A MODEL OF ORGANIZATIONAL CHANGE
Henry Frechette
Senior Vice President, Client Research
Forum Corporation
Frank Spital
Associate Professor
College of Business Administration
Northeastern University
Abstract
This paper reports on the initial results of a research effort to improve the long run effectiveness of
organizational changes, The literature suggests that organizations frequently experience periods of
stability that are punctuated by bursts of large scale change. Our objective is to understand what causes
this pattern, and what policies are likely to improve organizational performance through the change
process. In this paper we present a causal loop diagram.of the structure of our model, and discuss the
reasons for this structure. This model does reproduce the behavior patterns to which the literature
refers. In addition, we report the results of some sensitivity analyses and policy tests of the model.
These results have implications for managers. Finally, we discuss the limitations of our model in its
current form, and the next steps that we intend to take.
The Problem
The ability to manage charige is an important part of organizational leaming. Increasingly the role
at which an organization responds to changes in its environment (leams) is becoming the most
sustainable competitive advantage of an organization. Understanding and managing the change process
for performance improvement will become a central task of the executive team. A system dynamics
perspective may provide insight into both the most frequent and the most effective patterns of
organizational change.
One of the most frequently cited patterns of organizational change over extended periods of time is
an oscillatory one. The literature hypothesizes that this pattem occurs because there are powerful forces
for stability in organizations, but that this same pressure for stability ultimately creates the need for
change (Tushman, Newman and Nadler, 1989). In general, past business success reinforces the current
strategy, systems and structures, and the longer the success continues, “the more dominate these internal
forces for stability become” (Tushman et. al., p.113). On the other-hand, “the very social and technical
consistency that are the sources of success may also become the seeds of failure if the environment
changes” (ibid). Since we can be assured that the environment of any business will change at some
point, this structure will produce, for any company, an oscillating rate of change variously referred to as
“punctuated equilibria” or the “boom-splat”.
The objective of our current work is to model the key components of a change system, to describe
the relationships between components, and to explore the policy options that may result in the most
favorable performance outcomes. Specifically discussed in this paper is the basic structure of the
organizational change model and the impact of changing policies on several key variables. The
relationship of this system model to organizational leaming is discussed.
The Model Structure
For our purposes, there are two major dimensions of organizational change that are important:
scope of change, and timing of change. Scope of change can be either incremental or strategic.
Incremental changes are small and occur within the current framework or systems of the organization.
They are continuous in nature. The objective of incremental changes is to maintain or regain alignment
between key components within the organization, such as the strategy, structure, systems and people.
Strategic changes tend to occur when the very survival of the organization is at stake. They require
breaking out of the current context or frame of the organization, and attempting to move it to a
completely new configuration or alignment. Strategic changes are addressed to the organization as a
whole, are large scale, and are discontinuous. Both types of changes are considered in this simulation.
Page 209
Page 210 System Dynamics '91
In terms of timing, changes can be either reactive or anticipatory. Reactive changes are initiated in
response to an event (or series of events) that has occurred. Anticipatory changes are initiated in the
expectation of events that are considered likely to occur. Systemically, choices between reactive and
anticipatory change will affect the delays that are introduced into the system. In our model, only
reactive changes are explicitly considered at this time.
The basic structure of our model is shown in the causal loop diagram in Exhibit 1. The top half
(incremental change) and bottom half (strategic change) of the diagram have similar structures, but there
are some significant differences. We will first describe the structure that is common to both types of
change, and then discuss some of the differences.
Exhibit 1
Start
7 Merman i
Suecesstul”* Changes my
Incremental
Improvements F fed ‘S; Readiness for, Resources for
ae Incremental >” Incremental
Incremental Change
Changes © ~___-=>=”
+
Co”
~~ Perceived
Performance
Failed —- * Readiness
Strategic for Strategic ~
Changes Change
Successful +, Resources for
Strategic + Strategic
Improvements ce
Seantie 2
on +
iS)
We assume that the organization’s environment is changing such that some low level of continuous
change is required to maintain the organization’s performance. Therefore, in the absence of any
organizational change, the gap between the organization’s perceived performance and its goals (its
perceived performance gap) increases. As the perceived performance gap increases, its readiness for
change also increases, and it initiates change experiments intended to increase its performance. To the
extent that these experiments are successful they close the perceived performance gap and the cycle
begins again. These simple negative feedback loops have a few other loops attached to them.
One loop concems the impact of a failed experiment. Not all the change experiments succeed, and
to the extent that they fail they would increase the perceived performance gap. When there is failure, the
performance gap continues to grow. This creates a positive feedback loop.
Second, readiness for change is affected by the organization’s experience with prior change efforts.
Successful change experiments increase the readiness for change, creating another positive feedback
System Dynamics '91 Page 211
loop. Failed change experiments decrease the readiness for change (Hess, Ferris, Chelte & Fanelli,
1988). The link between experiments, successful or not, and readiness is an important one. We believe
that this simulates the organization’s ability to learn from its experience and to make connections
between policies, actions and results. Without this ability, an organization can only be reactive to
business performance outcomes.
Finally for both incremental and strategic changes, there is a limit to the number of experiments that
can be implemented during any time period. This limit is imposed by resource availability, which must
be allocated between getting the current job done and experimenting with changes which alter the job in
the future. As an organization improves its ability to learn, we would expect these two roles to merge.
The major differences between the two change strategies are longer delays and increased inertia in
the strategic change loop. Because strategic changes are large scale; discontinuous, and directed to the
organization as a whole, they have longer delays while they are being initiated and planned, and longer
delays while they are being implemented.
Building readiness for strategic change requires a larger perceived performance gap than for
incremental change (Nadler, 1989)), and usually does not occur unless the organization is in severe
crisis. Also, because they are large changes requiring large amounts of resources, the same level of
readiness leads to the initiation of fewer strategic changes than incremental ones.
Even though incremental changes are easier and faster to implement than strategic changes, they are
not a solution that can be relied upon to provide long term advantage. Kilman (1988) and Nadler
(1988) argue that incremental changes to a system can provide improvement in productivity for only
three or four years. As time goes on, the rétumis of working within the system diminish. The
fundamental strategic system becomes the constraint to performance improvement. At this point the
system itself should be changed. Once a new strategic system has been put in place, incremental
improvements to this new strategic system can again be effective. There is much literature which
supports the idea that over the long term, both types of change need to be operating (Imai, 1986; Miller
& Friesen, 1984 Nadler, et. al., 1989). It is also clear that there is a limit to how much change an
organization can absorb (Nadler 1989).
Model Performance.
A model has been built in ithink to begin testing the relationships described by the causal loops.
The model was initialized in steady state, and then subjected to a pulse in perceived performance gap
(the amount by which the organization is missing its goals), in order to simulate an organization that
fails to meet its goals. The model output, shown in Exhibit 1a, replicates the referent behavior that
Tushman et. al., 1989 have described; long periods during which there is no organizational change,
interspersed with bursts of organizational experimentation and change. Given the values that we used to
initialize the model, the duration of the “convergent” period (Tushman et.al., 1989), during which there
is no change, is approximately 30 years. The bursts of change occur over 7-10 years. Note that as the
perceived performance gap decreases, the actual performance of the business increases. In the exhibit
we have platted the performance gap measure, therefore lower values indicate better business
performance.
The model also produces other behaviors that are consistent with the literature. When the
convergent period comes to an end, the organization first tries to close its performance gap by making
incremental changes to the current strategic system. These can provide performance improvement for
only a few years, after which they fail and detract from performance. Only after these incremental
changes have failed to close the performance gap does the organization embark upon strategic changes.
These strategic changes are assumed to have a success to failure ratio greater than unity so they install a
new strategic system. The new strategic system acts both to close the perceived performance gap and to
enable incremental experiments to be successful again, and so we see both types of experiments
proceeding in harmony (Miller & Friesen, 1984).
Page 212 System Dynamics ‘91
a ived 1: PERCEIVED PERF GAP 2 Srategic Change Experiments 3: Incremental Chango Experiments.
Performance Gap Hl OR gp case Shewe cob ewew pele e CONTE Te TS eng
1
3] :
‘
‘
1
'
eee
1
‘
'
tb.
Pulse to Readiness for
Incremental Change
Testa toan 1: PERCEIVED PERF GAP 2 Suatogic Change Expeiments
Incremental Change 44. ito, :
Experiments 3] nia
1 4.50,
Hi
4
System Dynamics '91 Page 213
A third referent behavior can be seen by comparing Exhibits 1a—l1c. Here the model demonstrates
that the sensitivity of the model’s response is substantially decreased as an equal sized initiating pulse is
applied first to the perceived performance gap stock, then to the readiness for change stock, and then to
starting experiments stock. In short, we are systematically moving the reason for change away from the
performance issue. This result is consistent with the change literature which indicates that the likelihood
of achieving change is increased when the pressure for change occurs closest to the business issue (Beer,
Spector and Eisenstat, 1990; Tushman et. al., 1989; Nadler, 1989). As the driver for change moves
away from the business issue, the impact is diminished, the feel of the initiative begins to be perceived
as change for the sake of change. An example of an initiative that is not based upon solving a business
problem is the executive who comes back from a seminar, or reads a book, and thinks that creating a
vision would be great for his company. The company may get a new vision, but it will probably have
little to do with how the organization does its work.
Policy Test
The causal loop diagram shows that readiness for change is affected by both the perceived
performance gap and by the organization’s memory with success or failure of experimentation. In the
model, we can change the relative magnitude of these flows simply, by dividing or multiplying the
appropriate coefficients. This allows us to test different policies that differentially emphasize these
drivers. One policy, for example, would emphasize the perceived performance gap as a driver of
readiness; we can think of this as the “If it ain’t broke, don’t fix it” policy. Under this policy, if the
organization is meeting its goals there is low readiness for change; and this would still hold even if the
experience with prior changes had been very successful. On the other hand, a large perceived
performance gap would increase readiness for change and lead to initiation of change experiments, even
if the experience with prior experiments had been unsuccessful. The attention of the organization is
directed primarily on how well it is doing and paying little attention to how it is getting there.
Using the causal loop diagram and the model output shown in Exhibit 2a we can understand the
effects of this policy. A pulse in the performance gap increases readiness for incremental change and
this, in tum, leads to incremental change experiments being initiated. For the first two or three years,
these experiments are normally successful. After three or four years, however, incremental
improvements to the current strategic system are no longer effective and instead lead to decreased
performance, and therefore, a larger perceived performance gap. Since this gap is the primary driver of
readiness, the organization continues to do incremental changes that continue to fail and therefore drive
the performance gap larger.
Page 214 System Dynamics '91
1; PERCEIVED PERF GAP 2: Srategic Change Experments. 3: Incremental Change Experiments.
72.00
15.009
ki
System Dynamics '91 Page 215
This positive feedback loop would lead to complete corporate collapse. It is saved from doing this
because, at some point, the perceived performance gap gets so large that it drives an increase in
readiness for strategic change, which in turn initiates strategic changes.
Since we have assumed that the strategic changes are generally successful, they start to close the
perceived performance gap and create a new strategic system within which incremental changes can
again make a positive contribution. This burst of strategic and incremental change drives the
performance of the business to high levels, eliminating the perceived performance gap and therefore
driving all readiness for change to zero. The result is a period of stability during which there are no
more changes. In the popular literature this is referred to as organizational complacency. Since we
assumed that the organization’s environment is changing enough to require some level of organizational
change, this period of stability ends when the unchanging organization no longer “fits” its environment
and the perceived performance gap opens up again. An unanswered problem is understanding how to
recognize the need for change early since anticipating and reacting to this need have been cited as
advantageous for the organization (Nadler, 1989).
A second policy would be to increase the organization’s memory of success and failure with change
experiments. Under this policy, if prior experiments had made a positive contribution, the organization
would be ready to initiate more change experiments, and this would be true even if there was no
immediate perceived performance gap. The focus here is on the processes or experiments as well as the
results or performance.
Exhibit 2b shows us that under this policy the system behavior is different. We see the same initial
behavior of starting incremental changes but, when these incremental changes start to fail, the readiness
for incremental change is reduced. Therefore the organization does not continue to initiate more failing
incremental changes, and the performance gap is not driven so large. However, the perceived
performance gap is still large enough to increase readiness for strategic change and therefore initiate
strategic changes. As these strategic changes succeed, they close the perceived performance gap and
create a new system within which incremental changes are again positive contributors to performance,
Overall, cumulative performance under this-policy is much better than in the “If it ain’t broke, don’t fix
it” case, because the organization does not inflict so much damage on itself by continuing to initiate
incremental changes that fail before it initiates strategic changes. There are fewer delays in responding
to errors.
We might assume that since overall model performance was improved by emphasizing the
organization’s memory of success and failure with prior experience, then we would further increasing
this emphasis even more would achieve even better performance.
Surprisingly, it is possible in our model to over-emphasize experience with change experiments.
Exhibit 2c shows what happens when experience with change is much more powerful than the perceived
performance gap as a driver of readiness to change, i.e., “If it works, do it again”. After the initial
incremental changes start to fail, the organization stops doing them so quickly that it keeps the
perceived performance gap relatively small. This is beneficial in the short term, but we noted above that
perceived performance gap is a powerful driver for change, so that the likelihood of achieving change
is increased in response to a real business issue. The longer term effect of the small perceived
performance gap that occurs under this policy is that it is insufficient to drive the big burst of strategic
change that would create very high business performance in subsequent periods. Overall, the
cumulative performance under this extreme policy is lower than under the more balanced policy that we
described above. Exhibit 3 displays the cumulative performance gaps across these three conditions.
Page 216 System Dynamics '91
Exhibit 3:_ Cumulative Performance Gap
4: ‘If it ain't broke, don't fix it” 2: Balance 3: “If it works, do it again.”
1335.76 9
213.03 4
-909.69
BS deacceee
3
t
125.00
Months
0.00
Q
This suggests the intriguing hypothesis that outstanding business performance over the longer term
may be less likely to be achieved as a smooth progression than as a series of cycles. The declining part
of each performance cycle provides the readiness for strategic change that drives performance into the
growth phase of the next cycle. By contrast, a smooth increase in performance is unlikely to create
sufficient readiness for change to establish the conditions for superb performance.
Another interesting aspect of the model is its sensitivity to changes in the success to failure ratio for
experiments. We conducted a series of sensitivity tests. Model output is most dramatically affected by
independent changes in the success ratios of either incremental changes or strategic change experiments.
Perceived performance gap, for example, takes.a range of values twice as large when each of the
success ratios are changed by 20% than it does when any other system element is changed by 20%. It
seems somewhat obvious to say that if you increase your success ratio then your performance will
improve, but the sensitivity of the model to small changes in the success ratios is impressive. The
lessons for companies about learning what creates successful experiments in their own environments are
potentially powerful. Since there is evidence that organizations generally are not able to learn from
failures (Hess et. al., 1989; Mirvis and Berg, 1977), and since leaming what creates successful change
requires examination of both successes and failures, there is clearly much work to be done here.
Since the output of the base model matches the expected reference behavior pattern, we are
comfortable with the hypothesis that we have identified key components of the system that produces
organizational change. We also want to test some policies that may lead to more effective change over
time.
Discussion
The model has shown behavior that is consistent with the literature in the field. It has also
demonstrated behavior that suggests interesting hypotheses about company policies. For example,
taking the three policies described above suggests that paying attention to only organizational
performance is an insufficient strategy for sustaining performance. An interpretation of the observed
System Dynamics '91 Page 217
behavior would suggest that only knowing how you are doing without knowing how you got there will
cause waste and ultimately reduce competitiveness. Financially driven companies may fall into this
category. These are organizations which manage the budget rather than manage the business; i.e. they
make decisions which make the financial performance of the organization look good for short term
results, but ignore the longer term impact of these decisions for carrying out:the. business of the
organization. ,
It is just as clear that putting too much focus on the process rather than the results is also an
insufficient strategy for long term success. This observed behavior in the model might explain how an
organization can successfully go through a change process, incremental or even strategic, but get the
unintended result of no performance improvement. This condition, which is reported quite frequently in
the quality and customer focus literature, describes organizations making changes in areas which are not
important to the business or to the customer. The changes are being made out of context with no links
between the process or system changes and the changes in organizational performance. At its worst,
organizations are changing for the sake of change.
Performance improvement seems to be sustained in the long run only when there is a balanced
perspective between how the organization is performing and how the changes or experiments impact
that performance. In other words, from the data observed in this model, an organization will improve.
faster and longer when they know how they are getting better. This seems to be the basic definition of
organizational learning.
The intention of the model has been to explore the relationships between key variables known or at
least suspected to be critical in the process of organizational adaptability or leaming. The development
of the model to its current state has been successful in this, highlighting the impact of several policy
decisions, as well as identifying key sensitivity areas which could help organizations be more successful
during the change process. It is also clear that the model is not complete.
At this point we recognize several limitations to the model which we will be addressing over time.
These include the following:
1, Our experience and the literature (Kilman 1988; Schneider, 1990) suggest the importance of the
culture of an organization, the degree of alignment within the organization for the changes, and the
desirability of the new direction. We need to model these explicitly.
2. The second issue of concem is the organization’s capacity for change while carrying on the
business. While we have a simple resource constraint in the current structure of the model, we
believe it to be inadequate for modeling and understanding an on-going change process.
3. Finally, the model in its current configuration, seems to capture adequately the phenomena
described in the organizational change and quality literature. This gives us confidence regarding our
selection of formulae for the various variables. However, we have not yet tested the model with
specific case data. This will be done several times in the near future.
Page 218 System Dynamics '91
References
Beer, M., Spector, & Eisenstat. The Critical Path to Corporate Renewal. Harvard Business School
Press, 1989.
Hess, Farris, Chelte, & Fanelli, Learning from an Unsuccessful Transformation: A “Perfect Failure”.
R.H. Kilman, T.J. Covin (eds.): San Francisco, California: Jossey
Bass, 1989
Imai, K. Kaizen: The Key to Japan’s Competitive Success: Random House Business Division: New
York, NY, 1986.
Kilman, R.H. Coven, T.J. & Associates.
Competitive World, San Francisco: Jossey—Bass Publishers 1989.
Miller D. & Friesen, P. Organizations: -A Quantum View. Englewood Cliffs, NJ: Prentice Hall, 1984.
Nadler, D.A. Organizational Frame Bending: Types of Change in the Complex Organization.
ion R.H. Kilman, T.J..Covin (eds.): San Francisco, California: Jossey
Bass, 1989.
Schneider, B. Organizational Climate and Culture. San Francisco: Jossey-Bass, 1990.
Stata, R. Organizational Leaming - The Key to Management Innovation, Sloan Management Review
30(3) 1989.
Tushman, M.L., Newman, W.H., Nadler, D.A. Executive Leadership and Organizational Evolution:
Managing Incremental and Discontinuous Change. Corporate Transformation R.H. Kilman, T.J.
Covin (eds.): San Francisco, Califomnia: Jossey Bass, 1989.
Womack, J.P., Jones, D.T., Roos, D. The Machine That Changed the World. New York: Rawson
Associates, 1990.
