social environment in which they work and the (Harvard University Press. 1938), Chester L Barnard
making process itself to make decisions well. Over the published an essay, based on a talk he bad given in
past 40 years, the technique of decision making 1936 at Princeton, entitled "Mind in Everyday
has been greatly advanced by the Affairs." The central motif of that essay was a contrast
development of a wide range of tools -in particu1ar, between what. Barnard called "logical" and "nonlogical"
the tools of operations research and management processes for making decisions. He speaks of "the
science, and the technology of expert systems. But wide divergence of opinion . . . as to what constitutes a
these advances have not applied to the entire domain proper intellectual basis for opinion or deliberate ac-
of decision making. They have bad their greatest tion." And he continues:
impact on decision making that is well-structured,
deliberative and quantitative; they have had less
impact on decision making that is loosely structured, By "logical processes" I mea n conscious
intuitive, and qualitative; and they have bad the least thinking which could be expressed in words
impact on face to face interactions between a or by other symbols, that is, reasoning. By
manager and his or her coworkers-the give and take
of everyday work. In this article, I will discuss these "non-Logical processes" I mean those not ca-
two relatively types of decision making: "intuitive" pable of being expressed in words or as rea-
decision-making and decision making that involves soning, which are only made known by a
interpersonal. Interaction. What, if anything, do we
know about how judgmental and intuitive processes judgment, decision or action.
work and how they can be made to work better?
And why do managers often fail to do what they Barnard's thesis was that executive. as contrasted. say, with
know they should do – even what they have decided to scientists, do not often enjoy the luxury of making their
do? What can be done to bring action into closer decisions on the basis of orderly rational analysis, but
accord with intention? depend largely on intuitive or judgmental responses to
My artic1e will therefore have the form of decision-demanding situations. .
a diptych, which one half devoted to each of these Although Barnard did not provide a set of forma!
topics. First, I will discuss judgmenta1 and intuitive criteria for distinguishing between logical and judg-
decision-making; then I will turn to the subject of menta1 decision making, he did provide a phenomeno-
the manager 's behavior and the influence of logical characterization of the two styles that make
emotions on that behavior. Sometimes the term them easily recognizable. at least in their more extreme
rationa1 (or logical) is applied to decision making forms. In logical decision making, goals and
that is consciously analytic, the term nonrational to alternatives are made explicit, the consequences of
decision making that is intuitive and the term pursuing di1ferent alternatives are calculated, and
irrational to decision making behavior that responds to these consequences are evaluated in terms of how
the emotions or that deviates from action chosen close they are to the goals.
"rational1y." We will be concerned, then, with the In judgmenta1 decision making, the response to
nonrational and the irrational components of the need for a decision is usually rapid, too rapid to
managerial decision making and behavior Our task. allow for an orderly sequential analysis of the
you might say, is to discover the reason that underlies situation. and the decision-maker cannot usually give a
unreason. veridical account of either the process by whicl1 the
decision was reached or the grounds for judging it
correct. Never


5: /
theless, decision makers may have great confidence in rational and intuitive processes are so different that
the correctness of their" intuitive decisions and are they are carried out in different parts of the brain.
likely to attribute their ability to make them rapidly to
their experience.
Most executives probably find Barnard's account Split Brains and Forms of Thought
of their decision processes persuasive; it captures their
OWD feelings of how processes work. On the other Physiological research on "split brains"-brai8 in
band, some students of management, especially those which the corpus callosum, which connects the
whose goal is to improve management-decision hemispheres of the cerebrum, bas been severed-
processes, have felt less comfortable with it. It appears provided encouragement to the idea of two
to vindicate map judgments and to cast doubt on the qualitatively different kinds of decision making-the
relevance of management-science tools, which almost analytical, corresponding to Barnard's "logical,’ and
all involve deliberation and calculation in decision the intuitive or creative, corresponding to his "non-
making. Iogical".
Barnard did not regard the nonlogical processes The primary evidence behind this dichotomy is that
of decision as magical in any sense. On the contrary, the two hemispheres exhibit a division of labor: right-
he felt they were grounded in knowledge and handed people, the right hemisphere plays a special
experience: role in the recognition of visual patters, and r left
hemisphere in analytical processes and the use
The sources of these non-logical processes lie language. .
Other evidence in addition to the split-brain search
in physiological conditions or factors, or in suggests some measure of hemispheric specialization.
the physical and social environment mostly Electrical activity in the intact brain can be measured by
impressed upon us unconsciously or without EEG techniques. Activity in a hemisphere is generally
conscious effort on our part. They also consist associated with partial or total suppression in the hemisphere
of the alpha system, salient brain wave with a frequency of
of the mass of facts patterns concepts, about ten vibrations per second. When a hemisphere is
techniques, abstractions, and generally what inactive, alpha rhythm in that hemisphere becomes strong.
we call formal knowledge or beliefs, which For most right-banded subjects, when the brain is engaged in
are impressed upon our minds more or less by a task involving recognition of visual pattern, the alpha
rhythm is relatively stronger in the left than in ~ right
conscious effort and study. This second source hemisphere; with more analytical tasks, the alphas’ rhythm
of non-logical mental processes greatly is relatively stronger in the right hemisphere. (See Doctor
increases with directed experience, study and and Hamilton, 1973, and Doctor, 1975. some experiments
education (p. 302) and a review of the evidence. 2)
The more romantic versions of the split-brain doctrine
extrapolate this "evidence into the two polar forms of
At the time I wrote Administrative Behavior' thought labeled above as analytical and creative. As an
(1941-42), I was troubled by Barnard's account of intu- easy next step, evaluative nuances I creep into the
itive judgment (see the footnote on p. 51 of AB), discussion. The opposite of "creative," after all, is
largely, I think, because he left no clues as to what sub- "pedestrian." The analytical left hemisphere so this story
conscious processes go on while judgments are being goes, carries on the humdrum, practical, everyday work of
made. 2 I was wholly persuaded, however, that a theory the brain, while the creative right sphere is responsible for
of decision making bad to give an account of both con- those flights of imagination that produce great music, great
scious and subconscious processes (see the end of p. 75 literature, great science, and great management. The
to the top of p. 76). I finessed the issue by assuming evideDC2 this romantic extrapolation does not derive from
that both the Conscious and the unconscious parts of physiological research. As I indicated above, tm: search bas
the process were the same, that they involve drawing on provided evidence only for some mea specia1ization
factual premises and value premises, and operating on between the hemispheres. It does any way imply that
them to form conclusions that became the decisions. either hemisphere (especial right hemisphere) is capable
Because I used logic (drawing conclusions from of problem solving, decision making. , Or discovery
premises) as a central metaphor to describe the deci- independent of the other. The real evidence for two
sion-making process, many readers of Administrative different forms of thought is essentially that on which
Behavior have concluded that the theory advanced
there applies only to "1ogical " decision making, not to Barnard relied: the observation that, in everyday affairs,
decisions that involve intuition and judgment. That was men and women I make competent judgments or reach
certainly not my intent. But now, after nearly 50 years, reasonable decisions rapidly-without evidence indicating
the ambiguity can be resolved because we have that have engaged in systematic reasoning, and ' their
acquired a solid understanding of what the judgmental being able to report the thought processes took them to
and intuitive processes are. I will take up the new evi- their conclusion.
dence in a moment; but first, a word must be said about
the "two brains" hypothesis, which argues that
There is also some evidence for the very lntuition in Chessplaying
plausible hypothesis that some people,
confronted with a particular problem, make One much studied class of experts is the grandmasters in the
more use of intuitive processes in solving it, game of chess. Chess is usually believed to require a high level of
while other people make relatively more use of intellect, and grandmasters are normally full-time professionals
analytical processes (Doctor, 1978). who have devoted many years to acquiring their mastery of the
For our proposes, it is the difference in game. From a research standpoint, the advantage of the game is that
behavior, and not the difference in the the level of skill of players can be calibrated accurately from their
hemispheres, that are important. Reference to official ratings, based on their tournament success.
the two hemispheres is a red herring that can From the standpoint of studying intuitive thinking, chess
only impede our understanding of. intuitive , might seem (at least to outsiders) an unpromising research
non-logical thought. The important question domain. Chess playing is thought to
for us are "What is intuition?" and "How involve a highly analytical approach, with players working out
is it accomplished? Not, "In which cubic systematically the consequenCe5 of moves and countermoves, 50
centimeters of the brain tissue does it take that a single move may take as much as a half hour`s thought, or
place?" more. On the other hand, chess professionals can play simultaneous
games, sometimes against as many as 50 opponents, and exhibit
New Evidence on the Processes of only a moderately lower level of skill than in games playing under
tournament conditions. In simultaneous play the professional takes
Intuition much less than a minute, often only a few seconds, for each move.
There is no time for careful analysis.
When we ask the grandmaster or master how he or she is
In the 50 years since Barnard talked about the mind in able to find good moves under these circumstances, we get: the
everyday affairs, we have learned a great deal about same answer that we get from other professionals who are
processes human beings use to solve problems, to make questioned about rapid decisions:
decisions, and even to create works of art. and science.
Some of this new knowledge has been gained in the It is done by "intuition," by applying one's professional "judgment"
psychological laboratory; some has been gained to the situation. A few seconds' glance at the position suggests a
through observation of the behavior of People who are good move, although the player bas no awareness of how the
demonstrably creative in some realm of human endeavor; judgment was evoked.
and a great deal has been gained though the use of the Even under tournament conditions, good moves usually come
modern digital computer to model human thought to a player's mind after only a few seconds' consideration of the
processes and perform problem-solving and decision- board. The remainder of the analysis time is generally spent
making functions at expert levels. I should l like to examine verifying that a move appearing plausible does not have a hidden
this body of research, Witch falls under the labels of weakness. We encounter this same kind of behavior in other
"cognitive science" and “artificial intelligence”, to see professional domains where intuitive judgments are usually
what light it casts on intuitive, judgmental decision subjected to tests of various kinds before they are actually
making in management. We will see that a rather detailed implemented. The main exceptions are situations where the
account can be given of the processes that underlie decision bas to be made before a deadline or almost instantly. Of
judgment, even though most of these processes are not course we know that under these circumstances (as in professional
within the conscious awareness of the actor using them. chess when the allowed time is nearly exhausted), mistakes are
sometimes made.
How do we account for the judgment or intuition that allows
the chess grandmaster usually to find good moves in a few seconds?
The Experts Intuition A good deal of the answer can be derived from an experiment that
In recent years, the disciplines of cognitive science and is easily repeated. First, present a grandmaster and a novice with a
artificial intelligence have devoted a great deal of attention position from an actual, but unfamiliar chess game (with about 25
to the nature of expert problem solving and decision pieces on the board). After five or ten seconds, remove the board
making in professional-level tasks. and pieces and ask the sub
The goal of the cognitive science research has been to
gain an understanding of the differences between the
behavior of experts and novices, and possibly to learn
more about how novices can become experts. The goal of
artificial-intelligence research bas been to build computer
systems that can perform professional tasks as
competently as human experts can. Both lines of research
have greatly deepened. our understanding of expertise.
ping requires applying the brakes. For an experienced in a coherent way and they must be supplied with reasoning
driver, the sight of the red light simply evokes the capabilities that allow inferences to be drawn from the
applications of brakes. How conscious the actor is of information retrieved. and numerous chunks of information
the inversely, how automatic the response is, may to be combined. Hence intuition is not a process that operates
differ, but there is no difference in the logic being independently of analysis; rather. the two processes are
applied. essential complementary components of effective decision-
making systems. When the expert is solving a difficult
problem or making a complex decision, much conscious
deliberation may be involved. But each conscious step may
Intuition in Management itself constitute a considerable leap. with a whole sequence
of automated productions building the bridge from the
Some direct evidence also suggests that the intui- premises to the conclusions. Hence the expert appears to take
tive skills of managers depend on the same kinds of giant intuitive steps in reasoning. as compared with the tiny
mechanism as the intuitive skills of chessmasters or steps of the novice.
physician. It wouId be surprising if it were It is doubtful that we will find two types of managers
otherwise. (at least, of good managers). one of who relies a1most
The experienced manager, too, has in his or her exc1usively on intuition. the other on analytic techniques.
memory a large amount of knowledge. gained More likely, we will find a continuum of decision-making
nom training and experience and organized in styles involving an intimate combination of the two kinds of
terms of recognizable chunks and associated skill. We will likely also find that the nature of the problem
information. to be solved will be a principal determinant of the mix.
Marius J. Bouwman bas constructed a computer With our growing understanding of the organization of
program capable of detecting company problems judgmenta1 and intuitive processes, of the specific
from an examination of accounting statements. knowledge that is required to perform particular judgmenta1
The program was modeled on detailed thinking-aloud tasks. and of the cues that evoke such knowledge in situations
protocols of experienced financia1 analysts in which it is relevant, we have a powerful new 0001 for
interpreting such statements, and it captures the improving expert judgment We can specify the knowledge
knowledge that enables analysis to spot problems and. the recognition capabilities that experts in adman need
intuitively, usua1ly at a very rapid rate. When a to acquire as a basis for designing appropriate learning
comparison is made between the responses of the procedures.
program and the responses of an expert human financial We can also. in more and more situations, design
analyst, a close match is usua1ly found. expert systems capable of automating the expertise. or
In another study. R. Bhaskar gathered alliteratively. of providing the human decision-maker with
thinking-aloud protocols from business school an expert consultant Increasingly, we will see decision aids
students and experienced businessmen, who were or managers that will be highly interactive. with both
a1l asked to analyze a business policy case. The knowledge and intelligence being shared between the human
final analyses produced by the students and the and the automated components of the system.
businessmen were quite similar. A vast research and development task of extracting and
What most sharply discriminated between the cataloging the knowledge and cues used by experts in
novices and the experts was the time required to different kinds of managerial tasks lies ahead. Much has been
identify the key features of the case. The experts learned in the past few years about bow to do this. More
did this; it was done slowly. with much conscious needs to be learned about how to update and improve the
and explicit analysis by the novices. These two knowledge sources of expert systems, as new knowledge
pieces of research' are just drops of water in a becomes available.
large bucket needs filling. The description detail, Progress will be most rapid with expert systems that
of the use of judgmental and analytical is expert have a substantial technical component It is no accident that
problem solving and decision making deserves a the earliest expert systems were built for such tasks as
high priority in the agenda of management designing motors, making medical diagnoses. playing- chess,
research and finding chemical synthesis paths. In the area of
management, the analysis of company financia1 statements is
adman where some progress bas been made in constructing
Can Judgment Be Improved? expert systems.
From this and other research on expert problem The areas of corporate policy and strategy are excellent
solving and decision making. we can draw two candidates for early development of such systems.
main conclusions. First, experts often arrive at What about the aspects of executive work that involves
problem diagnoses solutions rapid1y and intuitively the managing of people? What help can we expect in
without being able to report how they attained the improving this crucial component of the man agreement task?
result. Second, this ability is best explained by
postulating a recognition and retrieval process that
employs a large number- generally tens of thousands
or even hundreds of thousands – of chunks or patterns
stored in long term memory.
When the problems to be solved are more than
trivial, the recognition processes have to be organized
Acodem)' o{ Manąemem EXECUTIVE

Knowledge and Behavior rary personal comfort at the expense of bad long-rut.
consequences for the organization.
What managers know they should do-whether by Behavior of this kind is "intuitive" in the sense
analysis or intuitively-is very often di1ferent from what that it represents response without careful analysis
they actually do. A common failure of managers, 8DĆ calculation. Lying for example, is much more
which all of us have observed, is the postponement of often the result of panic than of Machiavellian
difficult decisions. What is it that makes decisions scheming. The intuition of the emotion-driven
difficult and hence tends to cause postponement? manager is very different from the intuition of the expert
Often, the problem is that all of the alternatives have whom we earlier. The latter's behavior is the product of
undesired consequences. When people have to choose learning and experience, and is largely adaptive; the
the lesser of two evils, they do not simply behave like former's behavior is a response to more primitive
Bayesian statisticians, weighing "the bad against the urges. and is more often than not inappropriate. We
worse in the light of their respective possibilities. must not confuse the "nonrational" decisions of the
Instead they avoid the decision, searching for alterna- decisions that derive from expert intuition and judg-
tives that do not have negative outcomes. If such alter- ment-with the irrational decisions that stressful
natives are not available, they are likely to continue to emotions may produce.
postpone making a choice. A choice between I have made no attempt here to produce a com-
undesirables is a dilemma, something to be avoided or prehensive taxonomy of the pathologies of organiza-
evaded. tional decision making, but simply have given some
Often uncertainty is the 50urce of the difficulty. examples of the ways that stress interacts with
Each choice may have a good outcome under one set of cognition to elicit counterproductive behavior. Such
environmental contingencies, but a bad outcome under responses can become 50 habitual for individuals or
another. When this occurs, we also do not usually ob- even for organizations that they represent a
serve Bayesian behavior; the situation is again treated recognizable managerial "style." .
as a dilemma. Organizational psychologists have a great deal to
The bad consequences of a manager's decision are say about ways of motivating workers and executives
often bad for other people. Managers sometimes have to direct their efforts toward organizational goa1s.
to dismiss employees or, even more frequently, have to They have said less about ways of' molding habits so
speak to them about unsatisfactory work. Dealing with that executives can handle situations in a goal-directed
such matters face to face is stressful to many, perhaps manner. When it comes to handling situations, two
most, executives. The stress is magnified if the em- dimensions of behavior deserve particular attention:
ployee is a close associate or mend. If the unpleasant the
task cannot be delegated, it may be postponed. " response to problems that arise, and the initiation of
The manager who bas made a mistake (that is to activity that looks to the future.
say, all of us at one time or another) also finds himself
or herself in stressful situation. The matter must be
dealt with sooner or later, but why not later instead of Respond to Problems
sooner? Moreover, when it is addressed it can be ap-
proached in different ways. A manager may try to The response of an organization to a problem or
avoid blame-"It wasn't my fault!" A di1ferent way is to difficulty, whether it results from a mistake or some
propose a remedy to the situation. I know of no other cause, is generally one that looks both, backward
systematic data on bow often the one or the other and forward. It looks backward to establish responsi-
course is taken, but most of us could probably agree bility for the difficulty and to diagnose it, and
that blame avoiding behavior is far more common than forward to find a course of action to deal with iŁ
problem-solving behavior after a serious error bas been The backward look is an essential part of the
made. organization’s reward system. The actions that have
led to difficulties, and the people responsible for that
action, need to be identified. But the backward look
The Consequences of Stress can also be a source of serious pathologies.
Anticipation of it-particu1arly anticipation that it will
What all of these decision-making situations have be acted on in a punitive way-is a major cause for the
in common is stress, a powerful force that can divert concealment of problems until they can no longer be
behavior from the urgings of reason. They are hidden. It can also be highly- divisive, as individuals
examples of a much broader class of situations in point fingers to transfer blame to others. Such
which managers frequently behave in clearly outcomes can hardly be eliminated, but an
nonproductive ways. Nonproductive responses are organization's internal reputation for fairness and
especially common when actions have to be made objectivity mitigate them. So can a practice of
under time pressure. The need to allay feelings of subordinating the blame finding to a diagnosis of
gu1lt, anxiety, and embarrassment may lead to causes as a first step toward remedial action.
behavior that produces tempo Most important of all, however, is the forward
look the process of defining the problem and identify-
ing courses of action that may solve iŁ Here also the
reward system is critically important. Readiness to

62 S: I
search problem situations and effectiveness in finding Looking to the Future
them need to be recognized and rewarded. Perhaps the
greatest influence a manager can have on the problem- With respect to the initiation of activity, the orga-
solving style of the organization as a role nizational habit we would like to instill is responsive-
model is making the best responses to ness to cues that signal future dit1icu1ties as well as
problems. The style the manager should aim for to those that call attention to the problems of the
rests on the following principles: moment. Failure to give sufficient attention to the
1.Solving the problem takes priority over looking future most often stems from" two' causes. The first is
backward to its causes. Initially, backward looks interruption by current problems that have more
are limited to diagnosing causes; fixing proximate dead1ines and hence seem more urgent; the
responsibility for mistakes should be postponed until second is the absence of sufficient "scanning" activity
a solution is being implemented. that can pick up cues from the environment that long-
2.The manager accepts personal responsibility for run forces not impinging immediately on the
finding and proposing solutions instead of seeking to organization have importance for it in the future.
shift that responsibility either to superiors or to sub- In neither case is the need for sensitivity to the
ordinates, although the search for solutions may, of future likely to be met simply by strengthening intu-
course, be a collaborative effort involving many itions. Rather, what is called for is de1iberate and
people. systematic allocation of organizational resources to
3.The manager accepts personal responsibility for deal with long-range problems, access for these
implementing action solutions, including securing resources to appropriate input from the environment
the necessary authority from above if required. that will at their attention to new prospects, and
4.When it is time to look backward, fixing b1ame may protection of these planning resources from absorption
be an essential part of the process, but the primary in current problems, however urgent they may be.
focus of attention should be on what can be Attention to the future must be institutionalized; there
learned to prevent similar problems from arising in is no simpler way to incorporate it into managerial
the future. "style" or habit.
These principles are as obvious as the Ten Com- It is a fallacy to contrast "ana1ytic" and "intuitive"
mandments and perhaps not quite as difficu1t to obey. styles of management. Intuition and judgment-at least
Earlier, I indicated that stress might cause departures from good judgment-are simply ana1yses frozen into habit
them, but failure to respond effectively to problems and into the capacity for rapid response through
probably derives more from a lack of attention and an recognition. Every manager needs to be able to analyze
earlier failure to cu1tivate the appropriate, The military problems systematically (and with the aid of the
makes much use of a procedure called “Estimate of the modem arsenal of ana1ytical tools provided by
Situation. " Its value is not that it teaches anything management science and operations research). Every
esoteric, but that through continual training in its use, manager needs also to be able to respond to situation
commanders become habituated to approaching situations rapid1y, a skill that requires the cu1tivation of intuition
in orderly ways, using the checklist provided by the and judgment over many years of experience and
formal procedure. Habits of response to problems are training. The effective manager does not have the
taught and learned both in the manager's one-on-one luxury of choosing between "ana1ytic" and "intuitive"
conversations with subordinates and in staff meetings. Is approaches to problems. Behaving like a manager
attention brought back repead1y to defining the problems
means having command of the whole range of manage-
until everyone is agreed on just what the problem is? Is
ment skills and applying them, as they become
attention then directed toward generating possible solution
and evaluating their consequences? The least often
challenged and most reliable base of managerial influence Herbert A. Simon is Richard King Mellon
is the power to set the agenda, to focus attention. It is one Unive1'3ity Professor of computer' science and
of the most effective tools the manager has for training psychology at Camegie-Mellon Unive1'3ity,
organization members to approach problems where he has taught since 1949. For the past 30
constructively by shaping their own habits of attention. years he has been studying decision-making and
The perceptive reader will have discerned that “shaping problem-solving processes, using computers to
habits of attention" is identical to "acquiring simulate human thinking.
intuitions”. The habits of responding to problems by Dr. Simon was educated at the University of
looking for solutions can and must become intuitive – Chicago (where he received his Ph.D.). He was
caused by the presence of the problem itself. A problem- elected to the National Academy of Sciences in
solving style is a component of the set of intuitions that the 1967, and has received awards for his research
manager acquires one of the key components of effective {ram the American Psychological--Association,
managerial behavior. the American Economic Association, and the
lnstitute of Electrical and Electronic