E S S A Y
Decision Making:
It's Not What You Think
Sometimes decisions defy purely step-by-step logic. To be effective, companies
also should embrace intuitive or action-oriented forms of decision making.
Henry Mintzberg and Frances Westley
How should decisions be made? Easy,
we figured that out long ago. First
define the problem, then diagnose its
causes, next design possible solutions,
and finally decide which is best. And,
of course, implement the choice.
But do people always make decisions
that way? We propose that this
rational, or "thinking first," model of
decision making should be supplemented
with two very different
models — a "seeing first" and a
"doing first" model. When practicing
managers use all three models,
they can improve the quality of their
decisions. Healthy organizations,
like healthy people, have the capacity
for all three.
Consider how a real decision was made, a personal
one in this case. It begins with a call from an aunt.
"Hi, kiddo. I want to buy you a housewarming present.
What's the color scheme in your new apartment?"
"Coior scheme? Betty, you've got to be kidding. Til
have to ask Lisa. Lisa, Betty wants to know the color
scheme of the apartment."
"Black," daughter Lisa says.
"Black? Lisa, I've got to live there."
"Black," she repeats.
A few days later, father and daughter find themselves
in a furniture store. They try every desk, every
chair: Nothing works. Shopper's lethargy sets in. Then
Lisa spots a black stool: "Wouldn't that look great
against the white counter?" And they're off. Within an
hour, they have picked out everything — in black,
white and steel gray.
The extraordinary thing about this ordinary story
is that our conventional theories of decision making
can't explain it. It is not even clear what the final decision
was: to buy the stool; to get on with furnishing an
apartment; to do so in black and white; to create a new
lifestyle? Decision making can be mysterious.
The Limits of "Thinking First"
Rational decision making has a clearly identified
process: define -*• diagnose -*• design •*• decide.
However, the rational approach turns out to
be uncommon.
Years ago, one of us studied a host of decisions,
delineating the steps and then laying them out. A
illuiiratlon:fflJonnS.
SPRINC 2001 MIT SLOAN MANACEMENT REVIEW 89
Insight: Groping Precedes Zeroing In Choices Looking for Problems
t
Insight 1 Insight 2 I
dcti^itin process for building a new plant was typical. The
process kept cycling back, interrupted by new events, diverted
by opportunities and so on, going round and round until finally
.1 solution emerged. The final action was as clear as a wave
breaking on the shore, but explaining how it came to be is as
liLird as tracing the origin of that wave back into the ocean.
Often decisions do not so much emerge as erupt. Here is how
Alexander Kotov, the chess master, has described a sudden
insight that followed lengthy analysis:
"So, I mustn't move the knight. Try the rook move again....
At this point you glance at the clock. 'My goodness! Already 30
minutes gone on thinking about whether to move the rook or
the knight. If it goes on like this you'll really be in time trouble.'
And then suddenly you are .struck by the happy idea —
why move rook or knight? What about B-QNl? And without
any more ado, without analysis at all, you move the bishop.
lust like that."
Perhaps, then, decision making means periods of groping
followed by sudden sharp insights that lead to crystallization, as
A. Limgley and co-authors suggested in a 1995 Organizational
Science article. (See "Insight: Groping Precedes Zeroing In.")
Or perhaps it is a form of "organized anarchy," as Stanford
professor James March and colleagues have written. They characterize
decision making as "collections of choices looking for
problems, issues and feelings looking for decision situations in
which they may be aired, solutions looking for issues to which
iliey might be an answer, and decision makers looking for
work." (See "Choices Looking for Problems.")
But is the confusion, as described by those authors, in the
process, or is it in the observers? Maybe messy, real-life decision
making makes more sense than we think, precisely because so
iiuich of it is beyond conscious thought.
"Seeing First"
Insight — "seeing into" — suggests that decisions, or at least
actions, may be driven as much by what is seen as by what is
thought. As Mozart said, the best part about creating a symphony
was being able to "see the whole of it at a single glance in
my mind." So, understanding can be visual as well as conceptual.
In W. Koehler's well-known 1920s experiment, an ape
struggled to reach a banana placed high in its cage. Then it
saw ihe box in the corner — not just noticed it, but realized
what could be done with it — and its problem was solved.
Likewise after Alexander Fleming really .sciwthe mold that had
killed the bacteria in some of bis research samples (in otber
words, when he realized how that mold could be used), he
and his colleague were able to give us penicillin. The same can
be true for strategic vision. Vi.sion requires the courage to see
what others do not — and that means having both the confidence
and the experience to recognize the sudden insight for
what it is.
A theory in Gestalt psychology developed by G. Wallas
in the 1920s identifies four steps in creative discovery;
preparation *• incubation -*• illumination -*• verification.
Preparation must come first. As Louis Pasteur put it,
"Chance favors only the prepared mind." Deep knowledge,
usually developed over years, is followed by incubation, during
which the unconscious mind mulls over the issue. Then
with luck (as with Archimedes in the bathtub), there is that
flash of illumination. That eureka moment often comes after
sleep — because in sleep, rational thinking is turned off, and
the unconscious has greater freedom. The conscious mind
returns later to make the logical argument. But that verification
(reasoning it all out in linear order for purposes of elaboration
and proof} takes time. There is a story of a
90 MIT SLOAN MANAGEMENT REVIEW SPRING 2001
mathematician who solved a formula in his sleep. Holding it in
his mind's eye, he was in no rush to write it down. When he
did, it took him four months!
Great insights may be rare, but what industry cannot trace its
origins to one or more of them? Moreover, little insights occur
to all of us all the time. No one should accept any theory of
decision making that ignores insight.
"Doing First"
But what happens when you don't see it and can't think it up? Just
do it. That is how pragmatic people function when stymied: They
get on with it, believing that if they do "something," the necessary
thinking could follow. It's experimentation — trying something
so that you can learn.
A theory for "doing first," popularized in academia by
organizational-behavior professor Karl Weick, goes like this:
enactment -*' selection -*• retention.
That means doing various things, finding out which among
ihem works, making sense of that and repeating the successful
behaviors while discarding the rest. Successful people know that
when they are stuck, they must experiment. Thinking may drive
doing, but doing just as surely drives thinking. We don't just
think in order to act, we act in order to think.
Show us almost any company that has diversified successfully,
and we will show you a company that has learned by doing, one
whose diversification strategy emerged through experience.
Such a company at the outset may have laid out a tidy strategy
on the basis of assessing its weaknesses and strengths (or, if after
1990, its "core competenverbal
(comprising words in linear order), the second is visual,
the third is visceral. Those who favor thinking are people who
cherish facts, those who favor seeing cherish ideas and those
who favor doing cherish experiences. (See "Characteristics of
the Three Approaches to Making Decisions.")
We have for some years conducted workshops on the three
approaches with midcareer managers sent by Asian, European
and North American companies to our International Masters
Program in Practicing Management {www.impm.org). We
begin with a general discussion about the relationship between
analysis, ideas and action. It soon becomes evident that practicing
managers recognize the iterative and connected nature of
those elements. We then ask small groups first to discuss an
issue for about an hour (one of their own or else what we call a
"provocative question." For example: "How do you manage customer
service when you never see a customer?" or "How do you
organize without structure?"), summarize their conclusions on
a flip chart and report back to the full group. Next we give the
groups colored paper, pens, scissors and glue. Each small group
must create a collage about the issue they discussed in the
thinking-first session. At the end of that second workshop, the
groups view one another's images and compare "seeing first"
with "thinking first" — in terms of both process and results.
Finally, each group, with only a few minutes of preparation time
permitted, improvises a skit to act out its issue. Again, the
groups consider the results.
Reactions to the approaches are revealing. Participants note
that in the thinking-first workshop, the initial discussions start
Thinking first" features
the qualities of
science
planning, programming
the verbal
facts
cies"), which it almost certainly
got wrong. How can
you tell a strength from a
weakness when you are entering
a new sphere? You have
no choice but to try things
out. Then you can identify
the competencies that are
really core. Action is important;
if you insist on "thinking
first" and, for example, doing
formalized strategic planning (which is really part of the same
thing), you may in fact discourage learning.
Making Decisions Through Discussion,
Coiioge ond Improvisotion
Thus the three major approaches to decision making are "thinking
first," "seeing first" and "doing first." They correlate with
conventional views of science, art and craft. The first is mainly
Characteristics of the Three Approaches to Making Decisions
"Seeing first" features
the qualities of
art
visioning, imagining
the visual
ideas
"Doing first" features
the qualities of
craft
venturing, learning
the visceral
experiences
off easily enough, no matter what the mix of nationalities or
work backgrounds. Participants list comments on flip charts and
spontaneously use bulleted items and numbers ™ with the occasional
graph thrown in. Almost no time is spent in discussing
howto go about analyzing the problem. Groups quickly converge
on one of several conventional analytic frameworks: cause and
effect, problem and solution, pros and cons, and so on.
Many participants observe that such frameworks, particuSPRING
2001 MIT SLOAN MANAGEMENT REVIEW 9 1
liiily when adopted early, blunt exploration. Quality and depth
o\ AUii\ys\s may be sacrificed for process efficiency. Thinkingllrsl
workshops encourage linear, rational and rather categorical
iU'guinciils. All too often, the result is a wish list, with disagreements
hidden in the different points. In other words, there may
he less discipline in thinking first than
wf believe. riiiiikiiiL; conies too easily
U> most ol us.
But when a group must make a picuiie,
members have to reaeh consensus.
Ihal requires deeper integration
of the ideas. "We liad to think more to
do this," a participant reported. The
artistic exercise "really forces you to
capture ihe essence of an issue,"
another added. People ask mure quesliuns
in ihe seeitig-llrst exercise; they
Ixxome more playful and creative.
"In 'thinking first,' we foeused on
ihe [iKiblems: in 'seeing first,' we
Ineuscd on ihe solutions," one person
said. One group believed it had agreemenl
on ihe issue after the ihinkingInst
workshop. Only when the picture
making begim did its members realize
hmv superficial thai agreement was —
mine tif a cnmproniise. In contrast,
when you really do see, as someone
said. "The message jumps out at you."
lUit lo achieve ihal. the group members
have lo luui out more about one
another's capabilities and collaborate
uiorc closely. "I felt il became a group project, not just my project,"
said a patiicipant who had chosen the topic for his group.
Ihe seeing-first exercise also draws out more emotions; there is
more laughter and a higher energy level. This suggests that being
ahle to see a trajectory — having a vision about what you are
doing — energi/es people and .so stimulate.s action. In comparing
ihe .seeing-first exercise with the thinking-first discussion, a parliiipaiit
remarked, "We felt more liberated." The pictures maybe
more ambiguous iban the words, but they are also more involving,
A liet|ueiit conunent:" They invite interpretation."
One particularly interesting observation about the pictures
was ihal "ihe impression lasts longer." Studies indicate that we
remember pictures much longer and more accurately than
words. As R. Haber demonstrated in Scientific American in
1470, recall of images, even as many as 10,000 shown at one.second
Intervals, is nearly 98% — a capability that may be
In the "seeing first" workshop,
midoareer managers find making a
picture of problems required deeper
integration of ideas.
linked to evolution. Humans survived by learning to registei'
danger and safety signals fast. Emotion, memory, recall and
stimulation are powerfully bundled in "seeing first." (^ontrasl
that with one comment after the thinking-first workshop:
"Twenty-four hours later, we won't remember what this meant."
In fact, although many participants
have not made a picture since
grade school, the art produced in the
seeing-first workshops is t)ften
remarkable. Creativity fiows freely
among the managers, suggesting that
they could come up with more creative
ideas in their home organizations
it they more often used symbols
beyond words or numbers.
Our multicultural groups may like
the art workshop for overcoming language
barriers, but groups of managers
from tbe same company,
country or language group have
responded equally well, t^ne British
participant who was working on a
joint venture with an American partner
found that out. He met with his
U.S. counterpart a few days after the
workshops. "We talked past each other
for two hours," he reported. When he
suggested they create a picture of their
common concerns, they finally were
able to connect.
The improvisation skits — "doing
first" — generate more spontaneity.
Participants respond to one another intuitively and viscerally,
letting out concerns held back in conversation and even in
artwork. For example, turf battles become evident in the way
people stand and talk. Humor, power, fear and anger surface.
(M. Crossen and M. Sorrenti discuss improvisation at lenglh
in a helpful article published in 1997 in Advances in
Strategic Management.)
Weick has suggested that a key aspect of effective action in
organizations is the ability to remain open to signals from others,
even under extreme pressure. He believes that such heedfulness,
as he calls it, is a finely honed skill among groiiji
improvisers such as jazz musicians. Organizations that recognize
opportunities for improvisation — and hone the skills required
— increase their capacity for learning. In improvisation, people
have to respond with a speed that eliminates many inhibitions.
"Having to just act gets rid of the fears," a participant said.
92 MM SLOAN MANAGEMENT REVIEW SPRING 2001
Anoihcr iultled, after watching a colleague play the role of a frustrated
bank customer, "The output can be scarily real."
Mere words, in contrast, feel more abstract and disconnected
— numbers, even more so — just as the aggregations of
marketing are more abstract than the experience of seUing. The
.skits bring out what the words and numbers do not say —
indeed, what problems they cause. "Not everything is unsayable
ill words," claimed playwright Eugene Ionesco, "only the living
truth." Or as Isadora Duncan, the modern-dance pioneer,
insisted, "If I could say it, I wouldn't have to dance it." Thus
"doing first" facilitates the dancing that is so lacking in many of
today's organizations.
Enough Thinking?
Ihe implications for our large, formalized, thinking-obsessed
organizations are clear enough: not to suspend thinking so
much as put it in its place, alongside seeing and doing. Isn't it
time we got past our ob.session with planning and programming,
and opened the doors more widely to venturing and
visioning? A glance at corporate reports, e-mail and meetings
reveals that art is usually something reserved for report covers
— or company walls. And when organizations separate the
thinking from the doing, with the former coming from the
heads of powerful formulators and the latter assigned to the
hands of ostensibly docile implementers, those formulators lose
the benefits of experimenting — and learning.
Each approach has its own strengths and weaknesses.
(See "When Each Decision-Making Approach Works Best.")
"Thinking first" works best when the issue is clear, the data
reliable and the world structured; when thoughts can be
pinned down and discipline applied, as in an established production
process. "Seeing first" is necessary when many elements
have to be combined into creative solutions and when
commitment to those solutions is key, as in much new-product
development. The organization has to break away from the
conventional, encourage communication across boundaries,
bust up cerebral logjams and engage the heart as well as the
head. "Doing first" is preferred when the situation is novel and
confusing, and things need to be worked out. That is often the
case in a new industry — or in an old industry thrown into
turmoil by a new technology. Under such circumstances, complicated
specifications get in the way, and a few simple relationship
rules can help people move forward in a coordinated
yet spontaneous manner.
That suggests the advantages of combining all three
approaches. In order to learn, a company group might tackle a
new issue first by craft, which is tied to doing; then, in order to
imagine, by art, which is tied to seeing; finally, in order to proWhen
Each Decision-Making
Approach Works Best
•THINKING FIRST" WORKS BEST WHEN:
• the issue is clear;
• the data is reliable:
• the context is structured;
• thoughts can be pinned down: and
• discipline can be applied
as in an established production process.
"SEEINC FIRST" WORKS BEST WHEN:
• many elements have to be combined into
creative solutions:
• commitment to those solutions is key: and
• communication across boundaries is essential
as in new-product deveiopment.
"DOING FIRST" WORKS BEST WHEN:
• the situation is novel and confusing:
• complicated specifications would get in the
way; and
• a few simple relationship rules can help
people move forward
for example, when companies face a disruptive
technology.
gram, by science, which is tied to thinking. In ongoing sitiui
tions, art provides the overview, or vision; science specifies the
structure, or plan; and craft produces the action, or energy. In
other words, science keeps you straight, art keeps you intei'
ested, and craft keeps you going. No organization can ilo willi
out any one approach. Isn't it time, then, to move beyond our
narrow thinking about decision making: to get in !oiuh, to sec
another point of view?
Henry Mintztxrg and Frances Westley are professors oi
management at McGilt University in Montreal. Contact them al
mintzber@management.mcgill.ca and westley®managemenl.mcgill.ca.
Reprint 4238
Copyright ©2001 by the Massachusetts Institute of Technology. All riglui rcscrvcil.
2001 MIT SLOAN MANAGEMENT REVIEW 93