TIME - A KEY FACTOR IN CORPORATE STRATEGY
Peter M. Milling
Massachusetts Institute of Technology
Alfred P. Sloan School of Management
Cambridge, Massachusetts 02139
ABSTRACT
Empirical analyses indicate that the firm which is the first in bringing new products to the
market has a major competitive advantage. The development time for sophisticated and high
quality products is shortening. The time span of the market cycle is decreasing, and for
high technology firms, even rather short delays can cause a deep cut in the overall profit
performance. In the "Factory of the Future" the capability for immediate and reliable deliv-
ery of custom designed products is a crucial aspect.
Speed is becoming a decisive factor for corporate management. In Management
Science, however, this development is not yet taken into account adequately. Different
stages of the same process are still analyzed separately. Models of research and develop-
ment e. g., do not investigate how delays influence the market performance of the eventu-
ally achieved product. Studies of innovation diffusion focus solely on the market cycle,
thereby neglecting the lengthy and costly R&D processes. With such a limited perspective,
those models must fail to support effectively decision making in a dynamic high technology
environment.
The paper discusses System Dynamics’ role in such a setting. It presents a model for
innovation management which integrates the stages of R&D with the production and mar-
keting cycle. It is designed as a microworld for learning about the system and for studying
possible ways of influencing its behavior. The model consists of two modules: a C-written
algorithm, based on biological evolution theory, maps the firm's research and development
processes; the second module is a Dynamo-representation of innovation policies a77 inarket
dynamics. Both modules are tightly coupled throug!: “lows of information. Their interac-
tions allow the testing of corporate strategies for R&D planning and innovation manage-
ment.
Although still in the development stage, the model provides insights into the timing of
decisions. The results from this integrative view underline the importance of speed in the
strive for competitive advantage.
769
770 System Dynamics ‘90
THE DYNAMICS OF TECHNICAL CHANGE
In corporate management the product life cycle frequently serves as a powerful heuristic. It
is applied to outline the dynamics of market growth and decay. Its logistic curve is used to
forecast expected sales. The portfolio approach in strategic planning focuses of the different
stages of product maturity to determine the competitive position of a firm.
Most formal models generating the cycle’s characteristic sigmoid curve, however, do
not reflect properly the factors causing this behavior. They either use specific mathematical
functions to produce a predetermined behavior mode or are based upon biological or physi-
cal analogies. They fail to take sufficiently into account the economic environment of cost
and prices, quality and in-time delivery, etc. Purchasing decisions do not follow the same
natural laws as do the spread of a disease, the dissipation of particles or similar processes,
which were the original study objects in the natural sciences.
Furthermore, most of these models reflect the diffusion, not the evolution of a phe-
nomenon. They concentrate on - and thereby isolate - the short term dynamics of an inno-
vation. They do not explain the substitution processes between a sequence of products at
different levels of sophistication which take place as technology advances. Figure 1 shows
this sequence of life cycles for the sales of PC micro processors. It emphasizes the dynam-
ics of change as the transition from one level of technology to the next takes place.
Sales of Intel-Processors
{in Millions PC-Units]
cg
45 ( (2B soss/eces ~(_Jeozes Ml s0386/a0486
3
1,5
0
1983 1985 1987 1989 1991 1993
Source: Dataquest
Milling, Innovation
Fig. 1: Transition between Levels of Technology: Sales of Intel Processors in Europe.
System Dynamics '90
Due to the dynamics of innovative products, transition between technologies can occur
rapidly, leaving the firm little time to sell the product and to earn its return on investment.
While in the above figure, only one corporation controlled and nearly monopolized the
market; in the areas of high technology there is frequently fierce competition. The sales and
price development of dynamic random access memory chips (DRAMs) in Figure 2 exem-
plifies such a situation.
RAM-Chips: Technology Cycles RAM-Chips: Price Development
Sales [1E6 Chips} ‘gg lee {47cm}
10000,
1900
100
10
1979 «1979=«1982,S«198S— 1988194 10781979 -=Ss0e2S198Sstvea= too)
anag, Mawraion
Fig. 2: Sales and Price Development of DRAM Chips.
Each new generation brought along a quadrupling of its storage capacity and a signifi-
cant increase in production volume. With large scale integration grew the number of pro-
duction operations, and the development cost increased progressively. In 1985 investment
requirements for a state of the art factory were in the range of $ 50 million, four years later
they reached $ 250 million. DRAMs are manufactured by more than a dozen companies,
and - despite their sophisticated design and delicate production processes - are traded like
commodities. Competition leads to a rapid decline in prices immediately after the new tech-
nology entered the market. Trailing the leader’s race to market by only a short time, makes
it very difficult to achieve a sound economic performance.
In general, high expenditures for R&D and manufacturing equipment, short product
life cycles and a dramatic decay in prices immediately after market introduction characterize
this field - a constellation which is called a dynamic environment. Here it is vital for the
companies to build up in time the production capacity required; delivery must be fast when
demand gains momentum. Only early in the life cycle can high prices be charged; during
this stage the firm must earn its compensation for research and development expenditures,
the investment in production equipment and facilities, and for the general risk taken.
In a dynamic environment, speed becomes a competitive factor of a strategic dimen-
sion. Time to market and time to volume are essential for innovation management. A
McKinsey study revealed that a six months delay in starting production of a new product
can reduce overall earnings of up to 30 % (Dumaine 1989). Similar results are reported
from other sources and are confirmed by own analyses.
Such kind of a problem description might suggest a strategy of high pricing together
with an early provision of production capacity. This, however, also implies high fixed cost
and little flexibility if demand were overestimated or the market is in its downswing. Fur-
thermore, a policy of skimming prices during the introduction of a new technology might
cause slow market penetration; it might impede innovation diffusion and the gain of market
share. Without rapidly expanding production volume, only little benefits can be drawn from
experience and learning processes. These contradicting aspects emphasize the need for a
thorough understanding of the complex decision making parameters and their mutual inter-
actions.
The speed of innovation diffusion and the risks of upcoming substituting products are
not solely and not even mainly determined by factors outside the firm. Corporate strategy
produces, or at least co-produces innovation dynamics. R&D resource allocation and the
timing of market introduction, investment and production planning, cost management and
pricing policies, product quality and delivery delays are key control variables. They are also
indicators for successful innovations.
Innovation management requires decisions whose effectiveness are fundamentally
important for the competitiveness and the viability of an enterprise. Decision making at this
level of complexity cannot be automated, but it is possible to support it by means of formal
models and computer based systems. These Strategy Support Systems link cognitive pro-
cesses to computer routines and gain synergy from this combination (Morecroft 1984). The
expected benefits justify the substantial efforts required. Computer simulation allows
insights into the behavior of systems. It combines the theory based investigation and the
practical research of laboratory experiments and constitutes a third pillar for rational deci-
sion making.
If management wants to play a major role in controlling the future course of the com-
pany - and not simply adjust to predetermined behavior modes -, it must have a thorough
understanding of the system. A comprehensive and causal approach to model building is
required. To achieve insights into the processes under investigation, the factors which
cause system behavior must be represented. Models must explain and help to understand
why specific behavior modes occur; they must explicitly link structure to behavior. Only
then, can they be meaningful tools to improve managerial decision processes.
Innovation planning and control is seen as a learning process about the system and its
environment. Insights into the dynamics of the system under investigation, not forecasts
nor predictions, are the objective of such an endeavor. It is a paradigm of this systems
approach that through the iterating processes of model design and model analysis the
System Dynamics ‘90
System Dynamics '90
intrinsic properties of the problem will be better understood (de Geus 1988; Hayes,
Wheelwright and Clark 1988; Milling 1989b; Stata 1989).
AN INTEGRATED INNOVATION MODEL
To study the implications of innovation strategies, a general causal simulation model was
developed. Frequently only the market stage, during which the product is sold, is associ-
ated with the notion of an innovation. However, before the availability of a marketable
product lies the costly, lengthy and risky period of research and development. While the
market cycle tends to become shorter and reduce the time for the corporations to earn their
money, the research and development phase requires increasingly more time, personnel and
financial resources. These diverging trends make it difficult to achieve a satisfactory profit
performance. A comprehensive investigation into innovation dynamics must cover both, the
development and the market cycle.
The innovation model consists of two modules: one reflecting the processes of R&D,
the other representing the market cycles. Figure 3 shows the structure of the overall model
and its components. Both modules are linked through flows of information to monitor the
resource allocation, the intensity of the R&D-processes, the required minimum quality level
before a new product is considered ready for market introduction, etc.
Model of R&D~process } Model of Fater rise
Core Model of
Product Life Cycle
f Competition
(New) Knowledge
SSD} Market Entry
Information
_—— R&D-Planning
[Variation]
Fig. 3: Coarse Structure of the Integrated Innovation Model.
Quality
The module of the research and development phase deals largely with intangible pro-
cesses. Many attempts were made to define a production function for research and devel-
opment, using as input the allocated resources like budget, number of people assigned to
774 System Dynamics ‘90
the task, time and equipment available, etc. In general, these attempts were not successful
in describing how the various factors operate together to achieve the desired results. In this
model a different approach is used. An analogy to biological evolution theory defines how
new concepts develop by the variation and mutation of existing and known solutions. The
respective results are evaluated on the basis cf viability. If they seem to be superior to pre-
vious combinations, they are selected for further development, i.e. as the basis for future
evolution. Otherwise they are discarded. This evolution module is a C-written algorithm
that is linked to and interacts with the production and market part of the model.
When the development algorithm is run as a stand-alone module - without interference
from the production and marketing part of the integrated model - it generates the chain of
subsequent products at an increasing level of technological sophistication. In the long run,
it creates the S-shaped curve enveloping the technological development of the individual
innovations .
The production and market structure is based methodologically upon the System
Dynamics paradigms, i. e. the feedback perspective of social systems and the use of com-
puter simulation for gaining a better understanding of its properties. Professional
DYNAMO plus was used to represent the module, to link it to the evolution algorithm
through the External Function facility, to simulate and analyze the total model.
The intrinsic structure of the market penetration processes is derived from the classic
epidemics model; the interactions between potential and actual buyers generate the charac-
teristic behavior of a life cycle. This part of the overall model was described in detail else-
where (Milling 1986). The kernel is extended to include - among others - such aspects as
investment in production capacity, quality control, cost performance and pricing strate-
gies!,
In the core model three categories of customers, their different buying behavior, and
the respective shift in dominance between them, are taken into account: innovators, imita-
tors, and - in the later stages - repeat or replacement purchasers. They differ in their reac-
tions with respect to the already achieved market penetration, prices charged, and product
quality. This differentiation is frequently found in innovation literature (Bass 1980, 1969;
Jeuland and Dolan 1982; Rogers 1983).
The model’s investment in production facilities is based on a comparison between pro-
jected demand and available production capacity, both measured in units per period. Since
primarily, the model is not designed to predict the market success of an innovation but
rather to study and comprehend the effects of management strategies on innovation perfor-
mance, projected demand and actual demand behave rather similar. This alleviates the task
of demand forecasting.
1 All extensions described in the following paragraphs are indeed implemented and were investigated
to study particular innovation strategies. The simulation results discussed in the subsequent section of this
paper, however, were obtained with not all sectors activated.
System Dynamics '90 775
Besides required capacity, the firm's willingness to accept risk is explicitly considered
in the investment decision. A cautious attitude towards market risks and opportunities will
reduce, a more offensive or aggressive behavior will increase the indicated change in capital
investment. There are no financial limitations for the desired investment level, assuming
that the requested amount of capital will be available.
Total investment serves two purposes, it provides the needed capacities for production
purposes and those allocated to quality control. Investment comprises all factors (capital
goods and human resources) in their required respective proportions according to an
(implicitly assumed) production function. At the investment stage, no distinction is made
between the two operations; the facilities can either be employed for manufacturing new
products or for assuring the desired quality level of the process. Even after personnel has
been hired or specific equipment has been acquired, a reallocation between production pur-
poses and quality control is possible. Such a shift, however, requires an explicit realloca-
tion decision and must allow a certain adjustment time to take place. In the economic theory
of technical change, an analogous situation is characterized by “putty-putty" factors of pro-
duction (Phelps 1963).
Resource reallocation between both ways of capital usage is monitored by a decision
tule based on the firm's readiness for delivery - a relationship between incoming orders and
the actual level of production to meet this demand. In a market with short life cycles and a
constant threat of substitution through upcoming new products, delivery delays can cause a
permanent loss of demand. A strategy to provide the capacities for immediate delivery --
even at the cost of less careful quality control - might be an effort worth to be considered.
However, as the level of perceived quality decreases, production and operating result will
stagnate; they will remain far behind the values of other courses of action. The otherwise
gradual transition of potential customers to a later technology becomes very rapid and leads
to a steep decline of sales volume.
Quality turns out to be a prime competitive factor during all phases of innovation dif-
fusion. In the early stages, it influences market growth and penetration, in the later stages it
exhibits an important impact on the rate of imitation and the speed of substitution. Short
term advantages resulting from the negligence of quality, e. g. the gain of additional pro-
duction capacity and reduced cost, turn into severe damages. In a medium and long range
perspective, they cause detrimental effects on all relevant variables.
High quality does not extend an innovation’s economic life time significantly, but it
causes demand and production to peak at much higher levels and to descend from there less
rapidly. This leaves time to adjust to the change in the rate of production. The profit situa-
tion is favorably influenced - an observation that corresponds fully with empirical data on
the role of product quality and its influence on purchasing decisions (Milling 1989a,
Huibregtsen 1989).
To study the impact of pricing, the model contains an elaborated set of strategies.
Although in the short nin prices can be set without explicitly striving for full cost coverage,
776
such a policy is not sustainable in the long run; prices must be related to cost. Manufac-
turing's performance influences cost, prices and demand, which then feeds back to cost.
The dynamhics of these interdependencies have specific consequences for innovation pric-
ing.
A central tenet for innovation management is the control of average unit cost ,;which in
turn depends heavily upon the selected pricing strategy. The model derives actual unit cost
_ on the basis of two variables: long term developments on one hand, which are dominated
by the impact of learning or experience, and short term capacity usage with its accompany-
ing fix cost influences on the other hand?. The long range standard cost per unit are derived
on the basis of accumulated production volume. The actual cost for each period are calcu-
lated from the respective standard value modified for production volume variances. The
dynamic cost function generated in the model determines actual unit cost in each period
c(t) = [X(), xO] ()
on the basis of gained experience X(t), which gives the standard unit cost, and the variance
resulting from the level of capacity utilization achieved with the output volume x(t).
In a dynamic environment where the number of customers is influenced by the level of
adopters and where technological substitution can occur very rapidly, it is crucial for the
firm to gain a large market share and achieve rapid diffusion of the product. Prices are set
according to a policy: .
p(t) = oP(ty*x*(1-ate/T) 2)
where 1 defines the profit margin on top of long run standard cost. The use of standard
cost prevents the otherwise typical effects of full cost pricing. Multiplication with the expo-
nentially decaying factor with time constant T reduces prices in the first months after market
introduction, and induces rapid innovation diffusion during introduction. This policy -
among others - is discussed in Milling (1988); it has been proven the most effective one in a
dynamic environment. A control theoretical analysis of dynamic pricing with similar results
but which derives at an optimal solution is provided by Jeuland and Dolan (1982).
2 ‘The experience impact on cost assumes a direct relation between cumulated production X (which
incorporates a firm's experience) and average unit cost &, adjusted for inflation; cP defines standard unit cost
at the planned level of production. Every doubling of X is associated with a cost reduction by a constant and
predictable percentage:
Po= o, (X(/a] 5
where c stands for the cost of unit n (n < X) and 6 represents a parameter which depends on the experience
rate. For many businesses experience rates of 15 % to 25 % have been observed. Empirical evidence for this
relationship and a discussion of the implications for corporate strategy are provided in (Henderson 1968).
System Dynamics ‘90
System Dynamics '90
The core model with its extensions forms the base of a strategic Decision Support
System. It analyzes the processes of innovation diffusion from the perspective of corporate
management. It serves as a simulator to investigate how strategies can foster or hinder mar-
ket diffusion and profit performance. With such an objective the problems of model quan-
tification and validation in the absence of sufficiently corroborated information are evident.
Management, however, has to decide about the timing and the strategy of innovations -
even if there are no “hard” data available. Whether this is done on the basis of intuition and
mental models or whether formal, computerized support tools are used is only an instru-
mental question. Computer modeling alleviates communication, enforces precision, and and
makes at least some of implicit assumptions explicit; it encourages discussions and builds
consensus towards improved decision making.
The model is a conglomerate of information and data from several sources. It draws
on the concepts of innovation diffusion as they are widely accepted in the scientific litera-
ture; it uses statistically corroborated data whenever they are available. Yet a large part of
the information comes from less rigorous sources; it results from presentations and discus-
sions with managers where the model had to pass kind of a Turing-test (Turing 1950): The
structure of the model was critically analyzed by experts in the field. It was speculated
about possible reactions to managerial actions. Results were compared and evaluated
against the respective experience. Finally the model’s actual behavior modes were consid-
ered to be meaningful representations of real world responses. Despite these endeavors, the
claim for model validity - especially of its development module - remains modest.
INVESTIGATING THE TIMING OF INNOVATIONS
The extent to which a firm's policies determine the performance of its innovation, how they
should be planned and controlled is analyzed and evaluated on the basis of the following
assumptions:
(1) The products under consideration can be thought of as a technologically sophisticated
consumer durable likes CD players, video recorders, etc. Competition can and will
occur at a micro level between different firms and at a macro level between different
states of technology. There will be substitution when an economically more attractive
product is available or when a superior technology is introduced at competitive terms
into the market.
(2) Production on stock - at least at a level significant for the analysis - is not possible.
When incoming orders stay below capacity, the level of output is adjusted accord-
ingly. Over the ten years time horizon of the analysis such an assumption is realistic.
(3) The overall market acceptance of the innovations is assumed. The model serves as a
simulator to study how individual strategies can accelerate or hamper market penetra-
tion and profit performance. It is not the objective to predict market success or failure
of an innovation.
778
System Dynamics '90
Figure 4 shows a simulation run of the model with both modules coupled. Although the
outcome is similar to Figure 1, the underlying scenario is quite different. It is assumed that
at the time the innovation is launched, three competitors are in the market. The firm under
investigation and its main competitor hold initially the same market share, and in this base
run they offer identical products and follow the same strategies. They thus allocate both the
same resources for R&D, they choose the same timing for bringing new technologies to the
market, they offer the same quality, etc. Consequently the life cycles for both firms virtu-
ally overlap3.
Sales [Units]
25000
Product
Product 2
20000 +
Product 3
Product 4
15000 4 Total Sales
10000 +
5000 5
Fig. 4: Base Mode of the Integrated Innovation Model.
The behavior of production or sales volume and the development of operating results
duplicate the usual characteristics of a product life cycle. In the early periods, a large num-
ber of potential customers is attracted by the innovation. Increased product awareness leads
to accelerating demand. In later stages, the remaining market volume is reduced from two
sides: the number of potential customers declines and even they start to shift towards the
substituting technology with its superior performance and product attributes.
The curves emphasize the importance - for corporations in general and high technol-
ogy firms in particular - to provide a steady flow of new and improved products. Without
3 The required computation time to produce this output is rather long. On an IBM PS/2 Mod. 70 386
with 25 MHz it takes approximately 45 minutes for one run. The time depends mainly upon the size of the
“Technology Matrix” used in the algorithm and the number of iterations per cycle.
System Dynamics ‘90
779
such a replacement of less sophisticated technologies, sales and profits will soon start to
deteriorate. Firms cannot successfully rely on one single innovation, they must be based on
a corporate philosophy, which understands innovation as a permanent task.
Entering the market with an innovation too early increases the risk to fail. Potential
customers might still be unprepared for the the new technology; quality might not yet be
fully mastered; shipping delays might occur because of disturbances during the start of pro-
duction, etc. On the other hand the early launch of an innovation offers attractive opportu-
nities. Decisive competitive advantages might be achieved and sustained. A dominant mar-
ket share is difficult to attack from competitors following a me-too strategy.
Sales [Units]
30000
tod 18
Product 1 Market Entry Enterprise 1 Product 2 tn Per!
Competitors Market Entry in Period 20
Product 2 Competitors Life Cycles with dotted Lines
Weighting Factor of cumulated Market Share = 1
Product 3
20000 4 Product 4
10000 -
Fig. 5: Earlier Market Entrance of Corporation 1.
In the run it was assumed that all competitors enter the market with their new genera-
tion of products exactly at the same points in time. In the following Figure 5 this assump-
tion is waived; Enterprise 1 enters now 5 months earlier, at t=15, while the two competitors
make their market appearance at t=20. All other policies and all future entrance dates remain
unchanged. Figure 5 shows only the curves for the main competitors. For the first innova-
tion both life cycles overlap; for second product with the higher level of technological
sophistication the differences in sales become very substantial.
780 System Dynamics '90
At its maximum, the delayed product reaches a sales volume which remains approxi-
mately at only one third of the value achieved by its faster competitor. It can never recover
from its late market entry. And - although the future entry times are the same for all com-
petitors - the firm has acquired superior resources. It can devote more means to developing
and marketing later innovations. A sustainable competitive advantage was gained through
speed.
IMPLICATIONS FOR CORPORATE PLANNING
The results of the model analysis hold true only if the innovations are indeed accepted
by the market. Unsuccessful products would leave the firm with ample excess capacity and
hurt its profit performance. The willingness to accept risk is a crucial prerequisite for the
high technology business. Attempts to reduce or even to evade these hazards lead to poor
results. If firms decide to launch innovative products, they have to provide the necessary
production capacities to assure smooth delivery. Time to market and time to volume are
strategic variables in the strive for competitive advantage (Bower and Thomas 1988).
The model emphasizes that the market performance of innovations is heavily influ-
enced by intra-firm factors. It is not only price and advertisement that have a profound
impact on demand. The life cycle - in its time pattern as well as in its absolute volume - is
significantly controlled by the corporation's internal decision making. Management of
research and development, production scheduling, shipment logistics etc. influence the per-
formance of technological innovations. Newer developments in production technology and
production planning, e. g. concepts like Computer Integrated Manufacturing or the
“Factory of the Future", will even accentuate these results.
Well designed computer simulation models allow an effective investigation into such a
complex problem situation. They contribute to an improved understanding of the multiple
interactions in a dynamic environment. Through computer-based decision support tools,
management can improve its perspective and understanding of reality and achieve more
effective decision making.
The model presented here is designed in a modular fashion, and offers the flexibility
to be adopted to different types of innovations. It provides - even for situations that exceed
the capabilities of analytical methods - the possibility to study different courses of actions in
the setting of a management laboratory. In the real world, already a very few variables suf-
fice for a complete misperception of a decision situation (Sterman 1989). Strategy Support
Systems combine human creativity and judgment with the the capabilities of the power of
high speed computing.
The concept of decision support is linked to notions like learning, interaction and
evolution. It emphasizes the process of learning in developing a strategy rather than the
final result; the interactions with different facetes of a problem lead to a better understanding
System Dynamics '90 781
than the application of prefabricated procedures; problem solving should be the result of an
evolutionary process not an automated choice (Checkland 1985). The capability of the firm
to adapt and to learn faster than its competitors might be the only defendable competitive
advantage it has in the long run (de Geus 1988).
The ultimate purpose of effective planning in complex systems does not lie in produc-
ing plans but in changing the the mental models of the decision makers. Decision support
systems function as catalysts. They help to clarify complex internal images and to analyze
them.The demonstrate how action and reaction, stimulus and response, or cause and effect
fall apart. The knowledge and the technical requirements for the development and applica-
tion of such kind of decision support is available; their application make individual decision
processes to become more transparent and to converge. Adequately used, strategy support
systems provide a better understanding of the problem under investigation, they allow a
faster reaction to market developments and the achievement of decisive competitive advan-
tages.
REFERENCES
Bass, F.M. 1980. The Relationship between Diffusion Rates, Experience Curves, and
Demand Elasticities for Consumer Durable Technological Innovations. Journal of
Business 53: S50-S67.
Bass, F.M. 1969, A New Product Growth Model for Consumer Durables. Management
Science 15: 215-227.
Bower, J.L., and T.M. Hout. 1988. Fast-Cycle Capability for Competitive Power.
Harvard Business Review 66(6): 110-118.
Checkland, P. 1985. From Optimization to Learning: A Development of Systems Thinking
for the 1990s. Journal of the Operational Research Society 36: 757-7167:
Dumaine, B. 1989. How Managers can Succeed Through Speed. Fortune, No. 4,
February 13: 30-35.
Henderson, B.D. 1968. Perspectives on Experience, Boston: The Boston Consulting
Group.
de Geus, A.P. 1988. Planning as Learning. Harvard Business Review 66(2): 70-74.
Hayes, R.H., S.C. Wheelwright, and K.B. Clark. 1988. Dynamic Manufacturing.
Creating the Learning Organization. New York: The Free Press.
Huibregtsen, F.W. 1989. Management of Quality: The Single Most Important Challenge
for Europe, Paper delivered at the European Quality Management Forum, Montreux,
Switzerland, October 19, 1989.
Jeuland, A.P., and R.J. Dolan. 1982. An Aspect of New Product Planning: Dynamic
Pricing. In Marketing Planning Models, ed. A.A. Zoltners, TIMS Studies in the
Management Sciences Vol. 18, Amsterdam: North-Holland, pp. 1-21.
Milling, P.M. 1989a. Production Policies for High Technology Firms. In Proceedings of
the 3rd European Simulation Congress, ed. D. Murray-Smith, J. Stephenson, and
R.N. Zobel, Edinburgh, pp. 233-238.
782 System Dynamics ‘90
Milling, P.M. 1989b. The Design of Strategy Support Systems. In Proceedings of the 2nd
International Symposium on Systems Research, Informatics and Cybernetics, Baden-
Baden (forthcoming).
Milling, P.M. 1988. Controlling the Market Performance of New Products. In Advances
in Systems Research and Cybernetics, ed. G.E. Lasker, Ontario: I.1.A.S., pp. 310-
319.
Milling, P.M. 1986. Diffusionstheorie und Innovationsmanagement. In Technologie- und
Innovationsmanagement, ed. E. Zahn, Berlin: Duncker & Humblot, pp. 49-70.
Morecroft, J.D.W. 1984. Strategy Support Models. Strategic Management Journal 5(3):
215-229.
Phelps, E.S. 1963. Substitution, Fixed Proportions, Growth and Distribution.
International Economic Review 4: 265 - 288.
Rogers, E.M. 1983. Diffusion of Innovation, 3rd. ed. New York: The Free Press.
Stata, R. 1989. Organizational Learning - The Key to Management Innovation. Sloan
Management Review 30(3): 63-74.
Sterman, J.D. 1989. Modeling Managerial Behavior: Misperceptions of Feedback in a
Dynamic Decision Making Environment. Management Science 35: 321-339.
Turing, A-L. 1950. Computing Machinery and Intelligence, MIND 59: 433-460.
