DISCONTINUOUS INNOVATION DIFFUSION ANALYSIS
M Govindarajan and N Ramaswamy
Mechanical Engineering Department
Indian Institute of Technology
Bombay 400 076, INDIA.
The market place derives its dynamism from the inherent
willingness of a consuming population to innovate. Many technological
firms have been exploiting the consumer markets with their technology
based discontinuous innovations. Several companies have been marketing
small computers that, in pricing and programming structure, are amenable
to adoption by individual consumers. This study is an attempt to study
the diffusion/adoption process of personal computers in the Indian context
from both a behavioural theory and marketing strategy perspective.
INTRODUCTION
The overall pattern of the technological innovation diffusion
process can be thought of as a complex net of communication paths
linking the various stages of the process. The basic constructs of
diffusion - the innovation, adopter categories, the adoption process,
personal influence and the pattern of diffusion have helped in conceptua-
lizing the information dissemination process and in suggesting the value
of particular change-agent interventions.
DISCONTINUOUS INNOVATION
Anderson and Ortinau (1988) have adopted an innovation classi-
fication scheme based on behavioural interpretations. The product
innovation, when perceived by consumers as being a new _ product
established by a major technological advancement, is known as a dis-
continuous innovation. Such product innovations are marked by the
severity of learning requirements for the adopters as they result in
entirely new product categories. The product search attributes can be
highlighted by product display and advertising. The experience attri-
butes can only be acquired by constant use of the said products. Hence
it is difficult to forecast the diffusion rates since it may be necessary
to ‘educate’ potential customers about the new technology before they
can evaluate it and deduce a judgement of desirability.
A technological innovation creates one kind of uncertainty in
the minds of potential adopters (about its expected consequences), as
well as representing an opportunity for reduced uncertainty in another
sense (that of the information base of the technology). The latter type
of potential uncertainty reduction represents the possible efficacy of
the innovation in solving an individual's felt need or perceived problem.
This impels an individual to exert effort in order to learn about the
innovation. Once such information seeking activities have reduced the
uncertainty about the innovation's expected consequences to a tolerable
level for the individual, a decision concerning adoption or rejection
will be made (Rogers, 1983).
456
System Dynamics '90
LITERATURE REVIEW
A brief review of relevant literature is presented. The funda-
mental diffusion model (Bass, 1969) can be written as
SAG) = aE NCH) ~ NCE) J + bNCE) ENCE) - NCE) I
where
a is the coefficient of innovation
b_ is the coefficient of imitation
{N(t) - N(t)] is the number of potential adopters available
at that time
The above equation can be rewritten as
aN(t) _ y
“ar = g(t) { N(t) - N(t) J
and
N = N at t = t
oO oO
The value of g(t), the product growth coefficient, depends on the speci-
fic product innovation, the marketing system in which the innovation
is diffused and other elements such as the channels and change-agents
(Mahajan and Peterson, 1979).
In dynamic diffusion models,
N(t) = f[S(t)]
where S(t) is a vector of all relevant exogenous and endogenous vari-
ables, both controllable and non-controllable, affecting N(t). These
factors include socio-economic conditions in the marketing system, price
changes, government actions, marketing efforts and so on.
Dodson and Muller (1978) incorporated the effects of adverti-
sing and word-of-mouth in the dynamic diffusion model. They suggested
that the population of potential customers [IN(t) - N(t)] be segmented
into a subgroup of consumers who know of the new product but do not
buy it and a subgroup that knows and buys. Advertising was hypothe-
sized to be responsible for the adoption decision of the second subgroup.
To represent the dynamic nature of the market potential N(t), or the
increase in the market potential, N(t) - N(t), two mechanisms by means
of which the group [P(t) - N(t)] be made aware of the product, namely,
advertising and word-of-mouth were suggested. The diffusion model is
a{P(t) - N(t)} + bN(t) EP(t) - Nt] -
CIN(t) ~ N(t)] - @{N(t) - N¢t)]
agIN(t) - N(t)]
dt
System Dynamics '90 457
where
is the coefficient of advertising
is the coefficient of imitation
is the effect of advertising
is the forgetting coefficient
(t) is the population of a marketing system at time t
yao oD
Shariff and Ramanathan (1982) studied the time pattern of the
spread of innovations when the population was no longer binomial but
of a polynomial type composed of adopters, rejectors after adoption,
disapprovers and the uncommitted.
Shlomo Kalish (1985) opined that the rate of adoption was
determined by awareness diffusion and the rate of growth of the poten-
tial adopters, each controlled by advertising and price respectively.
The process of becoming aware was modelled as a simple epidemic type.
The adoption, being conditional on awareness, occurs when the value
of the product exceeds its selling price. In the case of durable goods,
the sales rate equals the adoption rate. Each individual adopts one
unit once. :
Homer (1987) developed a system dynamics model to study
the diffusion of medical technologies. The simulation model addresses
both the adoption and extent of use of a technology product and endo-
genously accounts for changes in actual and perceived performance.
PC MARKET - THE INDIAN SCENE
The personal computer appeared on the Indian scene sometime
during the end of 1985. Though IBM PC is not marketed in India, it
is widely copied by over 50 vendors who sell the IBM PC clones. They
account for 50 per cent of the PC market.
In terms of volume, the number PCs that have been sold in
India during 1985-86 is 15000; 1986-87 is about 20000; 1987-88 is 22000;
1988-89 is 50000. While the world-wide trend in PC sales is expected
to cross the 10 million mark in 1990-91, the Indian manufacturers would
be lucky if their combined sales cross the one lakh mark.
The PC prices which crashed in 1986 when the then government
declared its attachment for the computers, started rising from late 1987
through the first quarter of 1988. By June they had begun to touch
the pre-1986 rates. In 1987-88 PC clone prices were up (due to budget)
by about 33 per cent with increased cost of peripherals and services.
The list price increased but the margin decreased for the manufacturers
and dealers. The hike had to do more with international compulsions
rather than local pressure. Due to a sudden and severe recession of
DRAM chips in world markets, their prices began to rocket upwards.
Memory costs went up by more than 5 times and while this had not
affected the larger systems much it was murderous on smaller systems.
Thus while more micros were installed in 1988 than in any other year
the number was well short of the 55000 to 60000 expected. However,
DRAM shortage helped the PC suppliers in a different way. From late
1986 and continuing through 1987 the sharp competition in the market
had pushed PC prices down to what was clearly uneconomical levels.
System Dynamics ’90
The DRAM shortage provided an ideal excuse to hike up prices once
again. While the PCs were becoming cheaper internationally, PC prices
in India were ironically charting an uptrend. Prices of Indian IBM-
compatible PCs had been rising without any attractive configurations
or even price incentive rebates. For example, one company's PC costs
(pre-budget) were approximately Rs.34000/- to Rs.37000/-; PC/XT
Rs.45000/- to Rs.48000/-; PC/AT Rs.58000/- to Rs.62000/-. Another
company's PCs were priced at Rs.27000/- to Rs.28000/-; PC/XT
Rs.36000/- to Rs.38000/-; PC/AT Rs.54000/- to Rs.58000/-.
The 1987-88 import-export policy provided a number of sops
for the hardware industry. However, the budget dealt a blow to the
software industry by a new excise duty. While the imported computers
attracted only 90 per cent duties, the duty level on imports of inputs
to Indian-made systems went upto 98 per cent.
The year 1988-89 saw a transition in product cycles, shifting
standards, takeovers and strategic realignment of products and markets.
The PC acquired multi-functionality during the year; PC-FAX and
PC-Telex becoming familiar. More computers were installed: Over 45000
micros and nearly 1500 minis were sold, The Indian computer industry
recorded a 45 per cent growth to touch the Rs.500/- crore mark. Yet
despite the increase in revenues, profits came down by the year end
‘and scores of companies were left fighting for survival. One company
slashed prices on their PC/AT to Rs.53000/- from Rs.66000/- inclusive
of a 132-column printer, while another company gave it for Rs.45000/-.
The market was dominated by just a few companies and quite
a few new products. In a market dominated by less than 10 major
companies, the remaining 50 odd PC vendors have had to fight for 40
per cent of the share. In other words, of the nearly 45000 PCs, XTs
and ATs sold last year, more than 30000 were accounted for by the
big ten companies alone. The fierce competition in the market had led
to polarisation of the market with the top four vendors between them
accounting for 54 per cent share of the total micro market.
Just before the 1989 budget there was a price rise of around
10 per cent. In the budget, the government increased the excise
component by 5 per cent which inflated the prices further. The policy
change envisaged a shift in the import of various peripherals, from
OGL to restricted list. The government also withdrew the 15 per cent
cash compensatory support for exports. This policy split the industry
into two distinct segments - the assemblers and the manufacturers, who
are respectively for and against the OGL imports. With regard to the
shifting to the restricted list, the PC cartel will have a firmer hold
over the domestic market.
The "“peoples' computer" was launched by a public sector
company during the last week of November 1989. The cost advertised
was Rs.10,500/- inclusive of 15.75 per cent excise duty. The price
quoted was for a PC equipped with a single floppy drive. The real
cost stood at Rs.15,493/- (inclusive of MS.DOS 4.01 system with manual
and three application packages, sales-tax and octroi). With a printer
added, the cost comes to about Rs.22,000/-.
As the indigenous computer market is a seller's market, the
need for efficiency, cost-effectiveness, increased productivity could be
System Dynamics '90
dispensed with and most computer manufacturers have got away with
it. The Indian market has a visible lack of demand for PCs in the
home and education sectors. In the home sector, the pricing levels are
a deterring factor. So also is the lack of sufficient home entertainment
and game programs. Indian households are likely to remain outside the
scope of most of the PC marketing effort.
Investments in terms of marketing and distribution have pushed
up costs to a level where the capital requirements are considerable.
This has wiped out a number of vendors with a lower resource base.
Selling PCs to Indian customers can be exceedingly complex.
The pattern of buying is significantly different across segments. The
customer support function has its own woes - changing products, mobile
Manpower, hostile environment (dust, frequent. interruption in power
supply) and not-so-confident after-sales service. On the manufacturing
side, obsolescence is the main problem. It: is also not possible to fix
any stable kind of a price/volume relationship, due to a continuously
varying price (governmental regulations). Then, of course, the ubiqutous
R & D problems.
According to the eighth plan approach paper, the computer
industry is supposed to grow at 6 per cent (5.1 per cent during the
seventh plan) and the installed base of micro computers alone would
need to touch one million units (against 1,19,000 units as at the end
of the seventh plan). Software exports would need to hit Rs.150 crores
today. The Indian computer market will grow from Rs.700 crores in
1989 to Rs.1000 crores in 1990. The PC market would witness rapid
price erosion and industry consolidation over the next two years.
In order to sell their products in an extremely competitive
environment, the computer companies have already started offering liberal
credit terms and other attractive packages. As the production capacity
in the industry is now far in excess of demand, the industry will
undergo a sweeping transformation. There will be a definite change of
direction. Only a few companies will remain in the field of manufacturing
computers and others will have to go in for value additions and specific
applications.
The PC market is being excessively price driven that may
result in PCs becoming commodity items. The first time user still needs
hand holding. However, this is the segment that needs to be tapped
if targets of 1,00,000 machines or more a year has to be met as pro-
jected for the peoples' computer.
Tables 1 and 2 sum up the Indian micro market.
MODEL DESCRIPTION
A generic system dynamics model is attempted to reflect the
dynamics of the Indian micro market. Details and data have entirely
been culled out from various Indian computer magazines and business
newspapers. whereverneeded, the data have been assumed in consistence
with the prevailing situation in the Indian market. The major ten
companies have been clubbed together to form the industry. The skeleton
of the model is shown in Figure (1). Several submodels have also been
defined. Sales sector submodel is shown in Figure (2). The model is
TABLE 1
INDIAN MICRO MARKET
1986-1987 1987-1988 1988-1989
CO ‘MICROS SALES UNIT MICROS SALES MARKET UNIT GROWTH MICROS SALES MARKET UNIT HROWTH
SOLD VALUE PRICE SOLD VALUE SHARE PRICE SOLD VALUE SHARE PRICE
(Nos) (Cr. of (Rs.) (Nos) (Cr. of (%) (Rs.) (%) (Nos) (Cr. of (%) Rs.) (%)
Rs.) Rs.) Rs.)
A 3920 12.92 33000 5900 20.58 5.60 34800 120 8284 44.10 9.90 49.36
B 4112 26.78 65000 4885 35.54 15,40 72700 32 7899 61.70 19.20 100.90
Cc 3344 17.26 52000 5210 26.59 11.60 57000 46 6834 47,80 15.50 72.83
D . 1184 4.40 38000 3053 12.04 5.20 39500 33 3750 15.00 4.70 & 17.55
E 950 8.07 85000 1800 NA 4.70 60000 3.5 2700 14.20 4.40 8 28.86
F 815 5.56 68000 NA NA NA NA 15 973 6.81 NA : 28.46
G NA NA NA 2100 6.80 2.90 32400 NA 4244 33.19 9.32 3 515.38
H NA NA NA NA NA NA NA NA 2754 12.02 NA 3 173.74
I NA NA NA NA NA NA NA NA 1533 7.93 2.20 ® 231.20
J NA NA NA NA NA NA NA NA 905 6.31 1.70 73.33
INDUSTRY 40000 200 49.50 50000 52000 340.13 69.0
NOTE: Company names are disguised. Data not available for companies established after
facturing micros.
1987 and/or not manu-
oor
06, soyureudg wa3shg
TABLE 2
INDIAN MICRO MARKET (CONTD.)
SALES TURNOVER(CRORES OF RUPEES)
co INSTALLED UNITS MICRO SALE 1985-86 1986-87 1987-88 1988-89 YEAR PRODUCTS
BASE SHIPPED PERCENT OF OF OPERA-
TOTAL REV. TION
(Units) (Units) (1988-89)
A 12569 690 81.92 1.85 16.40 36.04 53.78 1986 MICROS
B 17000 658 50.98 25.89 45.54 60.24 121.02 1978 MICROS ,COMMON PRODUCTS
Cc 15900 570 50.87 19.58 37.31 54.36 93.95 1981 MICROS,MINIS,S/W,
COMMON PRODUCTS
dD: 5000 250 54.74 12,45 17.56 23.31 27.40 1971 MICROS,CAD/CAM,S/W,
COMMON PRODUCTS
E NA 225 50.00 15.73 22.01 23.81 28.35 1972 MICROS,H/W,S/W,
PROCESS CONTROL
F NA 81 33.15 8.19 13,90 15,99 20.54 1980 MICROS,H/W,S/W,
- LANS
G 6400 354 82.97 NA NA 6.50 40.00 1987 MICROS ,COMMON
PRODUCTS
H 9058 307 38.96 NA 4,23 11.27 30.85 1979 MICROS, MINIS,
COMMON PRODUCTS
I 1564 102 95.77 NA NA NA NA 1987 MICROS,MODEMS,
TC CONTROLLERS
J NA 75 89.88 2.24 3.37 4.05 7.02 1980 MICROS,LANS,
COMMON PRODUCTS
INDUSTRY 234° (347 510 1,095
06, Soyureufq ura3shg.
Top
462 System Dynamics '90
JPER GROWTH RATE | — —_
~ POLICY
COMPETITION NS i
ee ' [oatsnane
cRowtn |] —» [aaaTsize] [wes] PabTS Hane eee
1 t
|
: iit i
¢ PROMEX SALREV | ———-» |pRODEV] —— [RATINP]
| }
°
FL |
|
[iwossc] -—-—. [eonsat] = [aersaq] ——~ [rerancea] ——
f I
ae ress] —
= a eZ
iad pmo
SATISFIED CONSUMERS,
“FIGS -SKELETON OF THE MODEL
AVERAGE SALE
CO'S REL PROM EXPENSES, CUMLATIVE ee
PERCENT OF SA
— a fo PERCENT OF SR
oS
a h
IMPLEMENTATION TIME — Sf KE TING és
Pi MAR}
———
PRODUCT MODIFICATION
oN,
PROQUC T IMPROVEMENT
Topps a
Vz. INFORMATION. REPLACEMENT
\L e ceva + PRODUCT CAPABILITY
EASE OF OPERATION
IMPLEMENTATION TIME
LATENT DEMAND,
PURCHASER ee Neue IN USE
OP venedithe,
neni PERFORMANCE
:
USER SKILL
‘ee FRACRTION
+
+
a nel OVERALL Sete val
RELATIVE PERFORMANCE
Garros) Gew eno0ue) “Nsranoano oF POTN
inwovar ORS =
PRODUCTINNOWAT!
IMITATORS aS NG pe) mee on
EFFECT
ABanoon a = ease [CONSUMER DELIGHT]
= ACCEPTOR FRACTIONS®
_ CO'S PROMQTIONAL PERFORMANCE
FIG.2-SALES SECTOR
System Dynamics '90
463
based on the philosophy of Morecroft's (1986) and Homer's (1987)
models, As the proposed model is still under progress and review,
only the salient features are described..
The model starts with an awareness diffusion module. 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 - FORGET.JK - NEGINT.JK - AADOPT.JK)
PBPOOL = Potential buyer pool (Buyers)
NRADVT = Number reached through advertising (Buyers/year)
NRSFCT = Number reached through sales force contact oe |
POSINT = Positive interaction between adopters
and potential buyers Coal
FORGET = Number fogetting Gad
NEGINT = Negative interaction between unhappy
adopters and potential buyers (G)
AADOPT = Actual adopters who have purchased
the product Gx}
The new product diffusion spreads through a combination 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 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 and the number of adopters depend on
the multiplier from cost which is modelled as a table function of average
cost of the product versus fraction of adopters.
A MLCOST.K = TABHL (MLCOS, AVCOST, 50E3, 75E3, 5E3)
Promotional expenditure(PROMEX) is modelled as a per centage
of total industry sales(INDSAL) which is equated to sales revenue
(SALREV). The marketing effort represents real expenditure on promo-
tional activities. The indicated marketing effort(IMKTEF) is a fraction
(indicated) of sales revenue to marketing effort multiplied by average
sales revenue(AVGREV). The indicated fraction is got by multiplying
the fraction normal(FSRMEN) by (POSINT/AADOPT). The equations are
R INDSAL.KL = (ECCOND.K)(PROMEX.K)(TEST1.K) (Rupees/year)
R PROMEX.KL = (DELAY3)(PROMXP.KL,BGDEL) (Rupees/year )
A PROMXP.KL = (PERSAL)(INDSAL.KL) (Rupees/year )
A IMKTEF.K = (IFSRME.K)(AVGREV.K) (Rupees/year )
A AVGREV.K = SMOOTH(SALREV.K,REVAGT) (Rupees)
A IFSRME.K = (FRSMEN)(POSINT.KL/AADOPT.KL) (0-1)
A ECCOND.K = TABHL(ECCON,TIME.K,1985,1990,1) (Gross domestic
product)
System Dynamics '90
Where ECCOND is economic condition which is modelled as
the growth of GDP over the years and its effect is thought of as resul-
ting in increased disposable income (Reserve Bank of India Report) and
is reflected in the company's share of the total market.
A COSMKT.K = (EFECON.K)(TEST2.K) (Dimensionless)
A EFECON.K = TABHL(EFECO,ECCOND.KL,1985,1990,1) (Per
capita)
The New Product Section is divided into three modules - New
product development, functional capability of the product and the reputa-
tion of the company. The company's share of the market(COSMKT) is
modelled as a function of product innovation effect(PRINN) and promo-
tional effectiveness(PROEFF). The PRINN is a table function of rate
of introducing new products(RATINP) and the PROEFF is a table function
of relative advertising expenditure(RADEXP). The RATINP is a delayed
variable of rate of start of new products(RSNEWP) over a period DELINP.
A TEST1.K = (GAMMA)(PRINN.K) + (PHI)(PROEFF.K) (Dimensionless)
A PRINN.K = TABHL(PRIN,RATINP.KL,O.1,.1) (44)
A PROEFF.K = TABHL(PROEF,RADEXP.KL,O,1,.1) (ea)
R RATINP.KL = DELAY3(RSNEWP.JK,DELINP)*PPC (Products/year)
The New product development section takes care of the product
developments, which are the outgrowth of projects costing thousands
of rupees in R and D activity and require years to complete.
L PRODEV.K = PRODEV.J+(DT)(PRODSR.JK-PRODCR.JK)
R PRODSR.KL = MAX(O,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 = PFCUR.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
rate Cen)
IPRODE = Indicated product development (Projects )
IFSRPD = Indicated fraction of sales i
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 (,.)
System Dynamics '90
The functional capability is modelled as
A FUNCAP.K = (PROCAP.K)(EEXPFC.K)
Where the product capability(PROCAP) is a level variable and the effect
of experience on functional capability(EEXPFC) depends on the relative
skill of the averageuser(REAUSR).
A EEXPFC.K = TABHL(EEXPF,REAUSR.K,0,.1)
A REAUSR.K XUSER.K/XSUSER.K (Dimensionless )
A XUSER.K = SUMEXP.K/AADOPT.K
where SUMEXP is the Co-flow of all experience
L SUMEXP.K = SUMEXP.J+(DT)(SXINCR.JK-SXDROP.JK)
R SXDROP.KL = (XSURSER*DEP)+(XUSER.K*REJECT.JK)
SXINCR Co-flow of experience increase rate
SXDROP. Co-flow of experience decrease rate
XSUSER = Experience of skilled user
The set of equations pertaining to the reputation module are
R ESTREP.KL
L AVPAGE.K
ESTREP
AVPAGE
TABHL(ESTRE ,AVPAGE,0,1,1)
AVPAGE..J+(DT)(1-(RATINP.JK*AVPAGE.J)/N)
Established reputation (Dimensionless )
Average Product age (Years)
oud
The model will have two distinct modules based on use innova-~
tiveness and product integration. The former may be split into creativity
and multi-use segments which measure respectively an individual's pre-
ference for variety (novelty) seeking within product usage and the ways
in which a currently owned product is used. The product integration
measures the utilization of the product, post adoption satisfaction and
expected future use based on experience of related products (Dickerson
and Gentry, 1983).
The post-purchase behaviour clarifies the rejectors(REJECT)
by differentiating between satisifed(PSBUY) and _ unsatisfied(PUSBUY)
buyers.
R REJECT.KL = (NORMAL)(USBUY.KL)
R USBUY.KL = (1-Q)*ADDOPT.KL
L PSBUY.K = PSBUY.J+(DT)(AADOPT.JK-USBUY.JK)
L PUSBUY.K = PUSBUY.J+(DT)(USBUY.JK)
System Dynamics '90
The sales lost to new competition and the growth of competition
are modelled in the competition sector.
R SLOSNC.KL = SALES.K/NORTIM*FMKTNC.K
A FMKTNC.K = TABHL(FMKTN,MKSTNC.K,O,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 (10)
NCOMPC = New competition capacity (Projects/year)
INCOMP = Increase in new (+)
competition
FNCOMP. = Failure of new competition (a5)
MNCOMP = Maturing of competition | Ga)
CONCLUSION
The suggested policy alternatives could be rationalization of
price, governmental policy, rationalization of excise duties, effect of
alternative configurations and software with a matching service. The
model, in its entirety, as contemplated will have several more sections
including a corporate growth sector. Partial model testing has been
done to debug the sales sector subsystem prior to the whole model
simulation. This was done to expose the intended rationality of the
decisions. The subsystem apparently responds. The nature of the
behavioural variables make it difficult to interpret the results. Further
work is on.
System Dynamics '90 467
References
Anderson, Robert L and Ortinau, David J. 1988.
Exploring Consumer's Podst-adoption Attitudes and Use Bheaviours
in Monitoring the Diffusion of a Technology-based Discontinuous
Innovation. Journal of Business Research 17:283-298.
Bass, F.M, 1969.
A new Product Growth Model for Consumer Durables. Managemert
Science 15(5):215-227.
Dickerson, Mary Dee and Gentry, James W. 1983.
Characteristics of Adopters and Non-adopters of Home Computers.
Journal of Consumer Research 10:225-234.
Dodson, Joe A and Muller Eitan. 1978.
Models of New Product Diffusion through Advertising and Woard-
of-mouth. Management Science 24(15):1569-1578.
Homer, Jack B. 1987. .
A Diffusicn Model with Application to Evolving Medicaf Techno-
fogies. Technological Forecasting and Social Change 31(3):197-218.
Jambekar, Anil B and Brokaw, Alan J. 1989.
A Tourism System Dynamics Modeg. System Dynamics: An Inter-
national Journal of Policy Modelling 2(1):1-11.
Mahajan, V and Peterson, R.A. 1979.
First Purchase Diffusion Modeks of New Product Acceptance.
Technological Forecasting and Social Change 15:127-146.
Morecroft, John D.W. 1986.
The Dynamics of a Fledgling High-technology Growth Market.
System Dynamics Review 2(1):36-61.
Reserve Bank of India Report on Currency and Finance. 1986-87. Vol.II-
Statistical Statements.
Rogers, Everett M. 1983.
Diffusion of Innovations. The Free Press, NY.
Shariff, M.N. and Ramanathan, K. 1982.
Polynomial Innovation Diffusion Modets. Technological Forecasting
and Social Change 21:301-323.
Shlomo Kalish. 1985.
A New Product Adoption Modet with Price, Advertising and Un-
certainty. Management Science 31(12):1569-1585.
CC BY-NC-SA 4.0