Marketing Letters 5:4.

(1994): 413-426
1994 Kluwer Academic Publishers. Manufactured in ihe Netherlands.

Understanding Managers' Strategic Decision-

Making Process
WILLIAM BOULDING'
Fuqua School of Business, Duke University, Bo.v 90121, Durhatu, NC 27708-1020

MARIAN CHAPMAN MOORE

RICHARD STABLIN
KIM P. CORFMAN
PETER REID DICKSON
GAVAN FlTZSiMONS
SUNIL GUPTA
DONALD R. LEHMANN
DEBORAH J. MITCHELL
JOEL E. URBANY
BARTON A. WEITZ

Key words: strategic choices, managerial decision making, decision errors, menial models

Abstract

This goal of this paper is to establish a research agenda that will lead lo a stream of research that
closes Ihe gap bclwcen actual and normative strategic managerial decision making. We start by
distinguishing strategic managerial decision making (choices) from other choices. Next, we pro-
pose a conceptual model of how managers make strategic decisions ihat is consistent with the
observed gap between actual and normative decision making. This framework suggests a series of
interesting issues, both descriptive and prescriptive in nature, about the strategic decision-making
process that define our proposed research agenda.

Is it possible for managers within organizations to have widely varying, and in-
correct, beliefs about market facts? Is it possible that consumers' utility for a
product and managers' perceptions (of consumer perceptions) of quality for that
product are inversely related? Is it possible that managers react to competitors
without considering customer behavior implications or competitors' subsequent
reactions? Is it possible that increasing the frequency or quantity of available in-
formation to managers about customer response leads to increasingly bad deci-
sions? According to empirical evidence, the answer to all these questions, at least
in some instances, is yes.^ It is these fmdings, along with other similarly puzzling
examples of managerial behavior, that lead us to believe there is great value in
better understanding managers' strategic decision making process.
The goal of this paper is to establish a research agenda that will ultimately lead
to a stream of research that closes the gap between actual and normative mana-
gerial behavior. The core of this paper is a conceptual model of how managers
414 WILLIAM BOULDING ET AL.

make strategic decisions that is consistent with the patterns of managerial deci-
sions mentioned above. This framework suggests a series of interesting issues,
both descriptive and prescriptive in nature, about the strategic decision making
process that defme our proposed research agenda.

1. What is strategic?

Our interest centers on the process of strategic managerial decision making. This
process is distinct from the often studied consumer choice problem (e.g., selecting
a particular brand of toothpaste) or a mundane managerial decision (e.g., where
to hold the company picnic). Strategic managerial decisions are typically "mes-
sier," i.e., they occur in a more complex environment, are difficult or expensive
to reverse, and the outcomes are, to a greater degree, contingent on other individ-
uals' or organizations' behavior. Moreover, strategic decisions often substantially
alter (and irrevocably so in the short run) the relationship between the manager's
organization and that organization's customers or competitors.
One characteristic of a strategic managerial decision that differentiates it from
other types of decision making is the particular setting or context. Figure 1 rep-
resents an abstraction of this context, in the form of the 3 Cs model. Managers
make strategic decisions within an organizational context, i.e., the company, and
these decisions affect a number of other players. Moreover, the reaction of these
other players normally affects the final decision outcome. These other players
include customers (with a further complicating factor that the company's cus-
tomer is often a channel intermediary that resells to an end customer) and com-
petitors (with a complicating distinction between current and potential competi-
tors), as well as broader publics such as society, regulators, and investors (i.e.,
the "world"). Thus, the viability of managers' strategic decisions depends in large
part on managers' knowledge about the current status and probable reactions of
their company, competitors, customers, and broader publics.

2. Overall conceptual framework

Figure 2 provides our conceptualization of how a manager makes decisions within

this strategic context. One should view the model in Figure 2 as completely nested
within Figure I. The key features of our model are (1) the decision trigger mech-
anism, (2) the manager's overall mental market model, which includes decision
rules and mental models of market and context response, (3) the decision, (4) the
actual market response function, (5) the recognition that context (i.e., the ele-
ments delineated in Figure 1) is pervasive in the decision-making process and can
interact with any of the elements of Figure 2, and (6) that learning occurs from
outcomes of the decision process. We briefly discuss these components to provide
an overview of the model.
UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 415

Figure I. The strategic context for managerial decision making

2.1. Elements of the framework

Strategic decision making requires a trigger to start (he process. There are many
possible triggers, such as an observed decrease in sales, a new product entry by
a competitor, a request for action from a higher level manager, or simply a regular
planning cycle. In addition, the outcome of a prior decision could itself trigger a
new decision making process.
The heart of our framework is the individual manager's mental model of the
context and market responses, represented by the functions ^,+ i = fi{X,, ^,) and
416 WILLIAM BOULDING ET AL.

Trigger
Mechanism

Overall Mental Model

Context Response
Models

Decision
Rules
Market Response
Models

"True" Market Model

Figure 2. A conceptual model of managerial decision making

UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 417

,, ^,), respectively. In this formulation ^,+ | represents a vector of

anticipated outcome measures (e.g., sales, profits, etc., at time t -\- I); X, is a
vector of possible actions that the manager can take (e.g., setting a new actual
price); 1, is a vector of the manager's estimates of factors that the manager be-
lieves are affected by his actions (e.g., cu^Xom^r perceptions of own-price, com-
petitors' prices, etc.), which in turn affect the outcome measures; and Y, are all
other factors that might affect S,^^. Finally, ft (•) and g (•) represent the manag-
er's mapping of the context and market response functions, respectively. Implicit
in this tnental market response model is what we label "what" and "how" knowl-
edge. "What" knowledge determines the values that managers assign the vectors
within the mental market response model, i.e., market facts. The "how" knowl-
edge determines the manager's mapping function, g (•), and describes how the
manager believes the market behaves. This "how" knowledge constitutes the
choice of relevant variables, (e.g., the Zs and the Ts), the functional form of the
relationship (e.g., linear, multiplicative, etc.), and the coefficients (e.g., price sen-
sitivity, etc.).
"What" and "how" knowledge is also implicit in the context response model.
The "what" knowledge consists of managers' beliefs about current levels of Z,,
while the "how" knowledge is reflected by the mapping function ft (•). Presum-
ably, the "what" and "how" knowledge embedded in the context and market
response models is generated from managers' prior experiences.
When faced with the task of making a strategic decision, the manager may use
these mental response models in concert with a set of decision rules to arrive at
a particular decision. For example, a manager may observe a price change by a
competitor. In some instances, a manager may invoke a decision rule that requires
that the manager have some concept linking the firm's actions with possible out-
comes. An example of such a decision rule is "choose the optimal values from
the set of possible actions found in X,, conditional on the new competitor price."
Alternatively, the manager's decision rule may be of the type "if the competitor
changes price, then we will match it," which can be executed without ever refer-
ring to the mental response model. The former decision process requires consid-
erably more cognitive effort and requires the manager to know the how's and
what's of a situation (e.g., the manner in which X, "causes" ^f+,, which in turn
affects ^,yi). The latter decision rule, however, only requires that the manager
knows to react. We refer to this latter knowledge as "if-then" knowledge. Note,
however, that this if-then knowledge is not immaculately conceived. If-then
knowledge must be at least implicitly generated by some manager's "how" knowl-
edge.
Following the flow in Figure 2, a decision is then made. The results of the
decision, .S',^ ,, are due to the "true" market response model, ^(•), which may not
match the manager's mental market model, § {•). Our conceptualization suggests
at least three major reasons why the manager's prediction (^,^ ,) can differ from
what actually happens {S, ^ |). First, the manager may have the wrong mental mar-
ket model (e.g., the model contains the wrong variables, the wrong functional
418 WILLIAM BOULDING ETAL.

form, etc.). Omitted variables (either unobserved or ignored by the manager) are
represented with the vector U in the "true" market model. Second, the manager
may have incorrectly estimated others' (e.g., customers, competitors, etc.) reac-
tions to the chosen decision variables (e.g., t., ,^, may differ from Z, +, because the
manager has a poor or non-existent h (•) function). Third, the included variables
in the manager's model g (•) may not have the correct "weights" or the correct
functional form (i.e., g 5* g). All three of these factors could play a significant
role in explaining managerial decision errors and observed anomalies in strategic
decision making.

22. The role of context

It is important to note that, in Figure 2, context (i.e.. Figure I) is all pervasive

and interacts with every component of this model. Context can clearly affect the
trigger mechanism. As noted above, actions on the part of customers, competi-
tors, or the company could all trigger the decision process. As examples, cus-
tomers could demand new products or services, competitors could take threat-
ening actions, or top management could provide the manager with new strategic
imperatives (e.g., increase market share). The relative sensitivity ofthe trigger
mechanism to these contextual factors is an interesting, and unresolved, research
question. Based on limited empirical and anecdotal evidence, we conjecture that
the trigger mechanism is most sensitive to competitive behavior but speculate that
it should be equally sensitive to customer and company behavior.
Context also influences the overall mentat model ofthe manager. The manager
organizes the history of behaviors of the company, competitors, and customers
into a set of current beliefs. In turn, this belief structure is codified into the indi-
vidual manager's mental model. Ultimately, this codification is driven by the in-
dividual manager's ability, goals, aspirations, etc. Thus, our conceptual model of
the decision making process explicitly accounts for individual differences. How-
ever, changes in the state of the individual manager's perceived world, as evi-
denced by new competitor, customer, or company behaviors (e.g., new organi-
zational structures/compensation systems), also lead to adjustments in the
manager's mental model. Currently salient examples of such changes include the
tendency of firms to move from individual-based decision making to group-based
decision making, and time compression in the decision making process brought
on by an increased rate of industry change. In these instances managers may come
to increasingly rely on if-then decision rules, and "group" decision rules, unsup-
plemented by their own mental response models, to speed their decision making.
Context is also directly incorporated into the "true" market model. For exam-
ple, competitive reaction will be contingent on the manager's decision. The out-
come of the action/reaction sequence appears exphcitly in the "true" market
model.
UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 419

2.3. The role of learning

Our model acknowledges that managers may "learn" from the decision process.
This is captured by the feedback loops from the true market model (i.e., actual
outcomes) to the trigger mechanism and the manager's overall mental model.
These feedback loops reflect the dynamic nature of strategic decision making.
The use of the word "learn" does not imply that the ongoing nature of strategic
decision making necessarily leads to better decisions over time. "Good" learning
would be reflected by the development of efficient decision rules, increasing pre-
cision in the two types of mental response model predictions, and perhaps even
changes in the market response objective function (e.g., a shift from market share
to profit as the objective). However, our model does not preclude the possibility
of "bad" learning. In fact, the messy nature of strategic decisions leads to the
very real possibility that managerial knowledge deteriorates rather than increases
over time. This deterioration in knowledge could occur because of, among other
reasons, incorrect causal attributions or overgeneralizing a correctly identified
relationship. The former problem is likely due to the age-old problem of spurious
correlation. The latter problem is perhaps best exemplified by behavior in the
airline industry. Specifically, a growing number of airlines have attached their
names, surely at great expense, to professional basketball arenas. Based on this
behavior, one suspects that the airline industry operates on the general, but po-
tentially inefficient, rule "match every competitive action."

2.4. Summary

In sum, the model in Figure 2 decomposes the strategic decision process into
several elements, which should help both researchers and managers. With respect
to managers, considerable evidence exists, both empirical and anecdotal, that
managers sometimes make bad decisions. Based on this evidence we ask not (/
managers make bad decisions, but rather, why managers make bad decisions. Spe-
cifically, can we articulate a process that explains apparent anomalies in mana-
gerial behavior? The model described above is our answer to this question. The
model we propose may help managers articulate, and therefore enhance, their
decision processes. In support of this belief, the decision theory literature (e.g.,
Schlalfer, 1969) tells us that problem decomposition increases the quality of de-
cisions. The model proposed herein accomplishes this decomposition and high-
lights points where errors may enter into the strategic decision making process.
With respect to researchers, the decomposition helps in two ways. First, it pro-
vides a means of linking a number of existing research streams to our particular
problem. Examples include knowledge structure research (e.g., Holland, Hol-
yoak, Nisbett, and Thagard, 1986), decision-making biases (e.g., Russo and
Schoemaker, 1989), managers' market response models (e.g., Chakravarti, Mitch-
420 WILLIAM BOULDING ET AL.

ell, and Staelin, 1981), competitor reaction models (e.g., Hanssens, 1980), mana-
gerial forecasting models (e.g., Clemen and Winkler, 1986), decision support
models (e.g., Lodish, 1971), and the emerging literature on learning organiza-
tions (e.g., see the 1991 Special Issue on Organizational Learning in Organization
Science).
Second, the decomposition helps identify the researchable components of a
very messy process. For example, one might focus on the dynamics that drive
good versus bad learning (Gupta and Steckel, 1993). Another area might be to
devise research methods designed to tap into and delineate managers' mental
models (Mitchell and Russo, 1993). Alternatively, one could look at the influence
of the group on decision-making biases (Janis, 1981; Corfman and Kahn. 1993).
Other possibilities include studying the process of competitive conjecture (Ur-
bany, 1993), or managers' conjectures about customers (Boulding, 1994; Dickson,
Urbany, and Kalapurakal, 1993). A further possibility is to study the trigger mech-
anism to the decision-making process (Moore and Bettman, 1994).
We believe the current research opportunities are close to limitless. The appen-
dix to this paper contains a representative, but far from exhaustive, list of research
questions. We hope that others generate additional research questions. However,
in the interest of starting this process we next turn our attention to the issue of
potential sources of error in strategic decision making, and in so doing generate a
series of testable research propositions.

3. Errors in managers' predictions

Recall that mental market models are implicitly contained in managers' decision
rules. Even when managers rely on a decision rule heuristic (if-then knowledge),
the origin of such rules can be traced to some manager's mental response models
(how knowledge). Consequently, we focus our attention on fallible "how" and
"what" knowledge. We make the implicit assumptions that errors in the response
models are negatively correlated with the quality of decisions, all else equal. More
formally, we propose

P|. Managers with better "how" and "what" knowledge make better decisions.

While this proposition seems noncontroversial, it can be argued that the process
of obtaining better "how" and "what" knowledge can hinder a manager's ability
to act, i.e., "paralysis of analysis." However, if P, is true, then we as academics
need to work toward enhancing managers' "how" and "what" knowledge.
Another general concern about errors in strategic decision making is whether
the learning that emerges from the dynamic nature of strategic decision making is
good learning or bad learning. We believe that the size of the prediction error will
be reduced only if the managers have both the ability and motivation to learn.
Therefore, a relevant line of inquiry is to identify the drivers for both ability and
UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 421

II
motivation. We note that, holding fixed the company context, increased external
pressures (e.g., greater competitive rivalry, more demanding customers, etc.)
should lead to decreased abiVity for managers to enhance their "how" and "v^'hat"
knowledge. Conversely, such pressures should lead to increased motivation for
managers to improve this knowledge. Such logic leads us to posit that external
forces lead to a negative observed correlation between motivation and ability
(Boulding and Staelin. 1993) for managers to develop "how" and "what" knowl-
edge, all else equal.
This negative correlation places the firm in a quandary if it wants to build
knowledge over time. Specifically, we propose the following:

Pi. External forces are unrelated to managerial learning since the motivational
effects of the external forces wit! be canceled out by decreases in managers'
ability to learn.

As a corollary to this, we propose the following:

p2A- Good learning will only occur within an organization if the firm provides
motivation from within.

The above discussion leads us to ask the more general question. "When are
managers motivated to learn?" We believe learning is most likely when managers
admit "ignorance" (Zaitman and Staelin, 1993) or "lack of knowledge" (Boulding,
1993). Thus, managers who perceive themselves to be "knowledgeable" will be
less likely to update their mental models over time. However, as the environment
changes, which is a defining characteristic of strategic decision making, this lack
of updating will lead lo increasing levels of prediction errors. The more managers
codify their mental market models into decision rule heuristics, the more serious
the lack of updating will become. Thus, we offer the following proposition:

Pj. Future prediction errors are increasing in the manager's current level of con-
fidence in his or her "what" and "how" knowledge.

We now focus on the three specific sources of the prediction errors identified
above (e.g., specifying the right variables, the correct functional form, and the
specific parameters). Although these sources of error quickly become interre-
lated, we start with the variable specification problem. The potential problem here
is one of spurious correlation; that is. the manager omits the variable that drives
the observed relationship. Further, it seems highly unlikely that, without formal
modeling, managers can control for omitted variable bias. This leads to the prop-
osition that managers cannot disentangle the influence of observed variables from
the unobserved error. More technically:

P4. Managers cannot perform two-stage least squares in their heads.

422 WILLIAM BOULDING ET AL.

A stronger version of this proposition is as follows:

P5. Managers' mental market models are single parameter models in the decision
variables.

This last proposition suggests that without formal modeling and estimation,
managers are certain to generate prediction errors in situations where relevant
omitted variables change over time and are correlated with the decision variable
of interest. Thus, even minimally complex, dynamic decision contexts will lead
to prediction errors.
These prediction errors will occur because the manager will infer biased
weights. Consequently, the "wrong variables" problem blends into the "wrong
weights" problem. One so-called solution is to increase the frequency of market
response feedback. Due to technology advances, the frequency of market feed-
back has increased dramatically in many industries. In static markets where
everything but response and the decision variable remains fixed, increased fre-
quency of feedback will help the manager hone in on the true weights. However,
in even moderately complex or dynamic environments (i.e., the "messy" context
of strategic decision making), increased feedback can lead to increased distortion
in the weights. This happens for at least two reasons. First, the increased time
pressure to make decisions will increase prediction errors. Second, according to
P4 and P5. managers will make incorrect inferences from the market feedback.
Instead of ignoring noise, managers will overreact to noise. Thus, we add another
proposition:

P(,. Increasing the manager's access to market feedback when the manager has a
fallible mental market model leads to increasing prediction errors.

We note that this proposition underscores the importance of identifying the

boundary conditions for when managers' mental models are or are not fallible.
Another potential source of error in developing market response weights
emerges from the behavioral decision theory and social cognition literatures. We
believe that many of the standard biases (e.g., availability bias, false consensus,
and so forth; cf. Russo and Schoemaker, 1989; Moore and Urbany, 1994) hold in
the context of strategic decision making. Other than suggesting the applicability
of these biases, we leave discussion of these effects, and the numerous related
propositions, to others.
We also acknowledge the importance of research directed toward understanding
competitor actions and reactions. Of particular interest are the weights managers
attach to competitor actions in their market response models. Based on limited
empirical and anecdotal evidence we propose the following:

P7. Managers overweight market response to competitor actions.

UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 423

We conjecture that this overweighting is due to the potential impact of competitor

actions on the firm, especially in the short run. It is usually the case that one
competitor can do more damage to the manager's firm, more rapidly, than a single
customer can. As a result, the salience of competitive actions results is quite dra-
matic. A corollary of overweighting the effect of competitive actions on market
response is as follows:

Pfi. Managers overreact to competitive actions relative to customer actions.

These propositions are consistent with our earlier conjecture that the trigger
mechanism in our process mode! is more sensitive to competitive behavior than
customer behavior.
Managers' conjectures about competitive behavior is another potential source
of prediction error in mental market models. A logical extension of P^ is

P9. Managers are better able to predict competitor response to an action than cus-
tomer response.

The logic supporting this proposition is that managers can effectively predict com-
petitive response via self-analysis. In particular, we believe that the manager and
his or her company are typically more similar to other competitors in the industry
than to the firm's customers. Managers' self-attributions to competitors, there-
fore, should be more accurate (and more likely) than self-attributions to cus-
tomers, resulting in more accurate predictions of competitor behavior than cus-
tomer behavior. Notice that the tendency to expect competitors to act the same
as the manager's firm would is a potentially important and even disastrous judg-
mental bias.
In contrast, it is also possible to justify the alternative proposition:

P9A Managers are better able to predict customer response to an action than com-
petitor response.

Support for the alternative proposition rests in the limited empirical and anecdotal
evidence suggesting that managers show little foresight in considering competitive
reactions to their decisions. Further, one might argue that the decision processes
for customers are sometimes more visible and accessible to the manager than
those of competitors. We suspect that whether P, or P^^ holds depends on this
issue of relative accessibility of decision processes.

4. Discus»on

Our intent is not to portray managers as poor decision makers or in any way be
critical of managers. We believe errors in strategic managerial decisions exist be-
424 WILLIAM BOULDING ET AL.

cause of the complexity and pressure of the decision environment. Often the sim-
plifying heuristics that managers employ are a rational response to an overwhelm-
ing task, decision errors notwithstanding. Nevertheless, we believe that managers
can improve their strategic decision making skills. Outlining and decomposing the
decision making process is a first step. Understanding the components of the pro-
cess allows managers to begin attacking the errors during the process. The same
decomposition serves to delineate a research agenda for researchers. We hope
this agenda leads to increased interest in, and understanding of, strategic decision
making.
An implicit part of this research agenda should be of mutual interest to both
academics and managers. Specifically, how do we move from mere description of
the process (i.e., this paper) to prescription? We reiterate the belief stated in our
first proposition that sophistication in "what" and "how" knowledge both matters
and helps. Only with such knowledge can one devise the right decision rules. In
contrast to this statement, we offer our final proposition:

P,o. Managers' decisions are more typically based on "if-then knowledge" (deci-
sion rules) than "how knowledge" (market response models).

Thus, driving errors out of the decision process Implies that managers need to
substantially alter their decision making behavior. Further, because of the dy-
namic setting, managers must possess a constant capacity for updating the deci-
sion making process. Therefore, a key issue to consider in future research is how
to get managers to repeatedly unfreeze and update their mental models.
Academics probably possess comparative advantage, relative to managers, in
terms of delineating the how's of market response. The question still remains as
to whether such knowledge leads to better decision making. If we hope to move
from a descriptive model of strategic managerial decision making to a normative
model, we need to first be sure that our own models are not flawed. Then we must
bridge the gap between academic model building and current managerial strategic
decision making practices. Building this bridge requires receptivity to change on
the part of both managers and academics.

Notes

1. The authors comprise Group 9 of the 1993 Duke Invitational Symposium on Choice Modeling
and Behavior. The authors ihank Joel Huber for his helpful comments.
2. For example, see Dickson. Urbany. and Kalapuraki (1993) for incorrect market facts. Boulding
(1994) for managers' misperceptions of customers' perceptions, Armstrong. Colgrove. and Col-
lopy (1992) for overreaction Io competitors, Moore and Urbany (1994) for lack of foresight in
considering competitive reaction, and Gupta and Steckel (1993) for more information leading to
poorer decisions.
UNDERSTANDING MANAGERS' STRATEGIC DECISION-MAKING PROCESS 425

Reference

Armstrong. J. Scott. R. Colgrove. and E. Collopy. (1992). "Competitiveness and Performance:

Impact of Objectives on Firms' Long-Term Profitability." Working Paper. The Wharton School,
University of Pennsylvania
Boulding, William. (1993). Traps lo Avoid in Becoming Customer Driven." Summary Worksheets,
Fuqua School of Business. Duke University.
. (1994). "Is Consumer Utility Increasing in Managers' Perceptions (of Customers" Percep-
tions) of Quality?" Working Paper. Fuqua School of Business. Duke University.
Boulding, William, and Richard Staelin. (1993). "A Look on the Cost Side: Market Share and the
Competitive Environment,'" Marketing Science, 12(2), 144-166.
Chakravarti. Dipankar. Andrew Mitchell, and Richard Staelin. (1981). "Judgment Based Market-
ing Decision Models: Problems and Possible Solutions," Journal of Marketing, 45 (Fall). 13-23.
Clemen. Robert T.. and Roberl L. Winkler. (1986). "Combining Economic Forecasts." Journul of
Biisine.s.'i and Economic Sltitistici, 4( I), 39—46.
Corfman. Kim P.. and Barbara H. Kahn. (1993). "Availability Bias in Groups: Are Two Heads
Better Than One.'" Working Paper. Slern School, New York University.
Dickson. Peler R., Joel E. Urbany, and Rosu Kalapurakat. (1993). "Decision-Makers' Mental
Models of Consumer Beliefs and Behavior." Paper presented at the Duke Invitational Sympo-
sium on Choice Modeling and Behavior.
Gupta, Sunil, and Joel H. Steckel. (1993). "Dynamic Decision Making and Response lo Change in
Distribution Channels: A Research Agenda." Paper presented at ihe Duke Invitational Sympo-
sium on Choice Modeling and Behavior.
Hanssens, Dominique M. (1980). "Market Response, Competitive Behavior, and Time Series Anal-
ysis."' Journal of Marketing Research. 27 (November), 470-485.
Holland. J.H., K.J. Holyoak, R.F. Nisbett, and P.R. Thagard. (1986). /«(/«c//on. Cambridge. MA:
MIT Press.
Janis, Irving R. (1981). GroupThink, 2nd ed. Boston: Houghton Mifflin.
Lodish. Leonard M. (1971). "CALLPLAN: An Interactive Salesman"s Call Planning System,"
Management Science. 18(4), 25-40.
Mitchell, Deborah J.. and J. Edward Russo. (1993). "Assessing Managerial Learning: Methods for
Looking inlo the Black Box." Research Proposal. Marketing Science Institute.
Moore. Marian Chapman, and James R. Beitman. (1994). "Competitive Moves and Countermoves:
The Role of Perceived Threat and Competitive Position.'" Working Paper, Fuqua School of Busi-
ness, Duke University
Moore. Marian Chapman, and Joel E. Urbany. (1994). "Blinders. Fuzzy Lenses, ami ihe Wrong
Shoes: Pitfalls in Competitor Analysis." Marketing Letters, forthcoming.
Russo, J. Edward, and Paul J.H. Schoemaker. (1989). Decision Traps. New York: Doubleday/
Currency.
Schlaifer, Robert. (1969). Analysis of Decisions Under Uncertainty. New York: McGraw Hill.
Urbany, Ji>el E. (1993). "Exploring Competitive Conjecture (or the Lack Thereof)." Working Pa-
per. University of South Carolina.
Zaltman, Gerald, and Richard Staelin. (1993). "Ignorance, Knowledge, Wisdom and Questioning
in the Developmeni of Decision Makers' Theories." Working Paper. Fuqua School of Business,
Duke University.

Appendix

1. What (riggers the initiation of a strategic decision making process?

2. How do managers scan the environment for information essential for strategic decision making?
426 WILLIAM BOULDING ET AL.

How do they decide whether to use the information? If they choose to use it, how does it get
incorporated into their decision model? How do we reliably uncover managers' mental models?
3. How do managers learn? Where do the "rules" in managers' decision models come from? How
much is individual-specific and how much is company-specific?
4. Under what conditions is a manager's mental model (decision rules, market model or context
model) updated or changed? Is the adjustment more sensitive to information based on compet-
itors or information based on customers? How do managers use feedback from the environment?
5. How can we help managers learn the "how" of decision models? What can be done to help
managers recognize and undo biases that pervade their strategic decision making process? What
form of information decomposition or aggregation is most helpful?
6. What aspects of the strategic decision making context affect the differential impact of different
kinds of information? Do certain aspects of context drive the updating or selection of mental
models? Do managers within the same company or the same industry have differing perception
of the same context? If so, how do they differ, why do they difTer, and what are the conse-
quences of the differing views?
7. What is the appropriate benchmark against which to compare managers" mental models and
decision making performance?
8. What are the appropriate decision aids for managers? Do these aids vary by level of managerial
knowledge? by context?