Beruflich Dokumente
Kultur Dokumente
STRATEGIC FOCUS
AND THE QUEST FOR TEMPORARY ADVANTAGE. Academy of Management
Proceedings, 1-6. Retrieved from Business Source Premier database.
NILS STIEGLITZ
Strategic Organization Design Unit
Department of Marketing & Management
University of Southern Denmark
Campusvej 55, 5320 Odense M, Denmark
THORBJØRN KNUDSEN
University of Southern Denmark
MARKUS C. BECKER
University of Southern Denmark
ABSTRACT
The paper addresses the challenge of how firms adapt to a turbulent environment and
reap temporary advantages. We contrast strategic flexibility and focus in the sampling of
alternatives. Our results suggest that strategic focus reaps temporary advantages in more
turbulent environments, while strategic flexibility is viable in less turbulent markets.
INTRODUCTION
How do firms adapt to turbulent business environments and reap temporary competitive
advantages? Market turbulence confronts organizations with a dilemma. On the one hand, they
must rapidly adapt to changing market conditions, requiring strategic flexibility. On the other
hand, they need time to learn about the performance implications of possible strategic actions.
Learning needs to be sustained, since first impressions could be misleading. This calls for
strategic focus in sampling alternatives. Our concern is primarily the adjustment of market scope
as a response to turbulent market conditions. Broad scope provides more options and thereby
supports flexibility. By contrast, a narrow focus restricts flexibility. This can be thought of as
putting Porter’s generic strategies of broad and narrow scope in a dynamic context.
Remarkable little is known about the way strategic focus, as opposed to strategic
flexibility, influences organizational adaptation. We develop a modeling structure to analyze
organizational adaptation in turbulent environments, drawing on the multi-arm bandit model
widely used in the literature on decision-making under uncertainty.
A turbulent environment is characterized by frequent and unpredictable changes of the
links between strategic actions and performance. To cope with uncertainty, decision-makers
form beliefs about action-performance linkages. In the absence of vicarious learning, decision-
makers have to rely on experiential learning by sampling strategic actions to form beliefs.
Two aspects seem to be of fundamental importance to understand organizational
adaptation. The first aspect relates to how much reliance decision-makers place on prior beliefs.
It corresponds to the classic trade-off of exploration and exploitation in organizational learning.
The decision-maker must choose whether to act upon current beliefs and to select the most
attractive action identified so far, or whether to seek out new information by choosing an action
currently believed to be inferior. Turbulence complicates the task, since entrenched beliefs about
action-performance outcomes may quickly become outdated. Intuition therefore suggests that a
more turbulent environment calls for less emphasis on prior beliefs and more exploration of
alternatives. This amounts to more strategic flexibility in sampling. In contrast, strategic focus in
sampling requires sampling strategic actions believed to be the most attractive one.
The second aspect is how decision-makers form beliefs. The question here is how
sensitive a decision-maker is to acquiring new information about the performance implications of
an action. Shall the decision-maker change estimates on first impressions, placing heavy
emphasis on recent experiences? Or is it better to be more insensitive to new information,
treating it as just another piece of evidence in finding a good strategic action? Strategic focus
calls for sustained learning about strategic actions, drawing on all information available. Again,
intuition suggests that a turbulent environment calls for more flexibility and higher sensitivity in
updating beliefs, since new information should be more reflective of the current state of the
world than previously obtained information.
While the first aspect looks at how much reliance is placed on prior beliefs, the second
aspects deals with forming these very beliefs. The interactions between these two fundamental
aspects of organizational adaptation have not been systematically explored in prior research. And
little is known about how they influence organizational adaptation in turbulent environment.
MODEL DESCRIPTION
X n ,t = X n ,0 + ! ( X n ,t #1 # X n ,0 ) + " (1),
with the random shocks ! i.i.d and normally distributed, with a mean µ = 0 and variance "2. The
variance determines the magnitude of environmental changes. The higher the variance, the more
do mean payoffs for a strategic action differ across two time periods. A higher variance therefore
signifies a more turbulent environment. The parameter #, bounded on the unit interval, is the
autoregressive coefficient and determines how quickly a trend reverses itself. It sets the pace and
direction of change. A lower value of # essentially makes the current payoff less dependent on
the previous time step. This translates into a more turbulent environment, with high volatility and
less discernable trends in the performance of strategic actions.
Every time a decision-maker samples a strategic action n, she receives a payoff drawn
from a normal distribution with the mean Xn,t. Decision-makers form beliefs about the
attractiveness of each strategic action. We draw on the reinforcement learning literature to model
belief formation. In principle, a decision-maker strengthens the belief in a strategic action that
yields a high payoff. Formally, the model uses the following rule for updating beliefs:
(! n ,t / Tn ,t # En ,t ) ! n ,t ! n ,t #1
En ,t +1 = En ,t + " t
+ (1 # " )( # ) (2).
Tn ,t Tn ,t #1
$T
i =1
n ,t
E is the belief about a strategic action n, T the number of trials in a time step, and $ is a
parameter bounded on the unit interval. The first part updates beliefs based on an estimate of the
average payoff of the strategic action by equally taking into consideration all payoffs received so
far. The estimate is fairly insensitive to recent payoffs and it thereby signifies strategic focus in
sampling strategic actions. The second part considers the marginal gains of a strategic action,
leading to a very high sensitivity to recent payoffs and more flexibility in sampling. The
parameter $ sets the weights placed on these two approaches to updating beliefs. A high $ tends
to make belief updating more accurate and slower, while a low $ corresponds to more sensitive,
but less accurate updating.
Managers sample strategic actions based on their beliefs about performance implications.
The decision-maker must choose whether to act upon current beliefs and to select the most
attractive action identified so far, or whether to seek out new information by choosing an action
currently believed to be inferior. A straight-forward approach to modeling sampling and resource
allocation based on these principles is provided by the Softmax algorithm. The probability pn,t of
choosing action n out of N options in time period t is given by
N
En ,t /! En ,t /!
pn ,t = exp / (" exp ) (3),
n =1
with E as the belief about action n in t. The parameter % is the proclivity to explore. With a low %,
a manager focuses sampling on the most attractive strategic action identified so far. Resources
get deployed to actions believed to be superior. A higher % results in a higher probability of
choosing an action so far believed to be inferior. It corresponds to more strategic flexibility in
allocating resources and more active exploration of the environment.
RESULTS
We use the analytical structure outlined above to study the search for temporary
advantages. Each business environment consists of ten strategic actions (N = 10). All simulations
are seeded with 10,000 agents and the mean payoffs of each action are firm-specific. Managers
have no prior beliefs about the environment they operate in. Each simulation runs over 250 time
steps. In each time step, an agent may allocate two samples (trials) to strategic actions.
We first compared strategic flexibility and strategic focus for the baseline case of a stable
environment. In a stable environment, the mean payoffs of strategic actions stay constant over
time. We find that strategic focus – a low proclivity to explore and very low payoff sensitivity –
reaps significant performance benefits (the results are available upon request).
------------------------
Figure 1 about here
------------------------
Figure 1 contains the main results for turbulent environments. In principle, a turbulent
environment does call for greater strategic flexibility compared to the stable environment. This
reinforces the basic intuition of prior research. Counter-intuitively, the results also show that
greater market turbulence requires less strategic flexibility and more strategic focus. Firms cope
better with an increasing pace of external change by focusing the sampling of strategic actions.
They have a lower proclivity to explore and are less sensitive to recent payoffs. In contrast, when
the environment is characterized by a lower pace of external change – with long, sustained trends
– strategic flexibility in sampling performs much better.
Increasing the magnitude of environmental change does not change the basic relationship
between strategic focus and market turbulence (detailed results are available upon request).
Strikingly, major random shocks make more insensitive updating of beliefs viable. In principle,
more market turbulence calls for more accurate learning. Managers gain from being more
insensitive to new information the more volatile the business environment becomes. Focus in
updating beliefs is more effective in turbulent markets. The important boundary condition here is
that highly accurate learning ($ = 1) combined with absolutely no exploration is a bad idea in
changing environments, regardless of the level of market turbulence. The results also suggest a
nuanced relationship between belief formation and the bases of learning. A higher proclivity to
explore goes hand-in-hand with more inaccurate learning, and vice versa.
These findings are very robust to changes in the model parameters. The principal results
hold for different specifications of prior knowledge about the business environment and for more
localized changes in the business environment (turbulence changes the performance of some, and
not all, strategic actions). The results are also substantiated by considering disruptive changes in
the structural conditions of the business environment (changes in Xn,0). Environmental
turbulence helps agents to cope with disruptive changes, since beliefs become less firmly
entrenched. However, our results do not apply to stable environments that experience disruptive
change. Under those conditions, responding to disruptive change calls for greater strategic
flexibility to quickly abandon old strategic actions and identify new attractive ones. These results
are available from the authors upon request.
REFERENCES
# =0 # = 0.25
1600 1.0 1600 1.0
1400 1400
0.5
0.5
1200 1200 0.5
0.5 0.5
0.5 0.1 0.1 0.1 0.1 0.1 0.5
0.1 0.5 0.5
0.1 0.1 0.1
1000 1000
800 800
" "
600 600
400 400
200 200
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
! !
# = 0.50 # =1
1600 1600
0.5
1400 1400
0.5
1200 0.5 1200
0.5 0.5 0.5
0.1 0.5 0.5 0.5 0.5 0.5
0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
1000 1000
800 800
" "
600 600
400 400
200 200
Figure
0
0
1:0.1Results
0.2 0.3
for
0.4
environments
0.5 0.6 0.7 0.8
with
0.9 1
minor random
0
0
shocks
0.1 0.2
(" 0.4
0.3
=1)0.5and
0.6
different
0.7 0.8 0.9
paces
1
of
environmental change (# = 0, 0.1,…, 0.9). Note that lower # signifies increasing paces of change.
! !
The bars report the average accumulated performance at the end of the simulation run (t = 250)
for the best-performing resource allocation strategy (%), as indicated by the label on top of each
bar. The absolute performance difference between the best- and the worst-performing resource
allocation approach is also indicated for each performance bar. Each panel shows a different
updating rule ($ = 0, 0.25, 0.5, 1).