Expert Systems with Applications 37 (2010) 8193–8200
Contents lists available at ScienceDirect
Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa
Incremental or radical? A study of organizational innovation: An artificial world approach Te-Tsai Lu a,*, Jong-Chen Chen b a b
Department of Business Administration, Kun Shan University, Taiwan, ROC Department of Information Management, National Yunlin University of Science and Technology, Taiwan, ROC
a r t i c l e
i n f o
Keywords: Organizational innovation Artificial world Consumer behavior Discrete-event simulation
a b s t r a c t The objective of this study was to construct a system that allowed organizations and consumers to interact with each other to investigate the effects of organizations adopting different innovative strategies on consumers. It was implemented by constructing an artificial world that abstracted some important features of organizations and consumers in the real world. The world consisted of four modules: organization, consumer, satisfaction (fitness) evaluation, and innovative change, which were linked into a model with discrete-event simulation techniques. Based on innovative strategies, organizations were divided into two categories: small (conservative) and large (aggressive) innovative changes. Two experiments were performed: ‘‘incremental innovation” and ‘‘radical innovation”. The experimental results showed that organizations adopting a comparatively conservative innovative strategy were likely to have better performance (in terms of consumer market share) than organizations adopting a comparatively aggressive innovative strategy. The result in the ‘‘radical innovation” experiment showed that there were limited changes of consumer market share for the organizations adopting small innovative change strategy. By contrast, innovation provided lesser-performing organizations a good chance of surpassing best-performing organizations if they adopted some innovative strategies, which could be roughly divided into four different categories. Ó 2010 Elsevier Ltd. All rights reserved.
1. Background and purpose of study Generally, we believe innovation is one of the main motives for economic growth and wealth building year round (Hargadon & Sutton, 2000), especially after the knowledge-based economy arrives, innovation becomes the key factor of success or failure for organization on competition (Jacob, Hellstrom, Adler, & Norrgren, 2000), and also the main source of competitive advantage (Sherwat & Fallah, 2005). And because of this, there have been increasing numbers of studies about innovation every year (Scott & Karl, 2004). Early scholars would face the common questions on the studies of characteristic of innovative organization and were not able to find the consistency from these characteristics (Downs & Mohr, 1976). On one hand, the complexity on the characteristic of the innovative organization is relatively higher than the others, and on the other hand, the variety of the attributes and operation variable while innovative are highly alterable. The main reason is because innovation phenomenon always is a ‘‘transient” process
* Corresponding author. Address: Department of Business Administration, Kun Shan University, No. 949, Da Wan Road, Yung-Kang City, Tainan Hsien 710, Taiwan, ROC. Tel.: +886 6 2050542x22; Mob.: +886 919786343; fax: +886 6 2050543. E-mail addresses:
[email protected],
[email protected] (T.-T. Lu). 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2010.05.067
(Carroll & Mosakowski, 1987). That is to say, the life cycle of an innovative enterprise is quite short, under a limited time, and if we want to do a long-term analysis to understand the whole innovation process that will encounter a natural difficulty (Khilstrom & Laffont, 1979). Thus, the innovation strategy that an organization adopts during the innovation process, may not the best choice under that situation. Wolfe (1994) indicated clearly, if we study innovation only from its result but not its process, it would cause bias. This is because organization innovation is a dynamic process, and all the attributes of organization, directly or indirectly, might be influenced by each other more or less. Slappendel (1996) pointed out further, the studies on organization innovation have to included individual and organization and their interaction process. But, the studies on these interaction processes are lacking presently, and it is making it difficult to perform these studies. Sendil and Daniel (2004) made appeals to strengthen and to discuss, design, manage and negotiate the interaction of a large-scared complicated system (for example, economic affairs, organizations, etc.), which should be the important core subject of future management and social science study fields. Besides organizational innovation, consumer behavior studies face a similar problem. Kotler (2005) said that consumers are independent individuals, and each consumer will be based on his/her independent will to make the consuming decision and behavior.
8194
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
Different consumers not only have different consuming behaviors, but also it will be more complicated if each consumer changes his/ her behavior when encountering other individuals or events. This will influence further not only the other consumers, also influence the original behavior of consumer him/herself and vice versa. Roughly, different decisions from the consumers and the result of their interaction might form a high dynamic phenomenon, and its final result normally cannot be controlled in advance. To sum up, we can say that both enterprise innovation and consumer behavior are very complex dynamic processed, and they are inter-worked. Traditional research approaches (including surveys and statistics which is cross-section study) are static analysis, and stress only on overall analysis. In other words, micro-behaviors of each consumer and organization usually will be ignored, but sometimes, one tiny change in an environment might lead to an extreme change under some special combination of environmental condition, this is the so called ‘‘butterfly effect”. Especially when the system condition reaches some critical point, one micro-vibration (interference) might touch off one macro change (Bak, Tang, & Wiesenfeld, 1988; Conrad, 1993; Kauffman & Levin, 1987). In the physical world, it is difficult, even impossible, to perform empirical studies of the dynamic interactions between organizations and their consumers. More important, empirical studies do not have good experimental ability and reversibility. Using simulation approach can fill the gap between real world and empirical study by setting different parameters to do different experiments. However, if we want to shrink this real world into a computer model, it is quite difficult or even impossible, in other words, some simplifications, assumptions and compromises are necessary and unavoidable (Fogel, Chellapilla, & Angeline, 1999). As what O’Callaghan and Conrad (1992) said, most of the systems are incomplete, which means that one system can match some certain assumptions, but not all. In this study, we are trying to establish a computer simulation system for the interaction between consumers and organizations. We create an artificial world, which is not to emphasize how to show every single element of the real world in details, but to catch some important characteristics (or parameters) which can be used to express or describe the objects in the real world, and set up some rules that can apply to these objects. And through these parameters and rules, we can establish the interaction relationship between objects and make the systems experimental by self-organizing to observe the process of dynamic behaviors. Also we can use the method similar to ‘‘Darwinian variation-selection” learning mechanisms where the system can have a stochastic approach to discover some unpredictable phenomenon. And eventually, we can analyze and trace some possible reasons and results. However, we emphasize that the purpose of this study is not only to practice in the computer system for some prior predicable results, but also to seek some possible phenomenon through the system of self-organizing learning. There are two advantages for the above artificial world approach: Firstly, it will not be limited to collecting the actual data which are difficult or even impossible to get, and also the study needs not to wait to proceed until after the controversial issues and also needs consistency. This offers some certain free and flexibility in the experiment. Secondly, it is a meso-system, thought ‘‘from simple to complex” approach which increases some system functions gradually, we can clearly analyze and know well by the experimental results, and to adjust the system when some new problem needs to be solved, and add some complicated and dynamic conditions to the system. Then, we can understand further the detailed dynamics inside the system. Technically and practicably, it has the relative advantage to adopt this approach with gradually increasing system function.
The researches of system dynamic using artificial world approach earlier, included ecosystem simulation (O’Callaghan & Conrad, 1992; Rizki & Conrad, 1985) and cerebrum data process (Chen & Conrad, 1997; Sipper, 1995). In recent years, this approach has been gradually applied to some social science studies, like Terano (2000), who used genetic algorithm and agent mechanism to establish an interaction system with an artificial world electrical community. Chen, Lin, and Kuo (2002) built an artificial world simulation system to research the fitness evaluation from the different leadership with subordinate adoptive different learning strategies, and the influence of short and long-term fitness when leadership changes. Sara, Roman, and Melanie (2005) studied the influence in the consumers market and the interaction for each other adopted with different market segmentation strategies. Lu, Chen, and Liao (2008) use a similar model to study organizations which adopt an aggressive or conservative strategy of learning in different competing environments. And from this article, we apply this approach to organization innovation and consumer behavior which is the very beginning of study in this field. As follows, we will describe the composition of the research model, system architecture and the related literatures, and then explain the experiment plan, experiment process and result analysis, and finally will be the conclusion for this study.
2. Research model The aim of this research model is to establish an environment (system) which can allow us to explore the interaction between organizational innovation and consumer behavior. In this interactive environment, when one or both have any change, it will influence directly the other’s fitness, and the change of the latter will influence the former again and then back and forth. The whole system will become an environment which elements connect and inter-influence each other. The main elements are composed of consumers and organizations. Consumers express their preference by choosing the organizations’ commodities, and organizations reply to consumers through the change for their commodities (we call it innovation change). There are roughly four modules in this model: ‘‘consumers”, ‘‘organizations”, ‘‘fitness evaluation” and ‘‘innovation change”, and through the ‘‘discrete-event simulation” techniques, we connect the interaction with all elements in this model, which means the interaction for each element of ‘‘consumers” and ‘‘organizations” which is completed by the ‘‘events” in discrete time. The application in the whole system for ‘‘who, which, when, where, what”, we can image that each consumer (who) will have consuming behavior (which) at some time (when) on some commodity (what) supplied by organization in respective area (where). From another view, if the commodity supplied by the organization is accepted by the consumer, this means this commodity is approved by the consumer. This model has four hypotheses: (1) Each organization only supplies one commodity at some certain time in each system clock with the same price (not to compete on the prices). (2) Each consumer only can choose one commodity at some certain time. (3) Each consumer can get all information of the commodity in order to evaluate his/her satisfaction with the commodity. (4) Each consumer can choose (buy) the most suitable commodity for him (her) without difficulty. Firstly, we will introduce the related research for the characteristics of ‘‘consumer” and ‘‘organization”, how to perform ‘‘consumer” and ‘‘organization”, and try to establish some relationship between them, and then explain how to execute consumer ‘‘fitness” and organization ‘‘innovation change”, lastly, we will describe how to make use of ‘‘discrete-event simulation” and correlate with all of them.
8195
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
2.1. Module of consumer and organization Module of consumer is composed of a group of consumers with different characteristics, and ‘‘organization” is composed of a group of organizations with different characteristics. For the study of consumers’ characteristics, Plummer (1974) proposed 36 factors that would influence consumer behavior, and divided them into four important constructs (activity, interest, opinion and population statistics variables). After that, Engel, Blackwell, and Miniard (1995) proposed 15 factors from the view of consumer’s decision process, and generalized it as three modules of environment influence, individual difference and psychological procedure. Recently, Kotler and Armstrong (2005) also pointed out 15 factors which would influence consumer behavior. From the study of organization characteristic to organization performance, some scholars believe organization performance and working environment quality are a positive correlation significantly (May, Lau, & Johnson, 1999), and Dyer and Reeves (1995) thought it is correlated with human resource management, while Chopra and Van Mieghem (2000) proposed enterprise electronics would influence its performance. The above scholars all study from a single level for the factors which would influence organization effects, the other group of scholars feel we should understand organization effect from all kinds of views, which means organization performance is a multidimensional concept, for example, Venkatraman and Ramanujam (1986) believe that organization performance needs to include finance, management and structure. Kaplan and Norton (1996) proposed balanced scorecard, and emphasized that we should study from the four perspectives of learning and growth, business process, customer and finance. From the above literatures we can see that, different scholars sum up the characteristics of consumers in that their consuming behavior will be not same. In the same way, in aspect of organization performance study, we still cannot find the characteristics agreed on by all scholars. Thus, we adopt a general concept for this study to describe the characteristics of consumer and organization management characteristic. And the meaning of ‘‘general” is to abstract the present study of consumer characteristic and organization characteristic, and conclude them into some factors. Yet, to avoid some possible controversial issues again on these factors, we do not emphasize the characteristic parameter to be the terms of the real world, in other words, we can call it a label for each parameter. In conclusion, we assume there are six characteristics for consumer and organization respectively. We now explain how to establish the relationship of consumer and organization. The relationship is hinted by Nicosia (1966) who proposed that in a consumer behavior model, he thought the interaction between consumer and organization is based on the message communication for each other, which means, consuming behavior comes from commodity characteristic and consumer attitude, and it is the most important procedure to decide the purchasing, and that the organization would present the ‘‘commodity” to the consumers through marketing, and consumers would express their willingness by buying the ‘‘commodity”. Under this situation of interaction, ‘‘commodity” is the media of communication for each other. In Nicosia’s model, it emphasizes that consumer behavior is some kind of interaction for organization and consumer characteristics, and ‘‘commodity” also can be considered as external manifestation for organization. Thus, we adopt one two-tier transmission mechanism, to transform organization internal traits (characteristic) to be external commodity traits, thus we get the interaction relationship between the commodity traits and consumer traits. The two-tiers transmission mechanism comes from some early ecological environment simulation systems (Rizki & Conrad, 1985). Recently it started to be applied in the socioeconomic field, like human resource management system (Chen
Fig. 1. The relationship of ‘‘genotypic traits” and ‘‘phenotypic traits” for organization.
et al., 2002), and organizational learning (Lu et al., 2008). And this two-tier transmission mechanism has the ability to improve evolution (Asselmeyer, Ebeling, & Rose’, 1996; Shipman, Shackleton, & Harvey, 2000). In Fig. 1 regarding the two-tiers structure for organization traits, the lower tier parameter represents the organization basic trait, and we refer to it as ‘‘genotypic traits”, and the upper tier parameter represents the organization external trait, which we call ‘‘phenotypic traits. In other words, ‘‘genotypic traits” is the abstract description for organization, and ‘‘phenotypic traits” is the externality of organization what consumers can feel, this means that consumers will evaluate this by ‘‘phenotypic traits” to see if it is what they expect to buy or not, and also it is the medium for organization to communicate and interact with consumers. Somehow, we can imagine ‘‘genotypic traits” as organizational constitution, like organization structure, technology ability, personnel training, leadership, strategy, etc., and we can see ‘‘phenotypic traits” as the commodities and service attribution supplied by the organizations, like function, quality, design, brand image and service, etc. And all these organization’s genotypic traits will influence the performance of their commodities (phenotypic traits) to the consumers. Therefore the two-tiers structure is fully connected with weighted relationship. We presume that each ‘‘genotypic trait” (gi) of organization has some certain influence on each ‘‘phenotypic trait” and labeled by weight (wij), Eq. (2) show the relations among them. It has two important hints: (1) Each ‘‘phenotypic trait” is determined by several ‘‘genotypic traits”, not by single ‘‘genotypic trait”, this makes every change of ‘‘genotypic trait” value will not be over-reacted on the corresponded ‘‘phenotypic trait”. (2) Each ‘‘genotypic trait” might influence several ‘‘phenotypic traits” at the same time, which means, when one ‘‘genotypic trait” value change, it might influence several ‘‘phenotypic traits” values at the same time.
pi ¼
n 1 X wij gj ki j¼1
where ki ¼
n X
wij
ð1Þ
j¼1
From above structure, for ‘‘genotypic trait”, ‘‘phenotypic traits” and weight variable, we adopt the value from 0.00000 to 1.00000 randomly in this study. The difference of these values is not represented as good or bad, it only represents the different traits among consumers and among organizations. The reason we adopt random is that there is not any acceptable value accepted by most scholars for the traits of organization and consumer, somehow it is unavoidable from some suppositions. Furthermore, the values produced randomly are uniform distribution, and in future studies, we can adopt the other distributions to do the experiment with different purposes.
8196
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
2.2. Fitness evaluation ‘‘Fitness evaluation” means each consumer will evaluate the fitness on the commodities supplied by the organization based on their own consuming traits, and decide if this will enable a ‘‘consuming change” event. The ‘‘consuming change” means that the consumers will change from one organization to another one when they buy a different commodity. The main evaluation index is from the concept of consumer fitness. This concept has been identified by most of scholars as very important influence for enterprise management, as in Berry (1995) and Reichheld (1996). The evaluation approach for consumer fitness in this study is partially the idea of ‘‘gap score” proposed by Oliver (1980), which means the satisfaction of consumers can get from the gap between ‘‘actual” and ‘‘expectation” quality of the commodities. When the two levels are matched, satisfaction of consumers reaches the highest, and by the gap becoming larger, the satisfaction of consumers will go down. For this point, we also take the concept proposed by Steenkamp and Baumgartner (1992), that during the purchasing procedure, a too high or too low stimulus will not arise any impetuous purchasing behaviors for the consumers, which means the ‘‘gap score” is a two-way function actually. They also describe the relationship between consumer’s satisfaction and marketing stimulus to be an inverted U curve. In other words, the excess marketing stimulus will not increase the satisfaction of consumers, and might actually reduce satisfaction. Similar literatures which supported this concept include Baumgartner and Steenkamp (1996) and Shiv and Fedorikhin (1999). The recent research Thompson, Hamilton, and Rust (2005) also clearly indicated that when a single commodity offers too many features, it might not be positive for consumers, and might have a negative influence, which we call it ‘‘feature fatigue”. To sum up above discussion, we presume in the relationship between consumer’s satisfaction and commodity traits, there will be a normal distribution. Like Fig. 2, we presume the relationship of one of consuming traits and its satisfaction (fitness) is a normal distribution of average value C1. And if the commodity supplied by the organization which ‘‘phenotypic traits” (p) is equal to C1, then the satisfaction will reach the highest value 1; when the commodity trait diverges C1 (higher or lower), and the satisfaction will be lower with the curve. In this study, we presume the variables of consumer normal distribution curves are all the same, only average value (consumer traits) is different. Thus, the consumers with two different trait average values can get different satisfaction on the same commodity. Like Fig. 3, if some ‘‘phenotypic trait” of the commodity supplied by organization is p, two consumers (c1 and c2) will get different s1 and s2 satisfaction separately. This design is
Fig. 2. Satisfaction of single consumer trait.
Fig. 3. Satisfaction of different consumers on one organization trait.
fitted with general practice concept, the same commodity will get different satisfaction with different traits of consumers. Above we get the evaluation of satisfaction of consumers on one trait of certain organization (or commodity). If we sum up the consumers’ satisfaction value of all organization traits (supposing there are six at present), it represents the total consumers’ satisfaction on the organization (or commodity). In each system clock, each consumer would have the evaluation of satisfaction, and choose the organization (or commodity) with highest satisfaction, and consumers will be bases on this for their consuming choice next time.
2.3. Innovation change In the theory of organizational innovation, Abernathy and Utterback (1978) divided the innovations into product innovation and process innovation, and also divided the innovations into improving the existed products and develop the new products, the prior is called ‘‘incremental innovation” and the latter is called ‘‘radical innovation”. The other scholars (Dewar & Dutton, 1986; Henderson & Clark, 1990) divided innovation into continuous innovation and discontinuous innovation. The former is similar to incremental innovation, and the latter is similar to radical innovation. Tushman and Anderson (1986) indicated that if the enterprise adopts incremental innovation and radical innovation step by step, it can destroy and rebuild core competence continuously and establish life cycle of technology innovation. In this study, we will explore how the enterprises that adopt incremental or radical innovations can influence their consumer possession rate. Practically, the enterprise that adopts incremental innovation strategy is to presume the worse enterprise will learn from better enterprises (copy their ‘‘genotypic traits”) continuously. The approach above is hinted from ‘‘market orientation” theory (Day, 1992; Narver & Slater, 1990), which means that the organization has to learn from their competitors (or clients). Since we presume the organizations with more consumers are better in their performance, thus, to learn from organizations with better performance is to learn from competitors, also we can say learning from their clients. And in the practice of radical innovation, we set the ‘‘genotypic trait” of enterprise with random change under some certain range. From the above approaches, the enterprise adopting incremental innovation strategy is to improve from existed base continuously and slightly, and adopting radical innovation strategy is to emphasize to jump the existing frame out, however, innovation with the limited resources (fund, resources and ability) will surely be changed in some certain scope (which means ‘‘genotypic trait” will have changed randomly in some scope). There are three steps for innovation change as follows:
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
1. Performance evaluation: We presume each consumer will choose the commodity supplied by one organization in each system clock to do the purchasing, from the other view, we can claim that this organization owns this consumer at that system clock, each organization performance is determined by the consumer number owned by each organization, the more consumers, the better the performance. And then, the top five organizations with good performance will become the learning targets for the other organizations with incremental innovation (change) strategy. 2. Incremental change: The ‘‘genotypic traits” and weighted values in all organizations learn from top five organizations, and the approach is to sum up the value of ‘‘genotypic traits” from the top five organizations and find the average value (gk). And then, we will do some certain change on this average value and copy to those learning organizations, Eq. (2) is the calculation of ‘‘genotypic trait” change, variable f represents change range whose value is from 0.00000 to 1.00000, the bigger the change range is, the more change ‘‘genotypic trait” is. The calculation for weighted value is similar with calculation of ‘‘genotypic trait” value.
g iðtþ1Þ ¼ g iðtÞ þ f fg kðtÞ g iðtÞ g
ð2Þ
gi: ‘‘genotypic trait” of organization i; gk: the average value of genotypic trait for top five organizations; f: change range; t: system clock. 3. Radical change: The organization adopting radical change will choose two of six ‘‘genotypic traits”, and change by random from the original value up and down 5%. 2.4. Discrete-event simulation The purpose of this study is to research the interaction between consumers and organizations through events simulation. For consumers, it includes two events, firstly is to evaluate on the commodities supplied by the organization (‘‘satisfaction evaluation” event), then to decide if to purchase the commodities or not, (‘‘consuming change” event). For the organization, it will evaluate the choice of all consumers (indirectly, it reacts the fitness to organization), and this is organization ‘‘performance evaluation” event, and based on these results, the organizations will learn from those which have better performance, that is adopting ‘‘incremental change” (event, or ‘‘radical change” event). System will automatically trigger ‘‘satisfaction evaluation” event of consumers in every system clock, and then, it might have a ‘‘consuming change” event. For organizations, system will execute ‘‘performance evaluation” event of organization, in the same way, system might trigger ‘‘incremental change” event or ‘‘radical change” event.
change by random. The other experiment is to presume the prior loser has the radical innovation.
3.1. Incremental change competition experiment We presume there are 12,000 consumers and 120 organizations, and divided these organizations into two groups which are ‘‘large” change and ‘‘small” change (the change range value is 0.5 and 0.3 respectively, please refer to Eq. (2)), and each group has 60 organizations. At the beginning of the experiment, each organization’s trait value and each consumer value were all initiated randomly (the value is from 0.00000 to 1.00000), that means, these consumers are equally distributed into all the organizations, and the consumer numbers owned by each organization is roughly the same. So, 12,000 consumers and 120 organizations mean each organization can own around 100 consumers. We do the experiment 10 times, and every time we will use different random numbers, which means that there will be different initiated trait values for organizations and consumers, and the average result we get is shown as Fig. 4. The organization adopting ‘‘small” incremental change attract around 7000 consumers, which means it owned more than half of the total 12,000 consumers, in other words, it does not mean the organizations adopting ‘‘small” incremental change strategies have absolute superiority compared with the ones adopt ‘‘large” incremental change strategies at the initial value setting. By the way, when the system clock is approximately 50, the organizations adopting ‘‘small” incremental change strategies suddenly rob a majority of consumers. After that, although organizations adopting ‘‘large” incremental change strategies gradually attract some consumers after system time 130. They are still the loser when the system reaches a total balance, it will not have any change until the 271 system clock. Why the organizations adopting ‘‘small” incremental change strategies can get most of the consumers compared with the ones adopting ‘‘large” incremental change strategies? This result is opposite to the expectation as we haven’t yet completed the experiment in this study. We expect the organizations adopting ‘‘large” incremental change strategies will tend to be more towards the superior organizations, and they will get most of the consumers rapidly. Yet, the truth is quite different, when organizations are alike more and more rapidly to the superior ones, all the organizations will become an ‘‘isomorphism” (DiMaggio & Powell, 1983), which means the commodities will all look the same, and in the same way, the commodity will lose their distinguishing features. Relatively, the group of organizations adopting ‘‘small” incremental change strategies will keep their diverseness because the
3. The procedure and result of experiment We perform two innovation change experiments: ‘‘incremental change” and ‘‘radical change”. ‘‘Incremental change” experiment is to divide all the organizations into two categories, one is to adopt ‘‘large (aggressive)” incremental change strategy, the other one is to adopt ‘‘small (conservative)” incremental change strategy, and make the two categories of organizations compete with each other until the system is stable and changes no more. ‘‘Radical change” experiment is to presume that under a stable situation of the above competing organizations, when one-side of them has radical change if this would change the balance? Regarding this question, we will take the stable result from prior experiment to do the two experiments, one is to presume the original winner has radical
8197
Fig. 4. The result of incremental change competitive experiment.
8198
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
change range is small and that means to keep the original commodity feature, under even distributed consumers, instead, they can get the majority of the consumers. The above result is similar to the real world, when the popular commodities are introduced in the market, some of the prior enterprises might get the majority of consumers, then, the other enterprises will follow up blindly to push a similar commodity into the market, on one hand, it makes ‘‘isomorphism” organization appear rapidly, while on the other hand, the commodity supply will exceed the demand and lower the profit of organizations. The genotypic traits of organizations adopting ‘‘small” incremental change tend to go to the best organizations more slowly, so their commodity features are more diverse and attract more diverse consumers. The above result also has the hint on some possibility that when consuming environment is diverse and multiplex, it may be a high risk action if the organization adopts the approach of imitating superior ones rapidly, because when the commodity is promoted in the market at the initial stage, consumers do not quite understand what they really need. With so many choices in the market, consumers are forced to make their decisions, at this stage, the organizations with better performance might only be a short-term phenomenon, after that stage, if there is one or more suitable commodities in the market, consumers will change their original choice. This concept can be identified again by next experiment (radical change).
3.2. Radical change experiment This experiment is based on the result of previous experiment, which is incremental change competition experiment at stable stage. We do the radical change of organization traits randomly from one of these two groups to observe the interaction between the two groups of organizations, the approach is to select 10% (six of 60 organizations) of organizations from the same group to do the change firstly, and we do a 5% change by random for 2 traits from 6 of these genotypic traits among these organizations. For example, if some genotypic trait value of the organization is 0.8, the change range is plus or minus 0.04, it is in between the biggest value of 0.84, and the smallest value of 0.76. And it is the same as in the previous experiment, this experiment will be observed until the system reaches balance. For each result of previous 10 experiments, this experiment will consist of 10 time experiments with different initiated traits valued randomly, therefore totally it will have 100 time experiments. The following experimental results are all presumed that organizations do radical change at system
Fig. 5. The winner radical change experiment.
clock 300. It means that the previous 300 clocks will be ‘‘incremental change experiment” result, and after that it will be the result of radical change experiment. First, as Fig. 5 shows, when the winner side (the organizations adopting ‘‘small incremental change strategies) in the incremental change experiment has radical change at system clock 300, and organizations with ‘‘large incremental change” strategies stay the same. The results show that in an enormous change, they can attract almost all the consumers because of radical change, however, as time goes by, some of the consumers will return to the organizations adopting ‘‘large incremental change” strategies, and lastly, they will reach another stable stage at around system clock 410. So, to sum up, the result is quite similar with the previous one, which means the winner is always the winner, and the loser is always the loser. This result also has a metaphor that the whole system has some attractors on the superior side, though all innovations (changes) will change the possession of consumers in a short time, but, these changes are temporary, the whole system will return back to the attractors finally and result that all changes are in vain. The next experiment is to presume that the loser side (organizations adopting ‘‘large incremental change” strategies) in the incremental change experiment has radical change at system clock 300, and the other side in which organizations adopting ‘‘small incremental change” strategies will stay the same. The approach is the same as the previous one. And the results are quite diverse compared to the experiments done above. We try to classify these 100 experiments into four categories as shown in Figs. 6–9. From 100 experiments, there are 31 times like Fig. 6 which we called ‘‘falls short of success”, 32 times like Fig. 7 which we named ‘‘a flash in the pan”, 28 times like Fig. 8 which we named as ‘‘to convert defeat into victory” and 9 times like Fig. 9 which we call ‘‘be equally matched”. And we will have further explanation based on the above categories as follows. From 100 experiments, there is almost 1/3 of results (31 times) which belong to ‘‘falls short of success” category (Fig. 6), and organizations adopting ‘‘large incremental change” strategies can get very many consumers while doing radical change, but also would lose almost half the numbers of consumers in a short time, afterward, they will have increased numbers of consumers owned by themselves slowly until the system clock is 456 and they will reach another high peak (owning 8401 consumers which occupies around 2/3 of consumer numbers), afterward, they might reverse all the way until after system clock 475, and then no more changes (4507 consumers left which occupy 1/3 of consumer numbers), then, they might lose the leading position. But, the final result
Fig. 6. ‘‘Falls short of success”.
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
8199
consumers (number is 0) within a quite short time (around 20 system clock), afterward, after the 400 system clock, they get some parts of consumers back (only 2221 consumers left which occupied 1/6 numbers), and this result is worse than the previous one before change (3025 consumers). However, when organizations of the loser side adopt innovation change, do they all stay in relative inferiority? From Figs. 8 and 9, this pattern of organizations shows the results of ‘‘to convert defeat into victory” and ‘‘be equally matched”, after radical change, they can get almost all consumers in a short time, and lose some consumers within a quite short time (around 20 system clock), afterward, to try to get some part of consumers back (Fig. 8 only) and reach the total balance situation, and the last result will be better than the original one which has not changed, this describes that the straggler has a chance to reverse the inferiority. Fig. 7. ‘‘A flash in the pan”.
4. Conclusion
Fig. 8. ‘‘To convert defeat into victory”.
Fig. 9. ‘‘Be equally matched”.
(4507 consumers) are still better than the previous result (3025 consumers), even if it makes a small difference small on the consumer numbers for each other. Secondly, 1/3 results from experiments (32 times) belong to ‘‘a flash in the pan” pattern (Fig. 7), and the initial result for this pattern will be similar with ‘‘falls short of success” after change, which means that after the organizations of the losing side can get almost all consumers in a short time after radical change, but, will lose all
Organizational innovation is a complicated dynamic process, if we want to study causal relationship with these interlaced factors among individuals it is quite difficult. This study will establish the interaction between organizational innovation and consumers in an artificial world, and do some controllable experiments through parameter choices and changes, to explore some dynamic problems which other research approaches are not easy to do. We have two types of experiments related to innovations in this study which are incremental change and radical change. The incremental change experiment is to presume the organizations with worse performance adopting the continuous imitate learning from the ones with better performance, and the result says that this type of learning approach will have isomorphism (the products are similar). In the multiplex consuming environment requested, if we adopt the fast imitating learning, it will be over ‘‘isomorphism”, and lose the diverse competition for the products, indirectly, it also shows the importance of adopting the market segmentation strategy. The result of radical change experiment says that when the winner adopts radical strategy, it is limited in the increase of consumer occupation, but, if the loser adopts radical change strategy, it has the chance to reverse the situation, and might have four different patterns (‘‘to convert defeat into victory”, ‘‘falls short of success”, ‘‘be equally matched” and ‘‘a flash in the pan”). In the real world, the winner usually has more resources and a better environment, it does not need to do any radical change relatively, but this offers the chances for other inferior organizations which want to survive, if the latter can adopt the more aggressive change (radical change) relatively, it is the chance to catch up and reverse the situation, and might appear as two patterns (‘‘to turn defeat into victory” and ‘‘be equally matched”) in our study. These results of dynamic process of competition might offer some basis to observe and research the possible reasons further in the future. Peter (2003) indicated that to understand the behavior of the whole, the only way is to understand the individual within the whole, then we can know how the individual makes his/her decision based the goal set by themselves. We simulate the individual behavior through an artificial world establishment in this study, to make these individuals show the outer behavior of the whole by self-organization, this approach also causes us to research some situations in the real world, and try to understand the reasons and the possible results on how the individual influences the whole. For the stage of present research, what the concepts we emphasize in this study is neither on the conclusions through experiments, and nor to shorten the distance between the real world and the artificial world established in this system, but to understand how we can make use of these models established in
8200
T.-T. Lu, J.-C. Chen / Expert Systems with Applications 37 (2010) 8193–8200
this artificial world and study some vague systems dynamic behavior, especially for some accidental or exceptional results from individual experiments. But, we need to emphasize these accidental, exceptional and micro-observational results are mostly not easy to associate with empirical or analytical social research approaches (Morgan, 2005). For the approach of artificial world, though the external validity is lower, but internal validity is high, and we can observe and explain the dynamic process for the interaction of system elements, and through the changes from different parameters which we can do many experiments which are impossible to do in the real world, there must are some certain contributions for the complex system problem. Sendil and Daniel (2004) assert that we should make use of this technology in the research of organization field, and emphasize the complex system interaction research is the one of core researches for management and socioeconomic in recent years. The result for this study is not to be disconnected with the real world, especially when we are in the real world, if there is any similar phenomenon like in this study, the model built in this study might offer a direction to understand, and then match up the approach with external validity to do the research. The above approach is like the case study in the social researches, and the external validity too low to do statistical generalization (Yin, 2003), but when exploring new problems and finding out the process and reason of some phenomenon, this approach has its own contribution. Lastly we would like to explain, the model established in this study is a quite simple model at present, especially in the choices on the traits of organizations and consumers, and this makes this model and experiments to be simple and idealistic. In the future, any specific issues will be based on this model to do the specific experiment to go closer to the real world. Besides this, it can also increase gradually the system function, for example, there will be some certain cost (obstruct) if the consumers move in different organizations, or after evaluation of consumers and organizations it will have time delay to do their action, or consumer cannot get all the information from the market, or consumers have different groups with different consuming behaviors, etc., these are all deeper issues in the future.
References Abernathy, W., & Utterback, J. M. (1978). Patterns of industrial innovation. Technology Review, 80(7), 40–47. Asselmeyer, T., Ebeling, W., & Rose’, H. (1996). Smoothing representation of fitness landscapes – the genotype–phenotype map of evolution. BioSystems, 39(1), 63–79. Bak, B., Tang, C., & Wiesenfeld, K. (1988). Self-organized criticality. Physical Review A, 38, 364–374. Baumgartner, H., & Steenkamp, J.-B. E. M. (1996). Exploratory consumer buying behavior: Conceptualization and measurement. International Journal of Research in Marketing, 13, 121–137. Berry, L. L. (1995). Relationship marketing of services-growing interest, emerging perspectives. Journal of the Academy of Marketing Science, 23(4), 236–245. Carroll, G., & Mosakowski, E. (1987). The career dynamics of self-employment. Administrative Science Quarterly, 32, 570–589. Chen, J. C., & Conrad, M. (1997). Evolutionary learning with a neuromolecular architecture: A biologically motivated approach to computational adaptability. Soft Computing, 1, 19–34. Chen, J. C., Lin, T. L., & Kuo, M. H. (2002). Artificial worlds modeling of human resource management systems. IEEE Transactions on Evolutionary Computation, 6(6), 542–556. Chopra, S., & Van Mieghem, J. A. (2000). Which e-business is right for your supply chain. Supply Chain Management Review, 4(3), 32–40. Conrad, M. (1993). Fluctuons – I. Operational analysis. Chaos, Solitons and Fractals, 3, 411–424. Day, G. S. (1992). Marketing’s contribution to the strategy dialogue. Journal of the Academy of Marketing Science, 20, 323–329. Dewar, R. D., & Dutton, J. E. (1986). The adoption of radical and incremental innovations: An empirical analysis. Management Science, 32, 1422–1433. DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147–160.
Downs, G. W., & Mohr, L. B. (1976). Conceptual issues in the study of innovation. Administrative Science Quarterly, 21, 700–714. Dyer, L., & Reeves, T. (1995). HR strategies and firm performance: What do we know and where do we need to go? International Journal of Human Recourses Management, 8(3), 656–670. Engel, J. F., Blackwell, R. D., & Miniard, P. W. (1995). Consumer behavior (8th ed.). New York: The Dryden Press. Fogel, D. B., Chellapilla, K., & Angeline, P. J. (1999). Inductive reasoning and bounded rationality reconsidered. IEEE Transactions on Evolutionary Computation, 3(2), 142–146. Hargadon, A., & Sutton, R. I. (2000). Building an innovative factory. Harvard Business Review, 78(3), 157–166. Henderson, R. M., & Clark, K. B. (1990). Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms. Administrative Science Quarterly, 35, 9–30. Jacob, M., Hellstrom, T., Adler, N., & Norrgren, F. (2000). From sponsorship to partnership in academy–industry relation. R&D Management, 30(3), 255–262. Kaplan, R. S., & Norton, D. P. (1996). Using the balanced scorecard as a strategic management System. Harvard Business Review(January–February), 75–85. Kauffman, S., & Levin, S. (1987). Towards a general theory of adaptive walks on rugged landscapes. Journal of Theoretical Biology, 128, 11–45. Khilstrom, R., & Laffont, J. (1979). A general equilibrium entrepreneurial theory of firm formation based on risk aversion. Journal of Political Economy, 87, 719–748. Kotler, P. (2005). Marketing management (12th ed.). New Jersey: Prentice-Hall. Kotler, P., & Armstrong, G. (2005). Marketing management analysis, planning, implementation and control (11th ed.). London, UK: Prentice-Hall International Inc. Lu, T. T., Chen, J. C., & Liao, G. X. (2008). Aggressive or conservative, general or specific? A study of organizations adopting different learning strategies in an artificial world. Expert Systems with Applications, 34, 1018–1027. May, B. E., Lau, R. S. M., & Johnson, S. K. (1999). A longitudinal study of quality of work life and business performance. South Dakota Business Review, 58(2), 1–5. Morgan, M. S. (2005). Experiments versus models: New phenomena, inference and surprise. The Journal of Economic Methodology, 12(2), 317–329. Narver, J. C., & Slater, S. F. (1990). The effect of a market orientation on business profitability. Journal of Marketing, 54(10), 20–35. Nicosia, F. M. (1966). Consumer decision process-marketing and advertising implications. New Jersey: Prentice-Hall. O’Callaghan, J., & Conrad, M. (1992). Symbiotic interactions in the EVOLVE III ecosystem model. BioSystems, 26, 199–209. Oliver, R. L. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research, 17(4), 460–469. Peter, C. O. (2003). Game theory and political theory. Cambridge: Cambridge University Press. Plummer, J. T. (1974). The concept and application of life style segmentation. Journal of Marketing, 38, 33–37. Reichheld, F. F. (1996). The loyalty effect: The hidden force behind growth, profit and lasting value. Boston: Harvard Business School Press. Rizki, M., & Conrad, M. (1985). EVOLVE III: A discrete events model of an evolutionary ecosystem. BioSystems, 18, 121–133. Sara, D., Roman, F., & Melanie, R. (2005). To segment or not to segment? An investigation of segmentation strategy success under varying market conditions. Australasian Marketing Journal, 13(1), 20–35. Scott, A. S., & Karl, T. U. (2004). Technological innovation, product development, and entrepreneurship in management science. Management Science, 50(2), 133–144. Sendil, K. E., & Daniel, L. (2004). Modularity and innovation in complex systems. Management Science, 50(2), 159–173. Sherwat, I., & Fallah, M. H. (2005). Drivers of innovation and influence of technological clusters. Engineering Management Journal, 17(3), 33–41. Shipman, R., Shackleton, M., & Harvey, I. (2000). The use of neutral genotype– phenotype mapping for improved evolutionary search. BT Technology Journal, 17(4), 103–111. Shiv, B., & Fedorikhin, A. (1999). Heart and mind in conflict: The interplay of affect and cognition in consumer decision making. Journal of Consumer Research, 26(12), 278–292. Sipper, M. (1995). An introduction to artificial life. Explorations in Artificial Life (special issue of AI Expert), 4–8. Slappendel, C. (1996). Perspectives on innovation in organizations. Organization Studies, 17(1), 107–129. Steenkamp, J.-B. E. M., & Baumgartner, H. (1992). The role of optimum stimulation level in exploratory consumer behavior. Journal of Consumer Research, 19(12), 434–448. Terano, T. (2000). Analyzing social interaction in electronic communities using an artificial world approach. Technological Forecasting and Social Change, 64(1), 13–21. Thompson, D. V., Hamilton, R. W., & Rust, R. T. (2005). Feature fatigue: When product capabilities become too much of a good thing. Journal of Marketing Research, 42(4), 431–442. Tushman, M. L., & Anderson, P. (1986). Technological discontinuities and organizational environments. Administrative Science Quarterly, 31, 439–486. Venkatraman, N., & Ramanujam, V. (1986). Measurement of business performance in strategy research: A comparison of approaches. Academy of management Review, 11(4), 801–814. Wolfe, R. A. (1994). Organizational innovation: Review, critique and suggested research directions. Journal of Management Studies, 31(3), 405–430. Yin, R. K. (2003). Case study research: Design and methods. London: Sage Publication.