Dynamic Consequences of Pricing Strategies for
Research & Development and the Diffusion of Innovations
Peter M. Milling and Frank H. Maier
Industrieseminar der Universitét Mannheim
P. O. Box 10 34 62, D-68163 Mannheim,
Federal Republic of Germany
ABSTRACT
The development and diffusion of innovations is a highly dynamic phenometion. It is
influenced by various factors like price, product quality, and market entry time. The
paper discusses the impact of pricing strategies on R&D performance and the diffusion
of innovations. It is based on a comprehensive decision support model in the field of
innovation management. The model consists of two components: (1) an evolution algo-
rithm modeling the processes of corporate R&D, and (2) a DYNAMO-based modul
mapping corporate policy making and the structural fundamentals of market dynamics.
The integrated model is used to analyze the dynamic consequences of different pricing
strategies on research and development, the readiness for market entry and the resulting
competitive advantages.
PROBLEM AND MODEL STRUCTURE
Since several years, the management of technology and innovation is an ongoing
research project at the Industrieseminar. Several papers were presented at System
Dynamics Conferences reporting on findings about appropriate policies and methodo-
logical developments. One of the first applications was devoted to pricing strategies in a
dynamic environment (Milling 1986a, Milling 1986b).
PRICING STRATEGIES AND THE DIFFUSION OF INNOVATIONS
Pricing new products is an essential but largely unresolved problem of innovation man-
agement. Peculiar difficulties result from the dynamics in demand interrelations, cost
development, and the risk of substitution through more advanced products. In an early
version of the innovation model, several price setting mechanisms were included for
direct investigation:
358 SYSTEM DYNAMICS '93,
(1) Myopic profit maximization where there is perfect information on the’ current
state of cost and demand. The optimal price is derived from elasticity of demand ¢, and
per unit cost c which depend on long run experience effects and on short term capacity
utilization:
é
opt std fi
=e: 1
pore i 1 (1)
1
(2) Skimming price strategy with the objective of serving first customers with high
reservation prices and subsequent price reductions. The-model applies a simple decision
rule modifying p?” through a function of market saturation ms:
skim
pe” =p" (1+ f(ms)) @
(3) Penetration: pricing aims at rapidly reaching high production volumes to benefit
from the experience curve and to increase the number of adopters. It sets prices accord-
ing to:
Pr" =p -(1- f(ms)) GB)
In the dynamic environment under investigation the classical pricing rule for profit
optimization turned out to be superior to the skimming strategy. The appropriate strat-
egy - as suggested by these results - constitutes the attempt to rapidly penetrate the mar-
ket. This objective is achieved by setting relatively low prices, especially in the early
stages of the life cycle, and by providing sufficient production-capacity for immediate
delivery. Temporary excess capacity hurts the financial performance less than longer
delivery delays. The combined price and diffusion effect stimulates the environmental
demand dynamics and reduces the risk of loosing. potential customers to upcoming sub-
stitution products.
COARSE STRUCTURE OF THE COMPREHENSIVE INNOVATION MODEL
Frequently only the market stage, during which the product is sold, is associated with the
notion of an innovation. However, before the availability of a marketable product the
costly, lengthy and risky period of research and ‘development has to be passed success-
‘fully. While the market cycle tends to become shorter and to reduce the time for the cor-
porations 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 (Milling 1991a).
The comprehensive innovation model consists of two modules: one reflecting the pro-
cesses of R&D, the other representing the market cycles. Figure 1 shows the structure of
the overall model and its components. Both modules are linked through flows of infor-
mation to monitor the resource allocation, the intensity of the R&D-processes, the
SYSTEM DYNAMICS '93 359
required minimum quality level before a new product is considered ready for market
introduction, etc.
Comprehensive Model of Innovation Generation and Diffusion
Module of the R&D Process Corporate and Market Module
‘Sector of Innovation Diffusion
incl.
Competition & Market Entry
‘Sector of Accounting
& Corporate Planning,
incl.
Sector of
R&D Controlling
Fig. 1: Coarse structure of the comprehensive innovation model
The module of the research and development phase deals largely with intangible proc-
esses. 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
the task, 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 solu-
tions. The respective results are evaluated on the basis of viability. If they seem to be
superior to. previous 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 (Milling 1991a, Milling 1991b, Maier 1992).
The corporate 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.
360 SYSTEM DYNAMICS '93
CONSEQUENCES OF PRICING STRATEGIES
IN THE COMPREHENSIVE INNOVATION DIFFUSION MODEL
ELEMENTARY FEEDBACK STRUCTURE OF PRICING,
R&D BUDGETING AND SALES
The first step in the analysis of the model behavior is the investigation of the feedback
structure of pricing strategies, R&D budgeting, market entry time and the diffusion of
innovations. (cf. Fig: 2). The central part of the market module is an equation that
determines a company's sales volume per period through addition of innovative and imi-
tative purchases and therefore the diffusion of innovations (Bass 1969, Milling 1986),
Innovative purchases are calculated as the product of the coefficient of innovation (INC)
— this is the fraction of innovators — and the number of potential customers (POTCUST).
Innovators buy a new product because they have a general interest in innovations. In
contrast, imitators. are influenced in their purchasing decision through the number of
customers who already bought the product, the so-called adopters. The imitators are
computed as the product of the. coefficient of imitation.(IMC), the potential customers
and the adopters. The coefficient of imitation (IMC) defines the probability that the
communication between adopters and potential customers — expressed through the term
(POTCUST*ADOPTER) — causes the purchase of a product.
i standard cost Prive
&] |
Pricing strategy
; rave pie
SALES ( =| INC()|* POTCUST (+ demand elasticty’
Dollar volume
Mee <=}! *— | micio|+ porcust «+ ADOPTER()
| fl
rultiplier of ;
hacthas Hey rel. competitive effect of price
ers is] advantage
se | 3
R&D volume and + flav technical < —_ technical Soinpedions pice ———
intensity know how
ses set readiness for
- market
(qnarket entry ime)
Fig..2: Feedback structure influencing the diffusion process
The first loop describes the feedback relations between the. sales of a: product, the
R&D process and the effect of relative competitive advantage. With an increasing. num-
ber of sales.volume and a growing dollar volume of sales the: R&D budgets and the size
of R&D personnel grow larger. By the way of enhanced higher technical knowledge this
SYSTEM DYNAMICS '93, 361
cause a stronger competitive advantage. The higher the sales volume, the better is the
resulting competitive position. That produces increasing coefficients of innovation and
imitation and finally leads to higher sales volume again. The sales oriented R&D budget-
ing strategies implemented and described here cause positive feedback (Maier 1992).
The second feedback loop shows the influence of pricing strategies on sales volume.
The actual price of a product is influenced by three factors. The first factor, standard
costs, is endogenous. The second and third element influencing the calculation of prices
are exogenous elements: the pricing strategy and the demand elasticity. Standard costs
are the basis for the calculation of the prices for each pricing strategy. They depend on
the. cumulative production of a product, influenced by the actual sales volume. Higher
cumulative production causes experience. effects that reduce the standard costs and
therefore the basis for pricing. Lower prices themself affect: relative price and improve
the effect of price on.the coefficients of innovation and imitation. Higher coefficients
again produce increased sales.
The price level depends on the pricing strategy. The model includes alternative pricing
policies like (1) the strategy of myopic profit maximization, (2) the strategy of skimming
pricing or (3) a penetration pricing strategy. Demand elasticity determines the profit
margin for the first three strategies and therefore the price. Feedback loop 3 shows the
effect of pricing on the dollar volume of sales. Higher prices cause, under the assumption
of a constant sales volume, an increasing dollar volume of salés, with all the conse-
quences on the R&D process, the technical know-how and the market entry time as
shown in the first feedback loop.
BASIC BEHAVIOR OF THE MODEL
To show the results of the analysis, first a short description of the model capabilities and
the general assumptions of the model runs will be given. The model maps the structural
fundamentals of two competing companies — including all policies of pricing, budgeting
for R&D and corporate planning — as well as the structure of the markets of successive
product generations.
For the following analysis of pricing strategies, it is assumed that the initial situation is
identical for both competitors. At the beginning of the simulation, both companies have
already launched the first product generation into the niarket. Corporate R&D influences
the technical knowledge of actual and potential products. Through corporate R&D it is
possible to develop improved and substituting product generations. New products are
introduced to the market if the technical know-how passes a threshold value. The
resources for research and development derive from older successful products. The total
amount of resources spent on corporate research and development is calculated as a
fixed percentage of dollar volume of sales.
This sales oriented R&D strategy — it is'activated in all model runs — produces posi-
tive feedback (Loop 1 in Fig. 2). With equivalent initial situation and the same set of
strategies both companies behave in an identical way for all product generations, except
some minor stochastic differences caused by the evolution algorithm modeling the R&D
process. If one company has a competitive advantage, e.g., through earlier market entry,
a concentration process will be initiated and continued that causes earlier readiness for
362 SYSTEM DYNAMICS '93
market entry and increased sales for all successive product generations. The competitors
with the advantage will improve continuously (Maier 1992).
CONSEQUENCES OF PRICING STRATEGIES IN THE
COMPREHENSIVE MODEL
The analysis of the following model runs will show the impact of different pricing strate-
gies on the process of R&D and the diffusion of an innovation. In the different model
runs the first company uses the strategy of skimming price; alternatively the competitor
uses skimming price strategy in the first model run (basic run). In the second. and third
strategy run he uses myopic profit maximization strategy and the strategy of penetration
prices. For all product generations the pricing strategies are the same. The demand elas-
x. : : z j
ticity ¢ = has the value -2.
x
P
P
Dollar volume of sales [1e6]
600
‘© Firm 1/ skimming © Firm 2/ skimming
4 Firm 1/ skimming + Firm 2/ optimal
= Firm 1 / skimming + Firm 2 / penetration
0 24 48 2 96 120
Time
(Months)
Fig. 3: Dollar volume of sales for the different pricing strategies
Exhibit 3 shows the time path of the dollar volume of sales for both competitors in
different model runs. The superior strategy is the penetration strategy of company 2. The
dollar volume of sales of firm 2 is nearly 44% higher than the first company's. The sec-
ond best strategy is the strategy of myopic profit maximization with a sales volume, that
is only 2% lower than in the run with penetration pricing. Compared to, the first com-
pany,.in this run the volume.is 21% higher. Exhibit 3 also shows that in the case of
SYSTEM DYNAMICS '93 363
skimming price strategy for both companies a relatively high volume is reached. The dif-
ference between the best strategy and the skimming price strategy is only 7%.
The variable market position gives an aggregation of a company's products market
share. Values greater than 1 mean that the market-position is better than that of the
competitor. Exhibit 4 the time path. of the market-position is shown. Running the model
with a parameter set where both competitors use the strategy of skimming prices, there is
no effect on the market-position, the value is 1 for both firms during the whole simula-
tion. If the second company is running a strategy of optimal prices or penetration prices
it improves its market-position until period 30 respectively 32 — when firm 1 launches as
a pioneer the second product generation. Firm 2 loses market-position ‘caused by the first
firm's competitive advantage of early market entrance for the second product generation.
After period 40 the second company is able to improve its market-position again through
the better effect of price and the higher value of the multiplier of competitive advantage.
In period 56 respectively 58 both competitors launch the third product generation. Tak-
ing the changing market-position as the measure for the quality of a strategy one can
state that again the strategy of penetration pricing is the best.
is market-position _
‘© Firm 1/ skimming © Firm 2/ skimming
Firm 1 /skimming - Firm 2/ optimal
1a Firm 1/skimming Firm 2/ penetration
Fig. 4: Market position for'the different pricing strategies
Looking at cumulative profits the result is different. In the basic run of the model,
where both companies are using a skimming price strategy, cumulative profits reach the
highest level. The second best solution in terms of cumulative profits is the strategy of
‘optimal prices. After period 110 the second firm passes the first company and finally
reaches a value that is only 9% lower than its. cumulative profits i in the basic run. Run-
ning the model with company 2 using the strategy of penetration pricing the first com-
pany is leading nearly almost period 120, At the end of the simulation, company 2-makes
up the first firms small advantage. The final level of cumulative profits is 24% lower than
in the run with skimming prices.
364 SYSTEM DYNAMICS '93
10
cumulative profits [1¢7]
‘© Firm 1/skimming -@ Firm 2/ skimming
8 3 Firm 1 / skimming * Firm 2 / optimal
= Firm 1 / skimming © Firm 2/ penetration
38 2
© Time
(Months)
96
Fig. 5: Time path of the cumulative profits
120
The last variable to investigate is the readiness for market entry. The companies
launch new products if the technical: know-how incorporated in a product exceeds a
critical value. In the basic run of the model both firms introduce their second product
generation to the market at period 29; the third product generation follows in period 55.
The model runs clearly show that pricing strategies have an impact on market entry time
(Fig. 6). In the case of the profit maximization strategy the second firm's dollar volume.
of sales is lower than the first firm's. That causes — compared to firm 1 - lower R&D
budgets and personnel, consequently reduced R&D volume and intensity and less techni-
cal know-how. This finally produces the delay in market entry time for the second prod-
uct generation shown below.
product generation 2 product generation 3
delay compared delay compared
pricing strategy firm 2 } firm 1 | firm 2 to firm 1 firm 1 | firm 2 | to skim.-strategy
skimming price 29 29 0 55 55 0
profit maximization 29 30 1 56 56 1
penetration price 29 32 2 58 58 3
Fig. 6: Consequences of pricing strategies on market entry time
Although the first firm has a competitive advantage resulting from its earlier market
entry, firm 2 realizes — due to the higher effect of price on the coefficients of innovation
and imitation — an increasing sales volume (see Fig. 3). Considering the third product
SYSTEM DYNAMICS '93
generation, the second company’s higher sales volume allows it-to make up the first firm's
advantage in market entry time. In comparison to the basic run of the model, there is a
delay in the readiness for market entry for both firms. From this point of view, the
skimming price strategy is the superior one, followed by the strategy of myopic profit
maximization and penetration pricing.
The last analysis shows that the results of the simulations and the profitability of a
strategy varies with initialization or parameterization. Under the assumption of skimming
price strategy for firm 1 and penetration price strategy for firm 2 the model has been run
with changed demand elasticity. Figure 7 summarizes the results of the different runs
were demand elasticity ¢ varies from -3.2 up to -1.2.
SE
aE
ae
By
ge
#8
aC
bE
Fig. 7: The impact of different demand elasticities on relative cumulative profit ratio
Due to the different profit margins — resulting from myopic profit maximization that is
the basis for price calculation — the use-of the absolute value of the cumulative profits is
not appropriate. For the evaluation of the runs, the relative cumulative profit ratio is
cum, profits firm 1 - cum. profits firm 2 a 0. The exhibit shows,
cum. profits firm 1
that with increasing demand elasticity the initial disadvantage of the second company but
also its chance of gaining an advantage rises. In the case of lower demand elasticities
(€ > -2) firm 2 has a diminishing but still existing disadvantage during the whole simula-
tion. : : “
computed. as
366 SYSTEM DYNAMICS '93.
CONCLUSIONS FOR PRICING
The model runs have shown that the judgment of strategies depends on the objectives of
a company. If a firm wants to enhance its dollar volume of sales or the market-position,
the strategy of penetration pricing is the superior one; but in terms: of sales volume there
is only a marginal difference between profit maximization and penetration pricing strat-
egy. Viewing cumulative profits and the readiness for market entry the strategy of
skimming prices is the best. The evaluation of an optimal strategy is not possible. The
outcome of a strategy and therefore the choice-of'a strategy depends on to many factors
that influence the diffusion process.
The results show clearly the relativity of the judgment of strategies. Neither’ optimal
solutions nor generally valid solutions can be found. Optimization algorithms must fail.
The model must always fit the unique characteristics of the problem under investigation.
REFERENCES
Bass, F. M. 1969. New Product Growth Model for Consumer Durables. Management
Science 15:.215 - 227.
Maier, F. H. 1992. R&D Strategies and the Diffusion of Innovations. In Proceedings
System Dynamics 1992, eds. Vennix, J.A.M.; J. Faber; W.J. Scheper; C.A.Th.
Takkenberg. Utrecht: 395 - 404.
Milling, P. M. 1986a. Diffusiontheorie und Innovationsmanagement. In Technologie-
und Innovationsmanagement, ed. E. Zahn, Berlin: Duncker & Humblot: 49 - 70.
Milling, P. M. 1986b.-Decision Support for Marketing New Products. In System Dynam-
ics: On the Move, eds. Aracil J.; J.A.D. Machuca; M. Karsky. Sevilla: 787 -.793.
Milling, P.M. 1991a. An Integrative View of R&D and Innovation Processes. In Mod-
elling and Simulation 1991, ed. E. Mosekilde. San Diego, CA: 509 - 514.
Milling, P. M. 1991b. Strategische Planungs- und Kontrollsysteme zur Unterstiitzung
betrieblicher Lernprozesse. In Systemmanagement und Mangementsysteme, ed. P. M.
Milling. Berlin: Duncker & Humblot: 11-31. ©
SYSTEM DYNAMICS '93 367
