Supporting Material is available for this work. For more information, follow the link from
the Table of Contents to "Accessing Supporting Material".
Table of Contents
DYNAMIC DIFFUSION MODEL FOR MANAGING
CUSTOMER’S EXPECTATION AND SATISFACTION
Seung-jun Yeon', Sang-hyun Park*, Sang-wook Kim*, Won-gyu Ha*
124 Flectronics and Telecommunications Research Institute (ETRI I)
#161 Gajong-dong, Yusung-gu, Daejon, 305-350, Republic of Korea
E-mail: sjyeon@etri.re.kr ', ubiquitous@etri.re.kr’, and wgha@etri.re.kr*
* Department of MIS, Chungbuk National University
#48 Gaeshin-dong, Heungdeok-gu, Cheongju, Chungbuk, 361-736, Republic of Korea
E-mail: sierra@chungbuk.ac.kr?
Abstract: Being successful can be just as dangerous to long-term health as being unsuccessful. Even
success can sow the seeds of failure by stressing and overburdening the current system. While
suppliers may be tempted to hype up their products to obtain additional sales in the short term, those
customers persuaded by ‘hype' are often disappointed with their experiences, which in turn bears a
negative impact in the long run. Starting from this point, this paper aims at answering to the generic
question on how suppliers make the suitable and well-timed decisions in diffusing new technology
effectively to adopters. To meet this research objective, the paper attempts first to investigate the
entire process of the adoption and diffusion of technology innovation, and then proposes an integrated
model by concatenating in structured manner the three prominent models for the management of
technology innovation such as diffusion model, adoption model, and customer satisfaction model. An
exploration of the dynamic mechanism underlying outward behaviors of the integrated model is
presented in the study by introducing the system dynamics simulation technique.
Keywords: diffusion, innovation, customer satisfaction, system dynamics
1. INTODUCTION
Diffusion is defined as a process that innovation is communicated through certain channels over time
among the members of a social system. Diffusion models attempt to analyze this process by which an
innovation is diffused throughout a determined social system (Rogers, 1993). Many researchers that
motivate the innovation and diffusion processes have used mathematical models in logistic functional
form to study dynamic diffusion processes (Blackman, 1974; Mahajan and Peterson, 1979; Mahajan
and Shoeman, 1977; Sharif and Ramanathan, 1981). Most of these models are deterministic, have a
binomial form (adopt or not), and result in a typical S-shaped diffusion curve (Fisher et al., 2000). In
the underlying diffusion process, the probability of a new user adopting technology depends on the
quality of experience enjoyed by the existing users.
Some problems may draw from the prior researches discussed above. First, there may be a dilemma
that suppliers have to consider whether they hype up customers' expectation to diffuse new technology
or not. Second, it needs to clearly define the behaviors created by the causal relationships and
feedbacks among variables in the system, prior to making decisions of ‘technology infusion and
diffusion.’ Third, the behavior of diffusion may be different according to which feedback is dominant.
Starting from these points, the paper aims at answering to the generic question on how suppliers make
the suitable and well-timed decisions in diffusing new technology effectively to adopters. In probing
the questions raised above, adopted is the systems thinking approach, an excellent tool for better
understanding this type of complex management problems. According to Sharif and Ramanathan
(1984), once a diffusion model takes on these factors, system dynamics should be used to ease the
mathematical and computational complexity. Within this system dynamics paradigm three major
attempts are made for the study: First, investigating the entire process of the adoption and diffusion of
technology innovation; Second, proposing an integrated model by concatenating in structured manner
the three prominent models for the management of technology innovation such as diffusion model,
adoption model, and customer satisfaction model; Third, exploring the dynamic mechanism
underlying outward behaviors of the integrated model proposed in the study which depicts the causal
relationships that influence technology adoption and diffusion behaviors. These attempts made for the
study and the results perhaps allow both researchers and practitioners to gain insight into the causal
factors influencing customers’ adoption decision making processes and thereby into the potential
diffusion patterns resulting from those adoption processes.
2. BASIC MODELS AND AN INTEGRATION MODEL
2.1. Diffusion Model
A diffusion model is based on the beliefs that good sales practice with hyped technology is expensive
but leads to a high proportion of satisfied users, which is positive for subsequent diffusion; and that
high choice probability makes new customers increase and in turn accumulate into total customers as
shown [Figure 1]. But there are some missing pieces that show the relationships between expectation
and word-of-mouth effect. It should be noted that the probability of a new user adopting technology
also depends on the quality of experience enjoyed by existing users. These relationships behind the
diffusion model appear in the customer satisfaction model discussed below.
2.2 Customer Satisfaction Model
Parasuraman began a series of systematic and multi-phased research program in the mid-1980s,
focusing on the concept and measurement of service quality. After the initial conception of their
service quality "gaps model" in 1985, they began the process of developing an instrument for
quantifying customers' assessment of service quality performance. SERVQUAL (Parasuraman et al.,
1991, 1994) has become a reasonably well-accepted model for measuring the extent to which a
company meets its customer’s expectations. As shown in [Figure 2], however, SERVQUAL lacks a
feedback structure which implies simply customer satisfaction depends upon the gap between
customers’ expectation and the service quality. It should be noted that as customers’ expectation and
the technology change, the method of assessing services quality and thus customer satisfaction also
change continuously. This means that static evaluation on customer satisfaction does not help to make
good decision any more. The missing implications behind the customer satisfaction model can perhaps
be explained by growth and under-investment archetype.
2.3. Basic System Dynamics Archetype: Growth and Under-investment
The "Growth and Under-investment" structure is a famous archetype to system dynamics researchers.
The story line of the archetype can be described as follows: A company experiences a growth in
demand that begins to outstrip the firm's capacity. When the capacity shortfall persists, the company's
performance (such as on-time delivery) suffers and demand decreases. The fall in demand is further
seen as a reason for not making future investments in capacity. This leads to a self-fulfilling cycle of
continued under-investment and falling demand. This scenario is an example of the "Growth and
Under-investment" archetype at work. At its core is a reinforcing loop that drives the growth of
performance indicator and a balancing force which opposes that growth (loops R1 and B1 in "Growth
and Under-investment Archetype") [Figure 3]. An additional loop (B2) links performance to capacity
investments, and shows how deteriorating performance can justify under-investing in capacity needed
to lift the limit to growth. Remainder of the paper is organized as follows. First, the feedback
structures are divided and analyzed in detail. Second, based on the causal-loop diagram (CLD), a
stock-flow diagram (SFD) is developed for the computer simulation runs.
Expectation
New
N
Customers Castors
Word-of Mouth
Effect
[Figure 1] Diffusion Model with unilateral causality
wi nui GY
, Cn ee
rae
is
Service
[Figure 3] Growth and Under-investment Model
° °
a
a cap Quality of
cassis 2 Se Tyee
Satisfaction ‘ °
avetect
Promotions is
Choice i
Probability New Cartes
Customers. ——-F
Expectation
N
ce)
Percieved
Quality
[Figure 2] Gap model with unilateral causality
Quatty ot $
ee Service —
@
s, _ Promotions <f.
Custer
Satisfaction SS Investment
Bopetaton iy
| @) Profi
: s
chalce
s tro
Wed of Mouth Customers
Tiect 4 oa 4
‘J New Customers
[Figure 4] Modified diffusion model with feedback loops
[Figure 5] Feedback loops between satisfaction and quality [Figure 6] Feedback loops with revenue pressure
do
°
setstacon oe predes
i % up 3
sy ~
P pecww op Bovamnt
wewaratn | i
et e >
Cactomors f ~ oe |
mere reas
st 8 eet
Cink %
pokiy .
Pte rz
Polen E coum
[Figure 7] Feedback loop with investment on quality
. 5
Customer “cop * ality of
satfacton i “Semiee
Ean \ ¢ . us
ud Sway : pr vestment
eae =
om aa '
Promotion: “8
2
—_
S Presure
ss
Choice %
ae
F
5 Sa 2
tle i
[Figure 8] Feedback loop of dynamic expectation
2
customer 2
2 settfaton a —_—
Customer ear BO 5, ‘Service
Salisfection gos Quality of f Ss
5, Saves ; : s
* s
Fs oO & * Expectation Investnent,
= ‘Wordof- Mouth
¥ © espectaton nvestment Eta 40 rsh
Word of Muth . een | 7
weet e Ade ’ Secnien
Customers
Secession ic
‘ Cnn, Pressure F
* , SS -_ : Le leas
vronany 2ST reity *
ey
© otis “ew eee § coer
[Figure 9] Feedback loops with price effect on expectation [Figure 10] Feedback loops with price effect on profit
2.4. Detailed Explanation of Causal-loop Diagrams
Again, the ultimate goal of this paper is to develop a scheme for finding tentative answers to the
generic question on how suppliers make a suitable and well-timed decision in diffusing new
technology effectively to adopters with holistic view. To begin with, let's take a look at the modified
feedback structure from the diffusion model to explain how to make new customers [Figure 4]. The
expectation makes the probability of a new user adopting technology increase. High choice probability
attracts new customers and in turn leads to total customers increase. Another important factor is 'word-
of-mouth' effects. The probability of a new user adopting technology also depends on the quality of
experience enjoyed by existing users. Therefore most of the market-oriented business firms in general
often spend a lot of money in service quality in order to gain word-of-mouth effects. During the initial
stage of the market, companies concentrate on the marketing through mass communications in order to
get the attention of customers. As time goes by, the relative importance on the image of companies
depends on the word-of-mouth effects. Moreover, as the market saturates, a company which has
passed the point of critical mass of the customers and has entered into self-reinforcing feedback loop
gains the word-of-mouth effects a lot greater than the others (Choi, 1996).
From the perspective of customer satisfaction model, the increase in the customers’ expectation
directly leads to the gap increase, which results from discrepancy between the expectation and the
perceived service quality. Moreover, the increase of customers which is influenced by new customers
makes service quality decrease naturally in under-investment archetype. Both high expectation and
low service quality make a big discrepancy and in turn lead to the decrease in customer satisfaction. It
means positive word-of-mouth effect does not work any more [Figure 5]. As dissatisfied customers
secede, total number of customers also declines. This decrease of customers put the financial pressure
on the company. Naturally they would be inclined to adopt different promotions to get new customers
through higher expectation as shown in Figure 6. It makes a vicious cycle to customers dissatisfied.
On the other hand, if the gap expands, to prevent the decrease in customer satisfaction, the company
get desperate in providing better service quality. Improved functional quality increases customer's
perception on quality with delay. As mentioned before, the increase in perceived service quality leads
to the gap reduction. If the gap decreases, customer satisfaction will increase. Figure 7 shows the
balancing loop that makes the gap decrease. Therefore the gap is narrower than before.
Another important loop, B3 is the relationship among gaps, customer satisfaction, and expectation.
Gap is reduced by improvement of perceived quality that makes customer satisfied. But the increase in
customer satisfaction makes the rise of expectation dramatically and in turn expands the gap between
perceived quality and the expectation. Satisfied customers expect more than before. This expectation
leads to the gap expansion repeatedly [Figure 8]. Moreover the expenditure spent to improve service
quality makes pressure on the price of service to increase. Price in turn gives two effects on both
expectation and choice probability in each. The higher price results in the higher expectation, which in
turn lowers choice probability on the opposite side. The quantitative effect depends on price elasticity
of demand. However, it draw that expectation also increases. It is a reason that makes the gap wider
than before [Figure 9].
At the same time, price also has at least two effects on profit. It determines how much profit is
generated per unit sold and it affects the number of units sold (Sterman, 2000). That is, higher prices
reduce sales. In this case, the price elasticity of demand determines which causal pathway dominates.
If demand is quite insensitive to price (the elasticity of demand is less than one), then price raises unit
profit more than the net effect of an increase in price is an increase in profit. Conversely, if customers
are quite price-sensitive (the elasticity of demand is greater than one), the lower path (B6) dominates
[Figure 10]. The increase in profit per unit is more than offset by the decline in the number of units
sold, so the net effect of a price rise is a drop in profit. Actually, separating the pathways specify
different delays in each. There may be to be a long delay between a change in price and a change in
sales, while there is little or no delay in the effect of price on profit according to the circumstance.
During the initial stage of the market, it is a very important problem in diffusion process while the
price of new products often falls significantly over time through learning curves, scale economics, and
other feedbacks.
3. SIMULATION RUNS
To understand the effects of the dynamics of expectation, the reference model was developed as
shown [Figure 11]. This model includes decision-making about how to manage expectation to diffuse
supplier's new technology successively. Decision-makers have a chance to change ‘fraction on
spending for increase of expectation’. At first, decision-makers determine the fraction on spending for
the increase of expectation. This fraction determines how much effort to put on both expectation and
quality, which has influence on the state of expectation and quality and in turn change the level of
customers’ satisfaction. Both expectation and customer satisfaction attract new customers, and thus
total adopters will increase. If that is the case, what about the results from the alternatives? In [Figure
13], there are four behaviors of computer simulation performed under different situations by changing
the values of ' Spending fraction on Expectation’. The initial simulation was run under the assumption
that 'Spending fraction on Expectation’ was 10%. It means 'Spending fraction on Quality' was 90% of
spending of total investment for increasing perceived quality. The level of customers increases slowly
and remains steady at the equilibrium as shown [Figure 13]. The other three results of simulation are
run comparatively under different situation. According to change of 'fraction on ES', they shows that
the number of customers grow sharply as 'fraction on ES' is getting bigger and collapse is getting
deeper as behavior of expectation change as shown [Figure 12].
traction tom expectation
tect ot popuarty
on purchases
hopping rate
topping traction
[Figure 11] Dynamic Diffusion Model
(Customers: | -2- 3-4.
Ouatty: 1-2
5m we 7 om = aio vt te oa
[Figure 14] Behaviors of Quality [Figure 15] Behaviors of Satisfaction
Spending fraction on Expectation: Line 1 (0.1), Line 2 (0.3), Line 3 (0.5), Line 4 (0.9)
4. IMPLICATIONS AND CONCULSIONS
Some implications may be drawn from this model. 1) It becomes clear that everything in the system of
technology diffusion is dynamic, complex, and interdependent. It reminds us that simplification,
structure, and linear thinking have their limits, and can generate as many problems as they solve. The
main point is that it is needed for decision-maker to be aware of all the system's relationships - both
within it and external to it. 2) Both expectation and customer satisfaction have to be measured for
successful growth. The more satisfied the customers are, the higher expectation they will have. That is
why customers are not satisfied as before though the service quality has improved continuously. 3)
Various behaviors of diffusion process are the results of complex feedback structure. The best
approach is to strike a balance, to consider both short-term and long-term options and to look for the
course of actions that encompass both. 4) The behavior of customers varies depending on which effect
is dominated between the expectation and the word-of-mouth effects. Generic S-shape appears when
the word-of-mouth effect dominates; while goal-seeking behavior does in case that the effect of
expectation dominates the system. This is why the various types of diffusion behaviors exist in reality.
5) The speed of diffusion is influenced by the combinatory effects of expectation and customer
satisfaction. The higher expectation, the faster speed of growth; and the higher customer satisfaction,
the higher level of adopters. The extreme expectation generated by 'hype-up' makes the behavior of
adopter overshoot and collapse.
In conclusion, a model from a single perspective is not enough to discover the source of a problem and
understand better the technology diffusion process. Integrating related models by some means with
wider perspective is prerequisite to find a more effective solution. An a possible alternative the paper
attempts first to investigate the entire process of the adoption and diffusion of technology innovation,
and then proposes an integrated model by concatenating in structured manner the three prominent
models for the management of technology innovation such as diffusion model, adoption model, and
customer satisfaction model. An exploration of the dynamic mechanism underlying outward behaviors
of the integrated model is presented in the study by introducing the system dynamics simulation
technique. The general scheme for dynamism of systems and the findings presented in the paper would
perhaps provide some ideas and directions for further study.
10.
11.
5. REFERENCES
Blackman, W. A., The market dynamics of technological substitutions, Technological Forecasting
and Social Change, 6, 1974, pp.41-63.
Choi, W. Y., Customer Retention Strategy in slow growth phase, LG Economic Research Institute,
1996.
Fisher, D. K., Norvell, J., Sonka, S., and Nelson, M. J., Understanding technology adoption
through system dynamics modeling: implications for agribusiness management, International
Food and Agribusiness Management Review, 3, 2000, pp.281-296.
Mahajan, V., and Peterson, R.A., First-purchase diffusion models of new-product acceptance,
Technological Forecasting and Social Change, 15, 1979, pp.127-146.
Mahajan, V., and Shoeman, R.A., Generalized model for the time pattern of the diffusion process,
IEEE Transactions on Engineering Management, 24 (1), 1977, pp.12-18.
Parasuraman, A., Zeithaml, V. A., and Berry, L. L., Refinement and reassessment of the
SERVQUAL scale, Journal of Retailing, 67 (Winter), 1991, pp.420-450.
Parasuraman, A., Zeithaml, V. A., and Berry, L. L., Reassessment of expectations as comparison
standard in measuring service quality: implications for future research, Journal of Marketing, 58
(January), 1994, pp.111-124.
Rogers, E. M., Diffusion of Innovations, 3rd., New York, The Free Press, 1993.
Sharif, M. N., and Ramanathan, K., Binomial innovation diffusion models with dynamic potential
adopter population, Technological Forecasting and Social Change, 20, 1981, pp.63-87.
Sharif, M. N., and Ramanathan, K., Temporal models of innovation diffusion, IEEE Transactions
on Engineering Management, 31 (2), 1984, pp.76-86.
Sterman, J. D., Business Dynamics: System Thinking and Modeling for a Complex World, Irwin
McGraw-Hill, 2000.
Back to the Top
CC BY-NC-SA 4.0