System Dynamics '91 Page 219
DIFFUSION ANALYSIS IN MARKETING
M.GOVINDARAJAN N.RAMASWAMY
Senior Lecturer, Professor
Management Division, Mechanical Engineering Dept.
College of Engineering, Indian Institute of Technology,
Madras 600 055, Bombay 400 076,
India. India.
ABSTRACT
The system dynamics technique has been adapted to reflect the dynamics of the diffusion
process from both behavioural and marketing strategy perspectives. The diffusion theory
has helped in conceptualising and explaining the information process and in evaluating
the various market interventions in the spread of new product innovations. The diffusion
and adoption process has been studied with particular reference to the personal computer
in India.
INTRODUCTION
Diffusion is the process by which an innovation spreads. It is the spread of a new
idea from its source of invention or creation to its ultimate users or adopters.
The conceptual foundations (Gastignon, 1985) of diffusion theory are:
* The concept of innovation
* The diffusion process over time
The personal influence and opinion leadership process
* The adoption process
The role of the innovators and other adopter categories
The social system or market segment within which diffusion occurs
To these could be added
* The role of marketing (change agent) actions
* The role of competitive actions
A firm that is preparing to introduce for the first time a new consumer durable,
would like to make an estimate of
* the market potential
* the shape of the likely approach to this potential and
the rate of approach to this potential
Two analytical frameworks are available. New product sales can be thought of in
terms of ©
*
1. Diffusion process
2. Adoption process
The adoption process, focuses on "the mental process through which an individual
passes from learning about an innovation to the final adoption". Adoption itself is
the act of buying the product in the case of nonrepurchasable products, or the
decision to use the product regularly, in the case of repurchasable products (Kotler,
1971).
Page 220 System Dynamics '91
A major difference between the diffusion process and the adoption process is that
diffusion occurs among (groups) persons while adoption is an individual matter.
UNDERLYING THEORY
Diffusion research in marketing has placed its major emphasis on the characteristics
of consumer innovators. Variables most likely to characterize consumer innovators
for technology-based products are higher income, higher education,.a favourable
attitude towards risk, related knowledge, experience-base for similar technological
innovations and so on. Another emphasis has been on the characteristics of innovations
that affect the speed of diffusion.
The marketing literature has pursued a further interest. The interest is in the
adoption process and the hierarchical stages of adoption through which the decision
maker passes.
A high involvement innovation (such as new technology) is a high risk/high information
need innovation. .For low involvement innovation, howeyer, adopters may skip the
information processing and evaluation stages and go directly to trial
Higher Involvement Low Involvement
Learning Hierarchy Hierarchy
Adoption Adoption
Trial Trial
Evaluation
Interest
Knowledge
Avvaseness
Acceptance of an innovation depends very much upon the prospective adopter’s
culture. Thus closed or traditional systems were found to hinder diffusion while
open or modern social systems appear to stimulate it (Baker, 1983).
Following Bernhardt and Mackenzie(1972) the diffusion process for economic goods
can be defined as the adoption over time of a specific product, by customers who
are linked by channels of communication to a given social structure and by a given
system of values or culture. The adoption process is influenced by the product, the
potential adopter’s characteristics, his linkage to a social structure, change agents and
the adopters culture. Different assumptions regarding these may lead to different
diffusion models.
If adoption of an innovation is plotted against time from introduction to complete
diffusion it will assume the characteristics of a normal distribution, or if plotted
cumulatively, assume the familiar S-shaped curve which characterises the product
life cycle curve.
System Dynamics ‘91 Page 221
The diffusion process can be characterized in terms of three dimensions:. the rate
of diffusion, the pattern of diffusion and the potential penetration level.. The rate
of diffusion reflects the speed at which sales occur over time. The potential
penetration level is a separate dimension indicating the size of a potential market.
That is, the maximum cumulative sales (or adoption) over time (Gastignon, 1985).
OVERVIEW OF MODEL AND FRAMEWORK FOR ANALYSIS
A generic system dynamics model is drawn to study the diffusion process and dynamics
of the personal computer market in India (Govindarajan, 1990a). The personal
computer market is not described here. However, the generic nature of the complex
interactions is depicted in the figure. The analysis was carried out using the following
framework (Govindarajan, 1990b).
AWARENESS DIFFUSION:
Awareness diffusion is controlled by advertisement and the rate of growth of the
potential adopter population is controlled by price. In the case of durable goods
the adoption process is related to observed sales.
The adoption rate X(t) is
X(t) = [N- X(t)] [a + bX(t)]
where
N is the population size
X (t) is the number of adopters by time t.
The implicit assumption in the above equation is that the population is homogeneous.
This is a. good approximation for diffusion of information. The conditional likelihood
of adoption is increasing linearly in the number of existing adopters a + b x (t),
where a and b are parameters.
= [N(@P)-X [a + bx]
The size of the potential adopter population is a function of price, while the
conditional adoption. rate still depends linearly on penetration level.
If a = O in equation (1), the information is spread: only by word-of- mouth, where
the seed is the initial adopter population. The rate of adoption is then determined
by information diffusion and price changes over time.
The degree of awareness is a discrete function. Either the customer is aware or
not aware of the product. In general, the awareness of a consumer in time ‘t’ may
be described as DA(t)=f (awareness in previous time period, competitors action,
word-of-mouth, Retailer efficiency, advertisement and demonstrations).
Diffusion is a reflection of the information propagation or learning process with firms
adopting a technology as they come to learn of it existence (epidemic learning
models).
The above equations can be rewritten as
axe = [a+ bN(t)] [NNO]
Where a, b and N remain constant over the entire substitution process. That is to
say, the diffusion spreads through a static environment. As stated earlier, this
assumption is tenuous. Due to changing characteristics of the potential adopter
Page 222 System Dynamics '91
population, technological changes, product modifications, pricing changes, general
economic conditions and other exogenous and endogenous factors, the parameters,
a,b and N are more likely to change over time.
The number of potential adopters N is influenced by the product’s price and the
effectiveness of promotion in establishing multiple uses for the product. The coefficient
b, the rate of influence of adopters on non- adopters will be influenced by the
product’s capacity to stimulate favourable word-of-mouth comments. The rate of
influence of advertisement, a, on nonadopters, will be influenced by the size and
effectiveness of advertising budget.
Buyer POOL
The number of buyers is influenced by the size of the pool of potential new adopters
and repeat buyers. It is also influenced by expected cost. The potential pool of
new adopters is increased by effective promotion and favourable word-of-mouth
interaction of past satisfied buyers with market place. However, the potential pool
of new adopters is also drawn down by the interaction of the likely buyers with
the past unsatisfied buyers. The number of buyers are then partitioned and accumulated
into satisfied past buyers and unsatisfied past buyers according to the quality of
experience index, which has a value between zero and one. The value of quality
of experience index implies the percentage of buyers satisfied.
PURCHASES
The purchase process can be shown diagrammatically as follows:
STATUS and STATUS OF PURCHASE
LOCATION —> |BUYER —»+ |FELT NEED} -* | DECISION
OF CONSUME! BEHAVIOUR
| Actual Purchase! | [sales] |
The status of the purchaser includes middle, upper and upper-upper class. The
states of buyer behaviour are awareness, knowledge, liking, preference and conviction.
The factors which affect the purchase process are of two types: external and internal.
The internal effects are due to the interaction of the classic 4 Ps of marketing
product, price, promotion and place. The external effects may be general economic
conditions, technical developments and so on. The felt need and actual purchase
can be described as
Felt Need = F (status and state of the mind of buyer, price of the
substitute product, Demonstrations, previous felt need and
status)
Actual Purchase = F (Purchase decision and Retailers efficiency)
Purchases are of two types: New and Replacement. New purchases reflect the selection
of some fraction of the stream of new purchase decisions in favour of the new
technology. Replacement purchases represent purchases to replace products discarded
for reasons of wear and tear. For one-time buyer replacement will not be indicated.
Given the product availability, the purchaser fraction will be identical to the fraction
of potential (acceptor fraction) purchases.
System Dynamics '91 Page 223
ADOPTION PROCESS
The new product is first adopted by innovators. Any adoption requires the actual
purchase of the product. The innovators’ social contact (or positive word-of-mouth)
induces others (a fraction) to adopt the product (imitators). In both the cases, if
the post-adoption behaviour and/or attitude is positive (favourable towards the
product), then the consumer remains delighted. Otherwise, the product is-abandoned,
thereby increasing the non-acceptor level. The company’s promotional marketing
efforts are aimed at keeping the non-acceptor level to a minimum and if possible,
increasing the acceptor level. The acceptance and extent of use of the product are
subject to change over time, mainly due to performance-related information, And
performance is related to changes in technology or its application (Homer, 1987). —
Acceptance is enhanced when information about the product and its performance 1s
favourable. The volume of information includes personal, professional and commercial
sources of communication in keeping consumer aware.
PROMOTIONAL EXPENDITURE
A percentage of sales revenue is spent on promotional marketing and product
development. Sales have an effect on the promotional budget that, is, the next
year’s plans depend very much on this year’s sales. Thus a company that is growing
due to its promotion efforts tends to reinforce its growth by expanding its budget
for the next year each time its sales improve. The company’s market share is affected
by the rate of introducing new products when early introduction gives the firm a
leading edge in promotional benefits.
PRODUCT DEVELOPMENT
The durability of a good has two implications. First, it can be stored in inventories.
Second, if it provides a stream of services to customers, they may wish to defer
purchases to take advantages of price fluctuations (David Rink, 1979).
New product strategy requires a historical base for assessment. Only an understanding
of a firms past product activities can provide the full context by which to evaluate
the challenges posed by its next set of new products. The individual product strategic
elements can be classified, as falling into four clusters involving the nature of
Products Developed
Markets sought
Technology Employed and the
New Product process
These four elements suggest that measurable comparisons of present to past products
may be performed along two basic dimensions. The first dimension- is the newness
of the technology within the new product relative to technologies already developed
by the firm. The second dimension is the newness of the market application for
which the new product is targeted compared with the users of past products.
The degree of newness in market application includes levels of newness regarding
product packaging, buyers, distribution channels, and support mechanisms. As each
new product comes on stream, the cumulative body of the firm’s technology and
market experience grows accordingly, and becomes much broader for the evaluation
of the next new product effort.
A market transition process for consumer durable takes place over time. In general,
product differentiation is likely to take place only after demand growth slows due
to product saturation. Competitive pressures are also likely to assist the differentiation
Page 224 System Dynamics '91
process. The introduction and diffusion of new products follow the maturation of
older, but related, products within the same market. The process occurs over time
and it is crucially important in understanding the differentiation process.
PRICING
Consumers cannot directly verify product quality prior to an initial purchase. Instead,
consumers base initial purchases on observed prices, which perfectly signal firm’s
qualities both in and out-of equilibrium. | Equilibrium quality is less than efficient,
but generally bounded away from the minimum quality. With free entry, observable
product. variety exceeds what-would prevail with perfect information.. As repeat
purchases become large relative to initial purchases, or as firms become small relative
to the size of the market, equilibrium product quality rises, and the market converges
to the full information equilibrium.
BEHAVIOURAL FACTORS
The inclusion of qualitative factors becomes necessary in an industry which is in the
process of consolidation on the one hand, while witnessing rapid changes on the
other, as new markets open and new technology appears with increasing regularity.
The image and credibility of the company, its marketing clout, reputation have all
to be included in the model. The model may attempt to include the various
dimensions of an innovation also.
MODEL DESCRIPTION
A generic system dynamics model is attempted to reflect the diffusion pattern and
the dynamics of the India micro market. Details and data have entirely been culled
out from various Indian computer magazines and business newspapers. Wherever
needed, the data have been assumed in consistence with the prevailing situation in
the Indian market. The model is shown in Figure in the form of causal loop diagram.
The model is based on the philosophy of Morecroft’s(1986) and Homer’s (1987)
models. As the proposed model is. still under progress and view, only the salient
features are described.
The model starts with an awareness diffusion model. With a slight modification to
the Tourism model of Jambekar ‘and Brokaw (1989), the potential buyer pool can
be thought of as
L PBPOOL.K = PBPOOL.J + (DT) (NRADVT.JK + NRSFCT.JK +
POSINT.JK - NEGINT.JK - AADOPT.JK)
PBPOOL = Potential buyer pool (Buyer)
NRADVT = Number reached through advertising (Buyers/year)
NRSFCT = Number reached through sales force contact (")
POSINT = Positive interaction between adopters and potential buyers. (")
FORGET = Number forgetting : (@)
NEGINT: = Negative interaction between unhappy adopters and potential
buyers (’)
AADOPT = Actual adopters who have purchased the product (')
The product diffusion spreads through a combinations of several communication
processes - advertising, peer group discussions and sales force contact. NRADVT
is a function of target audience and advertising expenditure. NRSFCT is a function
of the number of sales-person allocated to promotion and their contact rates. The
satisfied users and the potential buyers positively interact in an additive way while
System Dynamics '91 Page 225
the unsatisfied users and potential buyers negatively interact in a multiplicative way
and the respective coefficients of interaction have been assumed. A fraction of the
people who are aware of the product become actual adopters.
The: New product development section takes care of the product developments, which’
are the outgrowth of projects costing thousands of rupees in Rand D. activity and
require years to complete.
L PRODEV.K = PRODEV.J+(DT) (PRODSR.JK-PRODCR.JK)
R PRODSR.KL = MAX(0,PRODCR.JK + (IPRODE.K-PRODEV.K)/PRODAT)
R PRODCR.KL = PRODEV.K/PRODCT.K
A IPRODE.K = (IFSRPD.K) (AVGREV.K)/SPRODT
A PRRTPD.K. = PECUR.K/RINCPD.K
A PFCUR.K = SMOOTH (CHFUR.JK/URATE,TPUCUR)
A RINCPD.K = SMOOTH (INCPD.K INCST) PRODEV product development
projects (Projects) :
PRODSR = Product development start rate (projects/year)
PRODCR = Product development completion (’)
IPRODE = Indicated product development (Projects)
IFSRPD = Indicated fraction of sales revenue to product development (0-1)
PRRTPD = Perceived return to product development (1/Project)
PFCUR = Perceived fraction of change to usage rate (0-1)
RINCPD = Recent incorporation of product developments (Projects/year)
CHGUR = Change in usage rate (Usage index)
URATE = Usage rate (’)
The functional capability is modeled as
A FUNCAP.K = (PROCAP.K) (EEXPEC.K)
Where the product capability (PROCAP) is a level variable and the effect of
experience on functional capability (EEXPEC) depends on the relative skill of the
average user (REAUSR).
The sales lost to new competition and the growth of competition are modeled in
the competition sector.
R SLOSNC.KL = SALES.K/NORTIM*FMKINC.K
A FMKTNC.K = TABHL (FMTKN, MKSTNC.K,1,.1)
A MKSTNC.K = (NCOMPC.K/ (MNCOMP.K + NCOMPC.K)* SWNCOM
L NCOMPC.K = NCOMPC.J +(DT)(INCOM.JK-FNCOMP.JK-MNCOMP.JK)
SLOSNC = Sales lost to new competition (Units)
FMKTNC = Fraction of market lost to new competition (Dimensionless)
MKTSNC = Marketing strength of new competition (@)
NCOMPC = New competition capacity : (Projects/year)
INCOMP = Increase in new competition (0)
FNCOMP = Failure of new competition (')
Page 226 System Dynamics '91
MNCOMP = Maturing of competition (‘)
In its entirety the model will consist of the. following sectors:
Buyer pool, company, promotional marketing, new product, product development,
competition, customer, corporate, user perception and summary statistics.
CONCLUSION
A system dynamics framework is suggested for modeling the innovation diffusion and
adoption process of a hightech product. Normally, using customer needs as the
foundation for marketing in hightech products is problematic, because potential
customers often cannot articulate what they need. Marketing efforts should therefore,
represent not only promotional but detailing activities also to overcome the marketing
and technological constraints faced by such products. Sales, though a significant
factor is seldom enough of a determinant about a company’s intrincsic value standing
in the market. Hence the behavioral aspects are emphasized. Inspite of ambiguities
and paradigmatic problems, a workable model has been evolved. The. parameter
fixations and the unknown values of TABLE functions are fraught with such problems.
However, the model responds, rather reasonably well and is able to reproduce the
state history.
System Dynamics '91 Page 227
References
Baker Michael. J, 1983 ‘Marketing Theory and Practice’, Macmillan press,
London.
Bernhardt. I and Kenneth. D. Mackenzie, 1972 ‘Some Problems in Using
Diffusion Models for New Products’, Management science, Vol 19,
No.2, 187-200.
David Rink and John Eswar. 1979, ‘Product Life Cycle Research -.A Literature
Review’, Journal of Business Research vol. 7, 219-242.
Gastignon Hubert and Thomas S. Robertson, 1985, ‘4 Propositional Inven-
tory of New Diffusion Research’, Journal of Consumer Research vol.
11, 849-867.
Govindarajan. M. and Ramaswamy. N. 1990a, ’Discontinuous Innovation
Diffusion Analysis’ , International System Dynamics Conference, Bos-
ton, 458-471.
Govindarajan. M. and Ramaswamy. N. 1990b, ‘Diffusion Analysis Using
System Dynamics’, Proceedings of the IV National Conference on
System Dynamics, December 1990, Tirupathi, 87-98.
Homer Jack B. 1987, ‘A Diffusion Model with Application to Evolving Medical
Technology’, TFSC, 31(3), 197-218.
Jambekar, Anil B and Brokaw, Alan J., ‘4 Tourism Systems Dynamics System
Dynamics’ An International Journal of Policy Modelling. Vol.II No.I.
1989. 1-11.
Kotler Philip, 1971, ‘Marketing Decision Making - A Model Building
Approach’, Holt, Rinehart and Winston.
Morecroft, JDW, ‘The Dynamics of a Fleding High Technology Growth Market
- Understanding and Managing Growth Cycles’. System. Dynamics
Review. Vol.2. No.1. Winter 1986. 36-61.
CC BY-NC-SA 4.0