Raising Implementation Effectiveness of Innovations by Considering Structural
Aspects of Organizations
Philipp Wunderlich'”, Andreas GréBler!
Submission to
International System Dynamics Conference 2011, Washington
March 2011
1 Institute for Management Research, Radboud University Nijmegen, P.O. box 9108, 6500
HK Nijmegen, The Netherlands
* corresponding author: p.wunderlich@fm.ru.nl, +31 24 36 11470
Raising Implementation Effectiveness of Innovations by Considering Structural
Aspects of Organizations
Abstract
The purpose of this study is to analyze the impact of informal communication networks on the
implementation process of innovations within organizations. Therefore, a System-Dynamics
model is built to simulate and analyze implementation-specific dynamics that influence
implementation effectiveness. The findings of this study suggest that senior management of an
organization can use its limited resources more effectively by focusing on employee groups
that are connected to each other and by isolating excluded groups from other groups that are
not influenced by senior management. In addition, managers should only apply pressure on
groups until a specific tipping point is reached after which the innovation diffuses by itself
within the respective group. Major limitations of the study are that only one network structure
was examined and that all groups are considered to be homogeneous.
Keywords: rational and ambiguous innovations, innovation implementation, communication
networks, diffusion, network structure, implementation effectiveness, management support
1. Introduction: Communication networks in organizational innovation
implementation processes
Due to the intensification of competition as well as the rapid evolution of technology,
innovations are vital to most organizations (Choi and Chan, 2009, p. 245). In addition, a
growing number of customers are expecting organizations to act ecologically and socially
responsible. Those circumstances force enterprises to adopt and implement innovations even
beyond their core businesses. Nevertheless, the results of innovations such as improvements
in efficiency due to total quality management, statistical process control, and manufacturing
resource planning are in many cases not satisfying (Klein, Conn, and Sorra, 2001, p. 811).
Several studies have shown that an organization’s failure to benefit from an adopted
innovation can often be attributed to a deficient implementation process rather than to the
innovation itself (Klein and Sorra, 1996, p. 1055; Aiman-Smith and Green 2002, p. 421;
Gary, 2005, p. 644; Karimi, Somers, and Bhattacherjee, 2007, p. 123). The implementation
process, as the critical interface between the decision to adopt and the routine usage of an
innovation (Klein and Sorra, 1996, p. 1057), has received increasing attention by scholars.
The degree of implementation success is considered a better indicator for innovation quality
than the degree of adoption success due to the fact that not all adopted innovations get
ultimately implemented (Karimi et al., 2007, p. 103).
Despite the growing number of studies which identify multiple causes of unsuccessful
implementation processes, literature is lacking multidimensional models that explain the
difference between successful and unsuccessful implementation efforts. Such models should
take into account multiple and to some extent interrelated drivers of implementation success
(Dean Jr. and Bowen, 1994, p. 393; Klein and Sorra, 1996, p. 1056; Klein et al., 2001, p. 811;
Repenning, 2002, p. 110). In addition, Choi and Chan (2009, p. 245) point out that existing
implementation studies tend to focus either on employee-related aspects, mostly on an
individual level, or on organizational aspects such as management support, structure, and
resources of the implementing organization. By combining these two approaches, Choi and
Chan (2009, p. 251) show that management support significantly improves the
implementation effectiveness as well as the innovation effectiveness by strengthening the
collective innovation confidence and the collective innovation acceptance of employees.
The present study aims to contribute to existing implementation literature by examining
the combined and interrelated influence of two organizational aspects (communication
structure and management support) on implementation success, which is characterized by the
2
employee-related aspect innovation acceptance and usage. This is achieved by combining the
organizational aspects predominant in diffusion literature and the employee-related aspects
mostly discussed in implementation literature by means of an informal social communication
network. In contrast to Choi and Chan (2009), this study does not focus on the strength of
causal relationships between factors of influence and implementation success. Instead, the
dynamics within and between interacting employee and management groups, which are
partially caused by the communication structure and which affect implementation success
over time, are of particular interest. Building on the derived knowledge of the underlying
dynamics, the effectiveness of different management policies is analyzed by means of
computer-aided simulation.
The structure of the paper is as follows. In the second section, we review the literature on
innovation implementation. We concentrate on the process of innovation implementation
within organizations and on the effects of communication networks on the success of
innovations. The third section discusses a system dynamics model that we use for subsequent
dynamic analyses of communication interactions and their consequences on the
implementation of innovations. The results of these analyses are described in the fourth
section, in which we investigate the influence of management pressure on innovation success
with and without migration between groups in an organization. The paper closes with a
discussion of implications for research and practice.
2. Literature Review
2.1. The Process of Organizational Innovation Implementation
Joseph A. Schumpeter (1996, p. 81-86) describes innovation as a process of creative
destruction which is continuously revolutionizing macro level markets and structures. The
widespread sub-categorization of the innovation process into the consecutive phases of
invention, innovation, as well as diffusion and imitation can also be attributed to Schumpeter
(1939, p. 84-102; Milling and Maier, 1996, p. 17). The invention phase is characterized by the
discovery of a previously unknown solution to a problem. In form of an innovation, the
invention is economically used for the first time during the innovation phase. In the
subsequent diffusion and imitation phase, the innovation spreads through the market, thereby
increasingly realizing the potential technological progress (Milling and Maier, 1996,
p. 17-18).
On a micro level, innovations diffuse between actors of a social system or an organization
through an existing or emerging set of relationships (Allen, 1977, p. 234-265; Roger, 2003,
p. 5). Everett Rogers (2003, p. 5-6) defines diffusion in the standard work Diffusion of
Innovations as a process by which information is exchanged over certain communication
channels between members of a social system. He differentiates between the five stages
knowledge, persuasion, decision, implementation, and confirmation. The knowledge stage is
initiated by the first encounter with the innovation and ends after a general understanding of
the innovation has been acquired. In the following persuasion stage, an affirmative or
negative attitude towards the innovation emerges. Within the subsequent decision stage, the
innovation is at least partially tested before it is decided whether the innovation will be
adopted or disregarded. In case of a positive adoption decision, the innovation will be used for
the first time during the implementation stage. Within the final confirmation stage, the
adoption decision is continuously challenged and where appropriate revoked based on newly
acquired information about the innovation (Roger, 2003, p. 168-169).
Within an organizational context, the innovation process is subdivided into two main
processes: the initiation process and the implementation process (Zaltman, Duncan, and
Holbeck, 1973, S. 58; Roger, 2003, p. 420), which are similar to the stages mentioned in the
previous paragraph (see Figure 1). The initiation process comprises the collection of
information, the creation of concepts, the planning of the adoption process, and the final
decision to adopt or disregard the innovation (Roger, 2003, p. 420-430). It consists of the two
sub-processes agenda-setting and matching. The former starts with the occurrence of an
organizational problem, which could lead to distress. This discrepancy between the desired
and expected performance of an organization can initiate the innovation process. Thereupon
the problem is exactly defined. Within the subsequent process matching, an innovation is
assigned to the problem in order to solve it.
In contrast to the initiation process, the implementation process comprises all events,
activities, and decisions which ideally lead to a routine usage of the innovation. It consists of
the sub-processes Redefining/Restructuring, Clarifying, and Routinizing. Within the first sub-
process of the implementation process, the innovation is adjusted to organizational needs as
well as to the organizational structure. During the second sub-process, the innovation is
increasingly understood and used by the members of the respective organization. Finally, the
innovation loses its autonomous character and becomes fully integrated into the organization
in the course of the last sub-process (Roger, 2003, p. 435).
Initiation Implementation
ee Matching Saal g Clarifying Routinizing
Optional Innovation-Decision Authority Innovation-Decision
Collective Innovation-Decision
Figure 1: The innovation process on an organizational level
Within the initiation process, Rogers (2003, p. 403) also differentiates between three kinds
of adoption decisions on an organizational level (organizational adoption decision). In the
case of the optional innovation-decision, an individual independent from the members of the
respective social system decides over the adoption or disregard of the innovation. If a
collective innovation-decision is underlying, such a decision is based on the consensus of the
members of the social system. In the case of an authority innovation-decision, a minority of
the social system, which is characterized by high social esteem, expert knowledge or power,
decides in favor or against the innovation. This decision must then be accepted by all other
members of the organization.
Even though both, the initiation as well as the implementation process, have a substantial
influence on the successful utilization of an innovation, this paper focuses on the internal
implementation process of an organization as highlighted in Figure 1. It is assumed that this
process is initiated by an authority innovation-decision, which was made by senior
management of the organization.
2.2. Influence of communication networks on implementation success
Before analyzing the implementation process, it is necessary to select at least one significant
measure of implementation success in order to distinguish between successful and
unsuccessful implementation efforts. Karimi et al. (2007, p. 108) evaluate implementation
success by measuring the effectiveness, efficiency, and flexibility of business processes,
arguing that the first-order effects of an implemented innovation occur at the operational level
of an organization. Since this study is not actually measuring the implementation success
within organizations, it is directly evaluating the performance of the implementation process
5
by using implementation effectiveness as a measure of implementation success. With respect
to existing studies, this measure is consonantly implying that there is a strong correlation
between implementation effectiveness and implementation success, the later being, among
others, characterized by visible benefits from the innovation as well as by the routinization of
the innovation among employees (Choi and Chan, 2009, p. 249-251).
After selecting implementation effectiveness as a measure of implementation success, the
question of how implementation effectiveness itself is characterized needs to be answered.
Aiman-Smith and Green (2002, p. 422) evaluate organizational implementation effectiveness
by means of user speed to competence and user satisfaction. The sooner the innovation can be
productively used and the more satisfied its users are the higher implementation effectiveness
is. According to Klein and Sorra (1996), implementation effectiveness describes “the quality
and consistency of the use of a specific innovation within an organization as a whole”
(p. 1059). In a study among US-hospitals, Douglas and Judge Jr. (2001) found a positive
correlation “between the degree of implementation of TQM practices and overall
organizational performance” (p. 165). Based on these three approaches, implementation
effectiveness in the present study is described by the intra-organizational diffusion speed, the
reached degree of overall adoption, and its sustainability. It is assumed that the innovation is
only used by the members of an organization if they are completely convinced that the
innovation is beneficial. Thereby, the proportion of adopters within an organization also
resembles the quality of use of an innovation, as mentioned by Klein and Sorra (1996,
p. 1059). Hence, implementation effectiveness can be interpreted as the extent of intra-
organizational acceptance and usage of an innovation over time.
The question remains how implementation effectiveness and thereby implementation
success can be positively influenced. In this context, many factors are discussed within the
literature. Certainly, innovation-related characteristics are among them. However, those
factors are already considered within the initiation phase. If the benefit of the respective
innovation is doubted within the initiation phase, the organizational adoption decision will
often be negative so that the innovation will not even reach the implementation phase. Apart
from this, several studies have shown that an organizations failure to benefit from an adopted
innovation can often be attributed to a deficient implementation process rather than to the
innovation itself (Klein and Sorra, 1996, p. 1055; Green 2002, p.421; Gary, 2005, S. 644;
Karimi, Somers, and Bhattacherjee, 2007, p. 123). Therefore, this study focuses on factors,
which are largely independent of innovation specific characteristics.
A large number of factors being discussed in implementation literature is concerned with
the acceptance and usage of the innovation by the members of an organization (Klein and
Sorra, 1996). Thereby, structural and institutional aspects are often not taken into account at
all or only in a very simplified manner. Damanpour (1996, p. 695), for example, examines by
means of a meta-analytic procedure the influence of organizational complexity on the
innovation process. However, only the extent of horizontal complexity, characterized by the
degree of functional departmentation and the extent of role specialization, is used as an
indicator for organizational complexity. Dynamics between the horizontal elements of an
organization are not considered. Similarly, Repenning (2002, p. 122) excludes interactions
between organizational groups in his analysis of implementation-specific dynamics.
While the connections and interactions between different groups of an organization have
been largely neglected in implementation literature, they are considered essential in diffusion
literature. Hence, Abrahamson and Rosenkopf (1997, p. 307) investigate the effects of
randomly generated network structures on the diffusion process of innovations within social
networks. Thereby, they focus on the bandwagon effect, which is based on the restrictive
assumption that members of an organization never change their opinion about an innovation
once they adopted it. The same assumption is made by Bohlmann, Calantone, and Zhao
(2010, p. 749) whose market-level study examines the diffusion process with respect to
different topologies of social networks. Gibbons (2004) analyzes the impact of innovation
networks, which change over time, distinguishing between clearly beneficial and ambiguous
innovations. However, the focus is on networks between organizations and not on social
networks within them. In contrast, Krackhardt (2001) examines the dynamics between
adopters and nonadopters of an innovation on an organizational level by not making the
restrictive assumption, that adopters never change their opinion about the innovation. Still,
Krackhardt (2001) focuses only on the diffusion process, neglecting characteristic factors of
the implementation process.
The present study aims to overcome the mentioned limitations of previous studies by
introducing a social communication model, which comprises employee-related mediators of
implementation literature as well as structural and institutional mediators of diffusion
literature in order to analyze the effects of those mediators on implementation effectiveness.
Thereby, this study coincides with Ford and Ford (1995, p. 561), who argue that
organizational change processes should always be placed within a context of communication
in order to understand them better. Kraatz (1998, p. 638), for example, states that
communication within social networks results in an adjustment of behavior among its
7
members. As an example he discusses that colleges organized in a network tend to implement
a particular bachelor degree if a network partner successfully implemented it beforehand
(Kraatz, 1998, p. 632). Kraatz (1998, p. 634) calls this effect social learning through
networks. Those indirect learning processes are not just taking place between organizations
but also within them (Wood and Bandura, 1989, p. 362). Consequently, communication
networks also influence the organizational opinion-forming process with regard to the
perceived advantageousness of an innovation (Abrahamson and Rosenkopf, 1997, p. 293).
Assuming that the members of an organization are free to come to their own decision whether
they adopt an innovation or continue to use the status quo instead (individual adoption
decision), the opinion-forming process also influences the proportion of adopters and thereby
the organizational implementation effectiveness. Building on a mathematical model of
Krackhardt (2001), the following chapter describes a dynamic model to analyze the opinion-
forming process within organizational communication networks.
3. A dynamic model of communication networks within the implementation
process
In a first step, the underlying model of this study will differentiate between adopters and
nonadopters of an innovation within several homogeneous and equally large groups
(Krackhardt, 2001, p. 250-251). Those groups can for example represent homogeneous
departments of an organization that are interconnected over an informal communication
network. This communication network thereby represents the communication structure of the
underlying theoretical framework. Within this network, an innovation diffuses on two levels.
On the first level, an opinion-forming process is taking place between the adopter camp and
the innovator camp within each group (Krackhardt, 2001, p. 251). The proportion of adopters,
represented by the variable C, will determine the degree of diffusion within one group. In
course of the opinion-forming process, a certain fraction of adopters and nonadopters gets
convinced by the opponent camp. In the following, this process will be called conversion. On
the second level, an opinion-forming process is taking place between groups. This process is
represented by the exchange of members of the same party between connected groups.
Assuming that five groups are interconnected in a row, a fraction of adopters of group 2,
for example, is migrating to the connected groups 1 and 3 in order to influence the opponent
camp in this group. In return, a certain fraction of adopters of these neighboring groups is
migrating to group 2. The same process is taking place between the nonadopter camps of
connected groups. In the following, this process will be called migration (Krackhardt, 2001,
p. 252-254). The migration difference between incoming and outgoing adopters of a group
over a certain period will be referred to as net migration rate. Figure 2 is illustrating the
interrelation between the two processes conversion and migration underlying a
communication network of five groups being interconnected in a row. In contrast to
Krackhardt (2001), who is assuming that conversion and migration are taking place in an
iterative sequence, in the system dynamics model of this study we make the more realistic
assumption that both processes are taking place simultaneously.
¥ migration al2 ¥ migration a23 ¥ migration a34 ¥ migration a45 ¥
Fraction Fraction Fraction Fraction Fraction
Adopters Adopters Adopters Adopters Adopters
Group | Group 2 Group 3 Group 4 Group 5
jconversion1 jconversion2 onversion3 lconversion4| conversions
Fraction Fraction Fraction Fraction Fraction
Nonadopters| Nonadopters| Nonadopters| Nonadopters| Nonadopters|
Group | Group 2 Group 3 Group 4 Group 5
‘ migration nl2 k migration n23 ry migration n34 k migration n45 A
Figure 2: Process of innovation diffusion within an organization using the example of five
groups organized in a serial structure
In the following, the mathematical structure of those two processes is outlined. According
to Krackhardt (2001, p. 250), the active search of organizational members for innovation-
related information and opinions drives the conversion process within groups. However,
taking into account the satisficing concept of March and Simon (1958, p. 140-141), those
members do not consider the opinions of all members of the group. Instead, they stop
searching after finding one other group member who is fortifying their own beliefs. Thereby,
the random search for members of the same camp is limited to a special part of the group.
Members of a group will only convert to the opponent camp if they do not find at least one
like-minded person in this part of the group (Krackhardt, 2001, p. 250-251). Based on Asch’s
(2003, p. 295-303) work, Krackhardt (2001, p. 250) assumes that adopters advocate the
innovation more ambitiously than nonadopters do with regard to the status quo. That means
that adopters scan a greater part of the group than nonadopters do (Asch, 2003; Krackhardt,
2001, p. 250). Hence, the search intensity of adopters (Search Intensity A) is greater than the
search intensity of nonadopters (Search Intensity N).
Equation (1) describes the proportion of adopters (C) in a group i that is converted by the
nonadopters of that group over a certain time period, because they were unable to find other
like-minded adopters in the part of the group they scanned:
dCy en (l -C ‘aa Intensity A
-or-G a)
dt Time To Convert
The term (1—C,)*”"""""""'“ represents the probability that an adopter only finds nonadopters
in the part of the group he or she scanned (Krackhardt, 2001, p. 253). The proportion of
adopters that does not find any like-minded group members and converts to the nonadopter
camp within a certain time period (Time To Convert) equals C -(1—C, )°"" """“, Following
the same logic, equation (2) calculates the positive change of the adopter proportion within a
group i due to the conversion of nonadopters:
dC. i= C. iy C Search Intensity N
iA = ( 2) i . (2)
dt Time To Convert
As assumed in section 1, the decision to adopt the innovation was made by senior
management. In line with Repenning (2002, p. 113) this study also assumes that senior
managers exert pressure on the employee groups in order to convert initial nonadopters to
adopters of the innovation. This third influence on the conversion process of a group i is
1
}4 (3)
The variable Cg represents the externally given goal of senior management concerning the
described in equation (3):
dy fe Dx:
G
n
dt
average proportion of adopters within the organization as whole. Hence, the first term on the
right hand side describes the discrepancy between this goal and the perceived average
proportion of adopters within the groups. This difference is divided by the variable 7, which
describes the time needed, until senior management develops and implements suitable actions,
until employees react to those actions, and until employees finally modify their behavior
(Repenning, 2002, p. 115). Therefore, the result of equation (3) represents the proportion of
nonadopters that convert to the adopter camp within a certain period due to the pressure
10
managers exerts on them. The total change in the proportion of adopters over time within a
group 7 is represented by equation (4):
aC aC ACiw + din : (4)
dt dt dt dt
The migration process, which is taking place between groups, depends on the structure of
the communication network. In the following, that structure will be described by an adjacency
matrix that maps the migration ties between groups (Krackhardt, 2001, p. 252). The so-called
migration fraction represents the proportion of adopters or nonadopters of a group i that
migrates to each congenial camp of all connected groups within one period (Krackhardt,
2001, p. 252). The multiplication of the migration fraction with the proportion of adopters or
nonadopters and the subsequent division by migration time result in a periodical migration
rate of the group i. The multiplication of that migration rate with the total number of
connected groups describes the proportion of adopters and nonadopters respectively that
leaves the group i. The proportion of adopters and nonadopters respectively that immigrates
from each connected group into group 7 is calculated analogously. The adding of those
migration rates equals the total proportion of adopters and nonadopters respectively that
immigrate into group i within one period.
4. Analysis: Management-caused dynamics of the diffusion process
4.1. | Dynamics of the conversion process within groups
The preceding chapter introduced Krackhardt’s (2001) social communication network and
extended it by taking into account senior management’s influence in an organization. In
contrast to Krackhardt (2001, p. 254), the main question of this study is not how a minority of
innovators can overcome a majority of nonadopters. Instead, the dynamics between and
within communicating groups being influenced by senior management are of main interest.
Thereby, management is initiating the implementation process by exerting pressure on
employee groups to adopt the respective innovation, assuming that the innovation-related
commitment of each group, and hence the proportion of adopters, is zero beforehand (C = 0).
The following analysis is based on a communication network consisting of five groups
organized in a row as depicted in Figure 2. Equation (5) shows an adjacency matrix, which
represents this structure:
11
(5)
5
|
ro)
ooror
0
1
0
1
0
r+ovrocoeo
oroocso
The matrix shows that the first group (first line of the matrix) and the fifth group (fifth line of
the matrix) only have one communication partner, whereas all other groups are influenced by
two communication partners. During the following simulation of the diffusion process, it is
assumed that the search intensity of adopters equals six while the search intensity of
nonadopters is four (Krackhardt, 2001, p. 255). Krackhardt (2001, p. 256) shows that only the
ratio between (not the nominal values of) the search intensities plays an important role. At the
beginning of the simulation the groups only consist of nonadopters. The migration time and
the time to convert are one week each. The migration fraction is assumed to be 12.5 percent.
In the following the influence of management pressure on the average commitment of groups
is analyzed. The time needed until senior management’s actions are implemented and show an
impact on innovation usage is assumed to be 12 weeks (7 = 12). For now, it is assumed that
managers apply the same pressure to all groups.
Figure 3 illustrates the interrelation between the duration of management pressure and the
fraction of adopters within groups assuming that management starts to take action in week
twelve, which is represented in the model by the management goal being raised from zero to
one (Cg = 1). The left part of Figure 3 shows the well-known logistic S-curve of the diffusion
process (Abrahamson and Rosenkopf, 1997, p. 295; Repenning, 2002, p. 116). If senior
management continues to exert pressure for 20 weeks until week 32, enough nonadopters will
be converted to adopters within all five groups so that they are able to convince the remaining
nonadopters in their groups from week 32 on, even without managements support. The right
part of Figure 3 illustrates what happens to the proportion of adopters in group 1 when senior
management stops exerting pressure earlier. Since all groups have the same proportion of
adopters before the start of the simulation and since managers are exerting the same pressure
on all five groups, those groups exchange exactly the same fraction of adopters and
nonadopters at each point of time during the simulation. Therefore, the net migration rate of
all groups is constantly zero. Hence, migration does not have an impact on implementation
effectiveness in this scenario. That means that the behavior of group | (right part of Figure 3)
is exactly equal to the behavior of all other groups.
12
Fraction Adopters Fraction Adopters
1 1
0.75 0.75
0.5 0.5
0.25 0.25
i) 0
0 8 16 24 32 40 48 56 64 72 80 88 96 104 0 8 16 24 32 40 48 56 64 72 80 88 96 104
Time (Week) Time (Week)
Fraction Adopters[grl] : mgmt_off32 ———________ Fraction Adopters[gr1] : mgmt_off32
Fraction Adopters[gr2] : mgmt_off32 —__________ Fraction Adopters[gr1] : mgmt_off30
Fraction Adopters[gr3] : mgmt_off32 ———______—_ Fraction Adopters[gr1] : mgmt_off28
Fraction Adopters[gr4] : mgmt_off32 ——_____—__ Fraction Adopters[gr1] : mgmt_off26
Fraction Adopters[gr5] : mgmt_of{32 —_________ Fraction Adopters[gr1] : mgmt_off24
Figure 3: Influence of the duration of management pressure on the fraction of adopters
If senior management only applies pressure until month 30, the innovation still reaches a
permanent degree of adoption of one (red graph in right part of Figure 3). Even though the
diffusion degree and its sustainability are the same as in the case when senior management
pushes the innovation two weeks longer (blue graph in Figure 3), the third dimension of
implementation effectiveness, namely the diffusion speed, is slower. If senior management
stops influencing the groups at week 28 (green graph in Figure 3), the diffusion only reaches a
temporary diffusion degree of 38 percent within group one. Afterwards, the remaining
nonadopters are still strong enough to convince more adopters of the superiority of the status
quo than adopters are able to convince nonadopters of the superiority of the innovation. This
initiates a self-reinforcing process that causes the nonadopter camp in group | to get stronger
and stronger until all adopters have been converted to nonadopters. An even shorter duration
of senior management’s influence further reduces implementation effectiveness with regard to
all three dimensions (grey and black graph in Figure 3).
These dynamics suggests that there is a system-immanent tipping point or threshold level
of the proportion of adopters (Morrison, 2008). After reaching this tipping point, the negative
influence of the communication process, which is driving out the innovation, turns into a
positive one, which causes a complete diffusion of the innovation. That means that managers
only need to apply pressure on groups until a certain fraction of adopters is converted. With
regard to the discussed scenario, this threshold is reached after 28.7 weeks when 41 percent of
the nonadopters have been convinced to use the innovation. The threshold is lower than 50%
due to the assumption that adopters are more committed to finding like-minded group
members than nonadopters are. That is, the search intensity of adopters is greater than the
search intensity of nonadopters.
13
4.2. _ Dynamics of the migration process between groups
After analyzing the dynamics of the conversion process within groups, the focus will now
be on the migration process, which is characterizing the dynamics between groups. Therefore,
it is examined how the implementation effectiveness is influenced if senior management
omits to exert pressure on one of the five groups. Thereby, the adopter-nonadopter ratio of the
omitted group will differ from that of the other groups. Due to the communication between
groups, which is represented by the migration of adopters and nonadopters, the proportion of
adopters will also rise in the excluded group. It is assumed that senior management influences
the addressed groups over the whole simulation period. All other parameter values stay the
same. Figure 4 is illustrating how the exclusion of group | (left part of Figure 4), of group 2
(middle part of Figure 4), and of group 3 (right part of Figure 4) is influencing the fraction of
adopters over time. Due to the symmetrical structure of the communication network, the
exclusion of group 4 has the same effect as the exclusion of group 2. The same applies to
group 5 and group 1.
Fraction Adopters Fraction Adopters Fraction Adopters
1 1 1
0.75 0.75 0.75
0.5 0.5 0.5
0.25 0.25 0.25
0 0 0
0 16 32 48 64 80 0 16 32 48 64 80 0 16 32 48 64 80
Time (Week) Time (Week) Time (Week)
Fraction Adopters[gr! ] : offgri
Fraction Adopters[gr2] : offgrl
Fraction Adopters[gr3] : offgrl
Fraction Adopters[gr4] : offgrl
Fraction Adopters[gr5] : offgrl
Fraction Adopters(grl} : offer2.
Fraction Adopters| gr2} : offgr2.
Fraction Adopters(gr3] : offgr2.
Fraction Adopters( gr4] : offgr2
Fraction Adopters[ gr5] : offgr2:
Fraction Adopters[gr!] : offr3:
Fraction Adopters(gr2] : offer3:
Fraction Adopters[gr3] : offgr3:
Fraction Adopters[gr4] : offgr3
Fraction Adopters{gr5] : offgr3:
Figure 4: Influence of selectively applied management pressure
Figure 4 illustrates that the exclusion of group 3 (right part of Figure 4) leads to the
smallest reduction of implementation effectiveness. From a management perspective, it would
hence be most effective to influence the peripherally situated groups 1, 2, 4, and 5. This result
is line with Krackhardt’s (2001, p. 260-261) principle of peripheral dominance, which is
stating that adopters are more likely to prevail against a majority of nonadopters the more
peripherally located the former are. However, the exclusion of group 2 or 4 (middle part of
Figure 4) has a bigger negative impact on implementation effectiveness than the exclusion of
the more peripherally located group 1 or 5 (left part of Figure 4). The dynamics that cause the
behavior depicted in Figure 4 will be examined in the following.
14
After introducing the innovation in week twelve, senior management is exerting pressure
on groups 2 to 5, which leads to a growing fraction of adopters in those groups (left part of
Figure 4). As a result, the adopter-nonadopter ratio of group 2 (red graph) differs from the
adopter-nonadopter ratio of group | (blue graph), which consists only of nonadopters due to
the lack of management pressure. The bigger this difference is, the lower (higher) is the net
migration rate of group 2 (1). This in turn hampers the increase of the fraction of adopters of
group 2, which is therefore smaller than the increase of group 3. Hence, the net migration rate
of group 3 is also different from zero. Due to the management-caused conversion of
nonadopters within group 2, the net migration rate of group 3 is less negative than the net
migration rate of group 2. The damping effect of the conversion process results in a less
strong impediment of the adopter-nonadopter ratio of group 3 (green graph in left part of
Figure 4). This effect is even stronger in group 4 and group 5 (grey and black graph in left
part of Figure 4). In general, the more influenced groups are between a group i and the
excluded group 1, the higher the implementation effectiveness of group i.
However, in case the excluded group is located in the centre of the row (right part of
Figure 4), its low adopter-nonadopter ratio is influencing two groups (red and grey graph),
which leads to a negative net migration rates in both groups. Compared to the exclusion of
group 1, the implementation effectiveness of the organization is therefore lower during the
first weeks in case group 3 is excluded. Due to the conversion process, the proportion of
adopters in group 2 and group 4 is still rising. As a result, a greater proportion of adopters
migrates from those two groups into group 3 leading to a growing fraction of adopters also in
the excluded group (green graph in the right part of Figure 4). The rising adopter-nonadopter
ratio of group 3 causes a lower negative impact of the net migration rate on the proportion of
adopters in group 2 and group 4. This in turn has an increasing positive impact on the net
migration rate of group 3, which results in an increasing fraction of adopters also of group 3.
The implementation effectiveness in case of excluding group 3 is higher than in case of
excluding group | because the greater negative influence of group 3 due to migration is more
than outweighed by the positive influence of the immigrating adopters of group 2 and
group 4. Thereby, the damper effect of the conversion process plays a vital role. Therefore, it
can be reasoned that the excluded group should be situated as centrally as possible so that the
damper effect of the conversion process is maximized and adopters from several groups
immigrate into the excluded group.
However, there is no linear increase in implementation effectiveness the closer the
excluded group moves to the center of the network. In case group 2 is excluded from
15
management pressure, the implementation effectiveness is even worse than in case senior
management disregards group 1 (Figure 4). The difference to the previous scenario is that the
excluded group 2 influences the peripheral group 1, which is not communicating with any
other group but group 2. In this case, the positive influence of the conversion process can only
dominate the negative influence of the migration process until week 25. After week 25, the
fraction of adopters of group | decreases (blue graph in middle part of Figure 4). This is the
case because the positive influence of the conversion process is decreasing before the tipping
point is reached (compare to left part of Figure 3). During that period, the negative impact of
the migration process is greater than the positive impact of the conversion process and
therefore leads to a decreasing proportion of adopters of group | (blue graph in middle part of
Figure 4). Since group 3 (green graph in middle part of Figure 4) benefits from a less negative
net migration rate because group 4 (grey graph in middle part of Figure 4) is also influencing
it, the migration process is not able to dominate the conversion process in group 3. Due to an
even smaller negative impact of group 2 on groups 4 and 5 (black graph in middle part of
Figure 4), those groups reach the tipping point most quickly. After doing so, the even greater
fraction of adopters immigrating from group 4 into group 3 is able to support the conversion
process of group 3 to an extent that also this group reaches the threshold level. This in turn
enables the adopter camp of group 2 to reach the tipping point and reverse the negative impact
on group 1.
5. Management implications and directions for future research
This study has analyzed implementation processes assuming that employees of an
organization decide whether they implement an innovation or stick with the status quo based
on the information and opinions they are exposed to as members of an informal social
communication network within an organization. In contrast to most studies, the present study
does not only focus on the diffusion of rational innovations, which show the typical
bandwagon-like behavior, but also on ambiguous innovations. Therefore, the underlying
communication model of Krackhardt (2001) was modified by making the more realistic
assumption that conversion and migration processes within communication networks happen
simultaneously instead of sequentially. While Krackhardt (2001) was focusing on the
outcome of diffusion processes characterized by the fraction of adopters, the purpose of this
study was to analyze the process itself also taking into account the diffusion speed and the
sustainability of the reached diffusion degree. Following Choi and Chan (2009) structural
aspects, predominant in diffusion literature, and employee-oriented aspects, mainly discussed
16
in implementation literature, have been combined in order to understand better why so many
implementation efforts fail. Hence, the structural aspects of a network have been combined
with employee-related aspects by considering the role of senior management as well as
different search intensities of adopters and nonadopters.
Assuming that senior management of an organization pushes the adoption of an
innovation, the simulation results of this study show that managers can use their limited
resources more effectively by considering the conversion dynamics within groups and the
migration dynamics between groups. First, senior management only needs to apply on a group
until the proportion of adopters of that group reaches a certain tipping point. When this
happens, the adopter camp of the group is strong enough to convert all remaining nonadopters
by itself. Therefore, management can save resources by stopping to push the innovation in a
group when this threshold level is reached. Second, senior management is able to save
additional resources while still ensuring high implementation effectiveness by not exerting
pressure on less vital groups based on their position in the network. Thereby, senior
management should ensure to influence peripheral groups because they are less likely to adopt
an innovation due to their relatively few migration ties to other groups. However, a central
position of an excluded group does not always a result in higher implementation effectiveness
than a less central position. Due to the greater number of migration ties of central groups,
these groups initially hamper the diffusion to a larger extent than peripheral groups if they get
excluded. Therefore, senior management should consider the damper effect of the conversion
process by applying pressure on connected groups and by not excluding several connected
groups.
Future research in this area can focus on the limitations of this study. Therefore, the
behavior of the presented communication network can be analyzed by considering other
network structures than the proposed one. In addition, further insights can be generated by
relaxing the assumption that all groups are homogeneous and that the ties between groups are
equally strong.
Literature
Abrahamson, E. and L. Rosenkopf (1997). “Social Network Effects on the Extent of
Innovation Diffusion: A Computer Simulation” Organization Science 8(3): 289-309.
17
Aiman-Smith, L. and S. G. Green (2002). Implementing New Manufacturing Technology:
“The Related Effects of Technology Characteristics and User Learning Activities”
Academy of Management Journal 45(2): 421-430.
Allen, T. J. (1977). Managing the Flow of Technology: Technology Transfer and the
Dissemination of Technological Information Within the R&D Organization, Cambridge,
Massachusetts.
Asch, S. E. (2003). “Effects of Group Pressure upon the Modification and Distortion of
Judgments” In: P. W. Lyman, H. L. Angle, and R. W. Allen (Eds.): Organizational
Influence Processes, New York, 295-303.
Bohlmann, J. D., R. J. Calantone, and M. Zhao (2010). “The Effects of Market Network
Heterogeneity on Innovation Diffusion: An Agent-Based Modeling Approach” Journal
of Product Innovation Management 27(5): 741-760.
Choi, J. N. and J. Y. Chan (2009). “Innovation Implementation in the Public Sector: An
Integration of Institutional Collective Dynamics” Journal of Applied Psychology
94(1): 245-253.
Damanpour, F. (1996). “Organizational Complexity and Innovation: Developing and Testing
Multiple Contingency Models” Management Science 42(5): 693-716.
Dean Jr., J. W. and D. E. Bowen (1994) “Management Theory and Total Quality: Improving
Research and Practice Through Theory Development” Academy of Management
Review 19(3): 392-418.
Douglas, T. J. and W. Q. Judge Jr. (2001). “Total Quality Management Implementation and
Competitive Advantage: The Role of Structural Control and Exploration” Academy of
Management Journal 44(1): 158-169.
Ford, J. and L. Ford (1995). “The Role of Conversations in Producing Intentional Change in
Organizations” Academy of Management Review 20(3): 541-570.
18
Gary, M. S. (2005). “Implementation Strategy and Performance Outcomes in Related
Diversification” Strategic Management Journal 26(7): 643-664.
Gibbons, D. E. (2004). “Network Structure and Innovation Ambiguity Effects on Diffusion in
Dynamic Organizational Fields” Academy of Management Journal 47(6): 938-951.
Karimi, J., T. M. Somers, and A. Bhattacherjee (2007). “The Impact of ERP Implementation
on Business Process Outcomes: A Factor-Based Study” Journal of Management
Information Systems 24(1): 101-134.
Klein, K. J., A. B. Conn, and J. S. Sorra (2001). “Implementing Computerized Technology:
An Organizational Analysis” Journal of Applied Psychology 86(5): 811-824.
Klein, K. J. and J. S. Sorra (1996). “The Challenge of Innovation Implementation” Academy
of Management Review 21(4): 1055-1080.
Kraatz, M. S. (1998). “Learning by Association?: Interorganizational Networks and
Adaptation to Environmental Change” Academy of Management Journal
41 (6): 621-643.
Krackhardt, D. (2001). “Viscosity Models and the Diffusion of Controversial Innovations” In
A. Lomi and E. R. Larsen (Eds.): Dynamics of Organizations: Computational Modeling
and Organization Theories, Menlo Park, California, 243-268.
March, J. G. and H. A. Simon (1958). Organizations, New York.
Milling, P. M. and F. H. Maier (1996). Invention, Innovation und Diffusion: Eine
Simulationsanalyse des Managements neuer Produkte, Berlin.
Morrison, B. J. (2008). “Putting the Learning Curve in Context” Journal of Business Research
61(11): 1182-1190.
Repenning, N. P. (2002). “A Simulation-Based Approach to Understanding the Dynamics of
Innovation Implementation” Organization Science 13(2):109-127.
Roger, E. M. (2003). Diffusion of Innovations, New York.
19
Schumpeter, J. A. (1939). Business Cycles: A Theoretical, Historical, and Statistical Analysis
of the Capitalist Process, New York.
Schumpeter, J. A. (1996). Capitalism, Socialism and Democracy, London.
Wood, R. and A. Bandura (1989). “Social Cognitive Theory of Organizational Management”
Academy of Management Review 14(3): 361-384.
Zaltman, G., R. Duncan, and J. Holbeck (1973). Innovations and Organizations, New York.
20
