Erik van Doesburg
University of Groningen Faculty of Economics and Business
MBA Change Management
Aweg 4-3 9718CS Groningen 06-29135247 e.p.van.doesburg@student.rug.nl S.2149524
Force of Habit
November
2014
How Incumbent Habits From A Legacy System Influence Individual Adaptation To New Information Systems
1 Acknowledgements
First and foremost I want to thank my girlfriend Elina and each and every member of my family, without your support and love I would never have written this thesis in the first place.
I would also like to thank my supervisor, dr. B. Müller for his faith and guidance during the creation of my thesis. At times when I got stuck in my thoughts one single remark or suggestion from his side could clear my head and make the penny drop to continue.
Last, but certainly not least, I want to thank B. Roosenthaler who offered me insight on the decision making within the UDT case, proofreaders T. Oost and W. Zijlstra and of course all of my colleagues who participated in this study.
2 Abstract
Purpose: To determine how incumbent habits that were formed in a legacy system influence individual adaptation behavior when using new information systems.
Methodology: This paper presents a case study performed at one of the mayor Dutch banks.
In the light of this study several employees were interviewed and observed.
Internal documents were consulted to give a broader perspective.
Propositions were formed and preliminary results were found regarding the theories of punctuated equilibrium (Eldredge & Gould, 1972), the CMUA model (Beaudry & Pinsonneault, 2005) and habit development / disruption strategies (Polites & Karahanna, 2013).
Findings: The findings of this paper show that incumbent habits not only influence the outcome of coping strategies, it also describes how they influence familiarity pockets and thus adaptation. The strength of habits that are formed within a legacy system are so strong that they can result in a complete or partial exit when it comes to coping strategies even if the user’s initial appraisal of the system is positive. Yet by providing an environmental trigger the incumbent habits can be triggered within the new environment which in return not only broadens the initial familiarity pocket, it also led to the formulation a new strategy regarding habit development strategies.
Keywords: Technological Adaptation, Habits, Familiarity Pockets, Coping Strategies, Punctuated Equilibrium, Legacy Systems
Paper type: Case study
Word count: 13.887
Supervisor: Dr. B. Müller
3 Contents
1. Introduction ... 4
1.1 Background ... 4
1.2 Research Question ... 6
2. Theoretical section ... 7
2.1 Legacy Systems ... 7
2.2 Habits, Routines and Experience ... 8
2.3 Familiarity Pockets ... 11
2.4 Coping strategies ... 12
2.5 Conclusion ... 15
3. Method ... 16
3.1 Background - Bank Case ... 16
3.2 Collecting data ... 17
3.3 Analysis of the data ... 18
3.4 Validity and Reliability ... 19
4. Results ... 20
4.1 Initial Appraisal followed by Inertia- Coping & Habits ... 20
4.1.1 Interpretations ... 23
4.2 Changing Habits – Coping & Habits ... 25
4.2.1 Interpretations ... 28
4.3 Familiarity Pockets – Coping & Habits ... 29
4.3.1 Interpretation ... 31
5. Discussion and Conclusions ... 32
5.1 Limitations and future research ... 32
5.2 Conclusion and Recommendations ... 32
References ... 35
Appendixes ... 40
Appendix 1 – CMUA model ... 40
Appendix 2- Interview protocol ... 41
Appendix 3 – Interview questions ... 42
Appendix 4 – Background Participants... 44
Appendix 5 - Coding Tree ... 47
4
1. Introduction
Recent research underlines that habits are important drivers when it comes to IS acceptance and continued usage (e.g. Kim & Malhotra, 2005; Limayem, Hirt, & Cheung, 2007; Polites & Karahanna, 2012, 2013). While the antecedents that influence the final adoption of new IS have been extensively researched in change management literature (e.g. Benbasat & Barki 2007; Morris & Venkatesh 2010;
Venkatesh et al. 2012; Davis et al. 1989; Beaudry & Pinsonneault 2005) and the field has come a long way in explaining user behavior, there are still some blind spots when it comes to the understanding of individual adaptive behavior1 (Chin, Marcolin & Newsted, 2003). This study will look at how habits that were formed in incumbent systems influence individual adaptive behavior. As a result it will build theory, through a case study, on the differences of adaptive behavior of IS users and will focus on the influence that incumbent habits that were gained within legacy systems2 have on individual adaptive behavior.
1.1 Background
Habits can form an obstacle to repeated use of a new system, especially if the incumbent system is still accessible to the user (Polites & Karahanna, 2012; Polites, 2009). This is an important notion since a lot of the current IS implementations are (partial) replacements of incumbent systems (Polites & Karahanna, 2013). Habits and routines are a relatively unstudied part within the field of technological change and habitual behavior. Getting used to new systems means that a user needs to get rid of long standing habits and create new ones. This is a hard task since the force of those habits and routines unintentionally force users back to their old ways of working (Verplanken & Wood, 2006). Disrupting these habits and routines and stimulating the development of new habits might prove an efficient way to increase system usage (Polites & Karahanna, 2013). A hindering or enabling role of incumbent system habits on individual adaptation behavior when using new technological applications has not been described in literature yet and will be addressed in this thesis. In order to achieve this, I will link the literature of behavioral habits to IS acceptance, coping and familiarity pocket literature.
The academic community has primarily focused on individual acceptance and intended behavior following technological change. Within the field of social ontology, research concerning technology mediated change, predominately points to human agency (Boudreau & Robey, 2005). This has manifested in well-established models like; The Theory of Reasoned Action (Ajzen & Fishbein, 1980), The Theory of Planned Behavior (Ajzen, 1991), The Technology Acceptance Model (TAM) (Davis et al., 1989) and Unified Theory of Acceptance and Use of Technology (UTAUT) (Visnawath Venkatesh et al., 2003). Traditionally speaking the field of IS acceptance has been classified as described in figure 1.
Figure 1: Traditional IS acceptance research - model adapted from Polites & Karahanna (2013)
1“The cognitive and behavioral efforts performed by users to cope with significant information technology events that occur in their work environment.” (Beaudry & Pinsonneault, 2005, p. 493)
2 “Software systems that we don’t know how to cope with but that are vital to our organization” (Bennett, 1994, p. 19)
5
As mentioned, Davis et al. (1989) and Venkatesh et al. (2003) are key figures in the field of acceptance literature. Both authors state that personal, as well as contextual factors play a large role in the adoption of a new system. Furthermore, both studies make a clear distinction in outcomes resulting in use or non-use. Models such as TAM and UTAUT however assume that intention leads to behavior, but also that once that behavior has occurred a single time, it will be repeated. It ignores that users might have to go through the entire circle again once they are confronted with the new technology for a 2nd, 3rd or even a 1000th time. In my opinion, this assumption is one of the major flaws within said models. While I do not contradict that humans have the capacity to make rational choices and evaluate those choices, I challenge that humans will always evaluate those choices, be it either intentional or unintentional. In that way human agency ignores that users have gained routines and habits3 from the use of (legacy) systems throughout the years. By looking at those actions that were unintentionally not evaluated, I will shed new light on how adaptive behavior occurs.
The actual (non-)usage of the system is part of adaptive behavior because people have a predisposition towards the system and use the system in different ways (Boonstra & Van Offenbeek, 2010). The degree to which all of the features of a system are utilized can be described as the assimilation of an information system (Cooper & Zmud, 1990; Volkoff, et al., 2007). A common result after implementation of new technology is partial adaptation (Jarvenpaa & Ives, 1993). To illustrate partial adaptation in everyday life, think of smartphones. While most users use their smartphone not only to have phone conversations, but also use it to check time, e-mails and for instant messaging, there are plenty of people walking around with physical agendas, instead of the integrated one on their smartphone or use mp3 players to listen to music. Beaudry & Pinsonneault (2005) are among the few that focus on the individual level and developed a model, called Coping Model User Adaptation (CMUA), which accounts for a wide range of user behaviors that are categorized as benefits maximizing, benefits satisfying, disturbance handling and self-preservation.
The responses resulting from the CMUA model can be classified as coping strategies4 (Beaudry &
Pinsonneault, 2005). In the available literature of coping strategies concerning technological change, models usually focus on the change as a whole. This means that the user in question is assumed to have one predisposition towards the entire implementation. Users will be inclined to act only according to one of the behaviors mentioned by the coping strategy theories. This assumption ignores that technology is supposed to be a multifaceted solution, when it aims to provide a more integrated way of working. It could well be that there are parts of the solution that will be met by the user with acceptance while other aspects will meet resistance and will be ignored. Furthermore, just like the earlier mentioned acceptance models, this model is presented in a way that once the user has gone through the different stages, it is a done deal. Even though the user appraises the event as an opportunity the sheer number of failed IS implementations suggests in my opinion that it does not always lead to individual efficiency and effectiveness. This calls for more research regarding this subject.
3“The non-deliberate, automatically incalculated response that individuals may bring towards the behavior of IT usage.” (Limayem et al., 2001, p. 275)
4“The cognitive and behavioral efforts exerted to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person” (Lazarus & Folkman, 1984, p. 141)
6
1.2 Research Question
Taking the limitations and gaps in to account where the academic field stands right now, this thesis will draw upon recent research in technology use and adaptation, together with research on routines and habits, coping behavior and the user adaptation of technology, to examine the following research question:
In what way do incumbent system habits formed within a legacy system environment play a role in individual adaptation behavior when using new technological applications?
To answer this question, first a review of current literature has been conducted and the findings of the different constructs have been synthesized in chapter 2. This same chapter will state the conceptual model and the propositions that have been examined during the case study. The propositions that are formed in chapter 2 are the results of inductive reasoning and reduce the research question to a testable and falsifiable form. The case study has been conducted at a company where the implementation of an organization wide IS is in progress right now. Side-by-side monitoring, interviews and reviews of internal documents have been used to come to insights concerning the matter at hand and have been held against the findings from the academic field and the propositions of this paper.
The purpose of this research is to offer contributions in an ostensive academic manner as well as in a practical managerial manner. This paper’s contribution to the literature will be an enrichment in the field of habitual behavior and technology acceptance. Furthermore it will make suggestions on the refinement of coping strategies and will be an extension of Polites & Karahanna's (2013) paper on the imbeddedness of IS habits in organizational and individual routines by providing additional propositions concerning habit disruption and habit development strategies. By linking the existence of legacy systems with the literature of IS habits, acceptance of change and coping mechanisms, a more integrated view of change literature will be established.
This study will provide managers insight on how the existence of incumbent system habits that were formed within a legacy system influence individual adaptation of new technological applications, and will show them new strategies about triggering the incumbent habits in a new IS environment. It will provide suggestions on the set-up of new systems that are not ready to be implemented, but are still in the development. Gaining better insight will offer managers the possibility to make better use of available tactics of IS development and implementing strategies concerning technological change.
7
2. Theoretical section
The purpose of the theoretical section is to tie the different constructs together in a cohesive story and show their underlying relatedness. This section will give a recap of the foundations on which this field has been build and will provide definitions of the terminology used throughout this paper. The success of technological implementations is significantly influenced by current practices and how the implementation unfolds (Lapointe & Rivard, 2007). Understanding the factors and dynamics that influence these behaviors is central to this work. To better understand the role of incumbent habits of a legacy system on the individual adaptation behavior of IS usage, it is important that the different constructs of this framework are based on both intentional as well as automatic determinants of user behavior.
As said, the field of IS acceptance and usage has been researched extensively and has resulted in several models that examine the variables leading up to acceptance, like the Technology Acceptance Model (Davis et al., 1989), but also models dealing with different behavioral mechanisms when users are confronted with change like the Coping Model of User Adaptation CMUA (Beaudry &
Pinsonneault, 2005). By explaining how routines and habits work within IS usage, the different models will be linked in a conceptual model and propositions will be formulated to showcase the relations within the conceptual model.
2.1 Legacy Systems
The implementation of new systems is often performed in order to (partially) replace systems that users have worked with for extended periods of time (Polites & Karahanna, 2013), such systems can often be described as legacy systems. Polites & Karahanna (2013) use the term incumbent system; I will focus on legacy systems. While a legacy system often is the incumbent system, it does not mean that every incumbent system is a legacy system. An incumbent system is the current system, but that does not automatically mean that it is used often, nor does it imply that the system has been in use for an extended period of time. Especially this last criteria is, in my opinion important, if it comes to changing habits and routines. A legacy system can be defined as “[large] software systems that we don’t know how to cope with but that are vital to our organization” (Bennett, 1994, p. 19).
Furthermore Bennett provides characteristics that apply to most (but not all) legacy systems:
1. The system is over 10 years old 2. Written in an old coding language
3. It performs crucial work for the organization
4. Hard to change the system
5. Has a long history of intensive main- tenance
6. Specialized knowledge
Another, yet similar definition, states that a legacy system can be defined as “a mission critical software system developed sometime in the past that has been around and has changed for a long time without undergoing systematic remedial actions” (Lucia et al., 2001, p. 1). This latter definition is also more in line with the laws of program evolution (Lehman, 1980), which state that the underlying principles of what a legacy system entails are: the law of continuing change, which states that a program must undergo continual changes or it will become progressively less useful in the real world.
The second law, the law of increasing complexity, argues that the structure of evolving software will degrade unless remedial action is regularly taken.
8 | P a g e These characteristics apply to the technical aspects of the system. From the user-perspective, this is the system that employees have worked with, often as long as they can remember and is part of their work identity. The mere existence of technology has social implications since it influences people’s interpretations of technology and their actual behavior (Boonstra & Van Offenbeek, 2010).
In the end there are but few options what do with legacy systems when the laws of program evolution have come true and the legacy system becomes too outdated to keep up with current developments. One of the options to handle this, is to encapsulate the legacy system as a component in a new system, when implementing change (Bennett, 1994). However users might not find it easy to switch from the legacy system to a new system, learn how to operate it, and break with their old routines resulting in inertia (Boudreau & Robey, 2005).
2.2 Habits, Routines and Experience
Many displays of human behavior have the tendency to be based on frequently exhibited goal- orientated patterns which are performed in a mindless manner (Aarts et al., 1998; Polites &
Karahanna, 2012). The usage of regularly used systems becomes habitual over time, so prior use has been described as a predictor of habit as well. When a new system is introduced, the gained knowledge through prior use cannot automatically be transferred to the new system because users need to learn again how to operate the new system (Polites & Karahanna, 2012). They have often used incumbent systems for multiple years. Throughout that extended period of use, habits and routines of use have been formed. Especially with legacy systems the strength of these habits and routines can be fierce. Habits can be defined as the non-deliberate, automatically inculcated response that individuals may bring to IS usage (Limayem, Hirt, & Chin, 2001). This notion is in line with Aarts et al. (1998) statement about the goal-directed nature of habitual behavior where they explain that in order to start walking, which is behavior we do not truly think about, we need a destination to reach. Only once a person has determined where he wants to go and how to get there, the process of automatic behavior of reaching the determined goal, in this case walking, takes over.
This is the same for usage of systems when a user is confronted with a specific task (Polites &
Karahanna, 2013).
There is a clear-cut difference between habits and routines. While habits are formed on an individual bases, routines are described as an executable capability for repeated performance in some context that has been learned by an organization in response to selection pressures (Hodgson & Knudsen, 2004). Both habits and routines are within the context of this case study, which is set in a working environment, and are both applicable since as an individual you can still perform a routine which was instigated through the group.
The main take away is that once users have formed routines using a specific system for a certain task, they automatically return to that same system over and over again if they have to preform that task.
So habits and routines can disrupt the cycle that is proposed in theoretical model concerning
9 | P a g e technology acceptance and adaptation behavior. If we were to place habits and routines and their interplay with acceptance within the field of IS acceptance literature, it can be modeled as depicted in figure 2.
Figure 2: The influence of habits on traditional IS acceptance research
If a program that is used on a daily basis will be replaced by another program, the habit of using the original program needs to be changed as well. Since habits are partially automated behaviors, it is hard to change them (Aarts & Dijksterhuis, 2000). The mere intention of a person to change his behavior might only be a successful strategy if the strength of the habit is either weak or moderate in terms of Verplanken, Aarts, & Knippenberg (1997), for strong habits it will take more effort. The strength of a habit can generally be determined by the frequency of performance in the past while it took place in a similar setting (Ouellette & Wood, 1998). With strong habits, like the multiple times a day usage of an application, the intention to change habits has been shown to be unrelated to the actual behavior (Holland, Aarts, & Langendam, 2006). And even if habits have been changed, the chances of relapse are high (Polivy & Herman, 2002).
Methods to effectively change both weak and strong habits have been linked to the punctuated equilibrium theory (figure 3), which was introduced by Eldredge & Gould (1972). This theory states
Figure 3: Punctuated equilibrium model adapted from (Burnes, 2009)
10 | P a g e that development is marked by isolated periods of rapid change between long periods of time with little to no change called stasis (Orlikowski, 1996), or as Polites (2009) describes it; inertia. The punctuated equilibrium theory was first described in the field of biology and its use was later linked to the field of habitual behavior by, among others; Aarts, Paulussen, & Schaalma (1997); Ouellette &
Wood (1998) and Verplanken & Wood (2006).
These papers suggest that a mayor change within the environment, while stimulating the formation of new habits, is one of the most likely strategies to break with old habits. This implies that without a change of the context, be it physical or mental change, it is hard to break habits and inertia will prevail. Orlikowski (1996, p. 64) states regarding revolutionary change within the model; “Punctuated discontinuities are typically triggered by modifications in environmental or internal conditions, for example, new technology, process redesign, or industry deregulation.” When it comes to period of relatively little or no change, Polites (2009, p. 151) states: “Inertia has a negative impact on intentions to use the new system, above and beyond its impact through perceptions. Thus an individual using a system in an inertial state may perceive a new system as useful and easy to use, yet not voice intentions to actually use it.” Reasoning within the line of thought of those two statements, my first proposition is:
Proposition 1: Implementing new system with a revolutionary approach will have a positive relation with actual system usage and therefore the rate of forming new system habits will be higher.
This proposition is more likely to be successful if access to the original system is limited.
So far literature has focused on the disruption of old habits (Ortiz de Guinea & Markus, 2009;
Ouellette & Wood, 1998; Polites & Karahanna, 2012; Verplanken & Wood, 2006; Webb, Sheeran, &
Luszczynska, 2009) from a, in my opinion, managerial perspective. The proposed techniques are mainly concentrated at changing the environment that limits the triggers that activate old habits. I want to shift this view towards how habits from the legacy systems can play an active and positive role in the formation of new habits in the new system and thus creating an approach that is more focused on facilitating the end-user. Even though gained knowledge through prior use cannot be automatically transferred to new IS (Polites & Karahanna, 2012), the occurrence of habitual behavior can be triggered if supporting features of the current environment are similar to those contexts in which the behavior was learned and practiced in the past (Ouellette & Wood, 1998). This implies that the occurrence of habitual behavior is context dependent.
Let me illustrate this with a personal anecdote. Some ten years ago I moved from The Netherlands to Cyprus. While in The Netherlands they drive on the right hand side of the road, in Cyprus they use the left hand side. Luckily driving on the other side did not pose a problem for me, but there was one habit that was constantly triggered which had some funny results. The car I was driving in Cyprus of
11 | P a g e course had its steering wheel on the right side of the car. The one thing that was different from the Dutch cars was the positioning of the windscreen wipers and the direction indicators; they were switched from the left to the right and vice versa. So the environment that I was operating in was similar to what I was used to, but the actual usage was different. As a result I often switched on the windscreen wipers when I wanted to indicate my direction. And once I moved back to The Netherland the same thing happened again, because I got used to the new configuration and developed a new habit.
When it comes to IS related change, I argue that by partially rebuilding the context, in terms of lay out and task sequence of the incumbent legacy system, within the new system users experience the same triggers to display habitual behavior as they did before. So in that sense I reason that through similarity of the systems and recognition by the user, habitual behavior from the past can be triggered in a new environment.
Proposition 2: When system designers use a similar interface, and use the same task sequences as were used in the incumbent legacy system, habits that were gained while using the incumbent system will be triggered within the new system.
2.3 Familiarity Pockets
Familiarity pockets are the construct that tie habits, IS acceptance and coping together. An IS user's familiarity pocket comprises work routines and components accumulated through situated interactive use of the system and can be roughly defined as a user’s sphere of action. Meaning that the focus of a familiarity pocket is not so much the actual familiarity with the system, but more so the routines and habits gained by the user through the interaction with the system and/or other users (Yamauchi & Swanson, 2010). In terms of my conceptual model, the familiarity pocket is made up by the boxes of “actual new system usage” and “new system habits” (see figure 4). This implies that users know how particular features of the system work, either through prior use of similar features or newly learned practices. Different studies have shown that users typically don’t use all possible features (technological infusion) of a system, but stick with a rather limited set of known practices (Japerson et al., 2005; Orlikowski, 2000). When faced with situations that are out of the boundaries of the familiarity pocket, a user can resolve to workarounds (Yamauchi & Swanson, 2010). These workaround need to mask the user’s inability to select the appropriate feature within the new system. This action can either be intentional as a form of resistance or unintentional when the user is not aware of particular features within the system. This notion is in line with the findings of the case study described in chapter 3. Besides that the familiarity pocket can be seen as a sort of save haven of all the features that a user knows, it is also a representation of all the features that the user doesn’t know (Yamauchi & Swanson, 2010). Routines that are performed by users within their familiarity pocket mask much that is not known by the users. While users achieve a level of competency with the features that are within their familiarity pocket, they can often completely
12 | P a g e ignore features that are outside their familiarity pocket. The workarounds that the users invent will eventually make sure that the users get the job done. Yamauchi & Swanson (2010) coined this phenomenon competent ignorance. As mentioned, familiarity pockets are closely related to learning behavior (Yamauchi & Swanson, 2010). Developing new routines and habits is part of learning behavior. This notion is in line with findings of Boudreau & Robey (2005) who state that what people learn is not so much about what they learn during formal trainings, but can also be largely contributed to what they learn from unplanned activities that spread knowledge among the users.
Figure 4: Familiarity pocket in relation to traditional IS acceptance literature
The (encapsulated) legacy system can be part of a familiarity pocket (Beaudry & Pinsonneault, 2005) from which the user can expand its knowledge about the system, but might also be an obstructer of infusion if the legacy system itself is regarded by users as superior in reliability and/or use (Bennett, 1994). Functions of a legacy that resurface in new IS are familiar for the user and routines obtained while using the legacy system can be transferred to the new system and help in forming familiarity pockets within the new system. Coping strategies might help to move outside the familiarity pockets, gain more experience and thus expand the familiarity pocket.
Proposition 3: When system designers use a similar interface, and use the same task sequences as were used in the incumbent legacy system, users will invent less workarounds since incumbent habits are triggered.
2.4 Coping strategies
While intentionally changing habits and routines is hard, it is possible. The mechanics that come in to play if (behavioral) change is imminent are classified as coping strategies. So once the IT event has been appraised by the users and intended behavior can be measured, several reactions can occur,
13 | P a g e and at this point of time coping mechanisms are set in motion. Changes in the environment produce uncertainty and as uncertainty grows, problems start to occur (Benamati, 2001). In reaction to these problems, coping strategies are deployed. Coping is defined as “the cognitive and behavioral efforts exerted to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person” (Lazarus & Folkman, 1984, p. 141). Although no undisputed definition of the different coping strategies exists (Ashford, 1988), the aforementioned definition will be used regarding this study.
Adaptation behavior describes how users can react and how the implementation of an IT event can change IT functionalities, the users routines and habits, or the user’s perception of work. When placed within the traditional IS acceptance model, it has a rather wide focus ranging from the first appraissal of the event up to the eventual behavior. Figure 5 depicts how coping is connected to the different stages of IS acceptance
Figure 5: Coping strategies in relation to traditional IS acceptance literature
One of the main drivers behind coping behavior is the desire to reduce uncertainty (Ashford, 1988;
Bradac, 2001). Problem focused adaptation is predominately focused on the external aspects of adaptation, but does not concentrate that much on the inner self of the individual undergoing the change. When it comes to emotion focused adaptation behavior, a clear distinction between avoidance and rapprochement can be made (Carver & Connor-Smith, 2010; Ebata & Moos, 1991;
Roth & Cohen, 1986; Skinner, et al., 2003). This is closely related to Beaudry & Pinsonneault's (2005) CMUA model (see Appendix 1)(Beaudry & Pinsonneault, 2005), where the user can see the implementation as an opportunity, so he will look for approachal, or the user will see the implemtation as a threat and will employ the avoidance method. The approach method can be divided in two types of behavior where the person either uses problem solving trying to deal with the problem directly or in the other case he/she will look for guidance and support. When a person reacts and displays the avoidance method he/she will either look for alternative source to achieve satisfaction (i.e. work-arounds) or he/she will try to reduce tension by expressing negative feelings.
Uncertainty can lead to avoidance, but uncertainty can be reduced by perceived similarities, which in return would lead to a situation where the person is more open to approach (Bradac, 2001).
14 | P a g e Proposition 4a: There is a positive relation between perceived similarity of the legacy system
and new IS and the approach method.
The user is more likely to move beyond the scope of his familiarity pocket and will show active exploration and information seeking behavior.
Proposition 4b: There is a positive relation between perceived differences of the legacy system and new IS and the avoidance method.
The user is more likely to stay within the confines of his familiarity pocket and will make display no active behavior in trying to expand it.
The way users appraise the situation, influences their path of behavior. Avoidance type behavior leads to a significant reduction of the possibility that infusion, a concept part of the four stages of assimilation of Cooper & Zmud (1990), is reached, but approach type behavior does not lead to a significant increase of the possibility of infusion (Fadel, 2012). “In other words, emotion-focused behaviors such as seeking social support and positive reappraisal may help users achieve a sense of emotional equilibrium but neither enhance nor diminish their degree of system use.” (Fadel, 2012, p.
7). Users that engage in problem focused adaptation are more likely to reach infusion and achieve individual efficiency and effectiveness due to their deeper use and knowledge of the system (Fadel, 2012; Goode, 2012) where users that primary display emotion focused coping behavior are less likely to reach infusion and more likely to opt out (Goode, 2012).
CMUA also implies that as long as the user’s primary appraisal sees the IT event as an opportunity, he or she will always achieve “individual effectiveness and efficiency” as an outcome. While I acknowledge the plausibility that a positive appraisal of an IT event is more likely to achieve an outcome with individual effectiveness and efficiency, I do not think that the process is neither linear nor rational. Habits are partially automated behavior and have little to do with rational intentions to use a system (Aarts & Dijksterhuis, 2000). The notion that this kind of behavior is automated also explains the concept of action slips5 (Norman, 1981) even if the user feels in control and has a positive attitude towards the change. Therefor I propose that the strength of a habit will moderate the eventual outcome of the coping sequence.
Proposition 5: The strength of a habit will moderate the outcome of the adaptation strategy as proposed in CMUA.
5“The performance of an action that was not what was intended” (Norman, 1981, p. 1)
15 | P a g e
2.5 Conclusion
When combining the pillars of this theoretical section; coping behavior, familiarity pockets, routines and habits, we can see the interconnectedness of these constructs. IS acceptance literature is an extensively broad field and the conceptual model (figure 6) depicts how these constructs interact with traditional IS acceptance literature and with each other. The proposed conceptual model visualizes where coping strategies are deployed within the different phases of IS acceptance. As described, there is an overlap between coping strategies and the formation of familiarity pockets.
Coping strategies are deployed up to the point where the user actually starts using the new system, while that same usage is determined by the user’s familiarity pocket. While the incumbent system habits are proposed to influence the outcome of the user’s coping behavior depending on their strength, it also influences the relative size of the user’s familiarity pocket when it comes to the degree of perceived similarity between the systems. As said, habits and routines are key determinants when it comes to IS usage behavior. If these routines and habits need to be changed due to IS change several habit disruption and development strategies can be used. These strategies will influence the new system habits and thus also the familiarity pocket.
Figure 6: Conceptual model
16 | P a g e
3. Method
Case study approach is considered to be the right approach when ‘how’ or ‘why’ questions are asked about a focal phenomenon over which researchers have little or no knowledge or control (Yin, 2009).
Case studies are also considered to be appropriate when researching contemporary questions in natural settings where little or no previous research has been done (Liu et al., 2011). These features of case study method fit well with the objectives of understanding how the existence of incumbent habits influences individual adaptation behavior. This study is set up in a way that it follows the guidelines for conducting case studies as prescribed by Eisenhardt (1989).
The foundation of this paper lies in grounded theory and focuses on the expressed thoughts and feelings of the users and their actual behavior. To gather the necessary information for the case study, side by side observations as well as oral and written interviews were conducted. Furthermore access was granted to internal documents about system usage and user behavior. The interviews are interpretations and opinions of individual users and the conclusions drawn in this paper are my personal interpretations of those statements. This means that this paper does not aim to present an objective truth for all situations, but rather tries to use storytelling as a lens to give a detailed examination of the observed phenomena within this case. These observed phenomena will be compared to the earlier stated propositions and additional insights will be shared.
3.1 Background - Bank Case
Every company starts with a simple operating system, but over the years acquisitions of new divisions and mergers might take place, making the once upon a time simple way of working, more and more complicated with numerous systems. These old systems can become legacy systems. To investigate the impact of the mere presence of legacy systems on adaptation to newly introduced IT systems, I conducted research at one of the major Dutch bank and insurance companies, which in this paper will be called Bank for Regular People (BRP).
In 2010 BRP made an organizational wide decision to update and simplify most of the systems that employees from different brands of BRP, but also inter-organizational departments, had to work with. A typical employee had to use up to 15 different systems a day, just to answer the customer’s questions. Most of the systems at hand were developed in the early 1990-ies and were not considered to be user-friendly anymore. These different systems were to be integrated in one unified desktop system (UDT). UDT would be accessible for every brand within BRP’s organization. To integrate the way of working, is one of the main strategic decisions to change an organization (Rugman & Hodgetts, 2001) and is a logical step for the organization to optimize their business.
BRP decided to build most of UDT in-house while adding custom build components. They acknowledged that building UDT would be an immense task and decided that evolutionary implementation would give BRP the best option to create UDT according to everybody’s wishes. The
17 | P a g e evolutionary approach made it possible to fine tune the program when needed, but also keeps the users involved in the development of the system. This particular setting forms a great opportunity to test the propositions mentioned in chapter 2 and to find an answer on this paper’s research question.
The decision to build UDT in-house and acquiring several custom build Kana components was made after an extensive selection process. One of the main advantages of the chosen package is its flexibility. Since most of the core functions of the different operating systems will keep on running in the background and need to be connected UDT a lot of flexibility of UDT is required. UDT is the umbrella that connects all the different systems in one single screen. UDT’s mission is to achieve that 80% of the information within UDT is available within four mouse clicks.
In the initial set up of this research two of BRP’s brands were selected for investigation, Alpha and Beta. Brand Alpha worked with a specific program that was not available for brand Beta, but would be implemented in UDT for both brands. During the research period it became apparent that there were some mayor differences between the implementation of UDT between the two brands. The gradual implementation strategy within Alpha was not replicated in Beta. Beta would experience a revolutionary implementation where all the systems would be replaced in a short period of time, which would be a perfect opportunity to investigate proposition 1. Unfortunately it became apparent that the implementation at Beta would be delayed several months. As a matter of fact the implementation would occur only after the deadline of this paper. I decided to continue the research while investigating just one of the brands and build a case study on the information obtained from Alpha’s users.
Due to the size of this project, Alpha opted for gradually implementing features one by one, instead of a revolutionary approach where the entire finished product was delivered at once. In terms of type of change, UDT can be classified as evolutionary change instead of revolutionary. This also meant, in combination with the flexibility of UDT, that if users were dissatisfied with certain features, there was time, room and budget to improve. The implementation process started in late 2010 and is still in progress with both major releases of new applications, as well as minor fine tuning within existing features. The use of this system is largely targeted at the call center agents and local branch office employees who are in direct contact with customers, but it is also available to employees of the different back offices. Within the framework of this study, I restrict myself solely to observing and interviewing agents of the call centers. To conduct this research also among branch offices and/or back offices is not feasible given the time frame of this study.
3.2 Collecting data
Multiple data collection methods strengthen the grounding of proposed theories through triangulation of the evidence (Eisenhardt, 1989). In order to achieve this, I used multiple sources of
18 | P a g e evidence. First of all side-by-side observations were performed, in total about 15 hours. After the observations, the same person was asked to participate in a semi-structured interview. The interview protocol and the interview questions can be found in appendix 2 and 3, while an overview of the participants’ backgrounds can be found in appendix 4.
Interviewees were assured that discussions were strictly confidential and the content of each interview was reviewed and signed off by each interviewee as being a truthful representation of the interview. Interviews typically started with open-ended questions about the system that was being implemented, followed by more specific questions about their involvement with, understanding of, and attitudes towards aspects of the system, such as the impact on working processes. In total 6 persons participated in the oral interviews and most of these interviews lasted for about 20 minutes.
Some of the employees that I asked to participate in the interview sessions did not feel comfortable with being recorded. They did however offer a lot of off the record information. They were also asked to answer the interview in written form. Eventually 4 of the additional users that were asked to answer the questions by a written reply complied with this request. Although this method did not give a direct option to ask follow up questions or elaborate on the answer, it aided in the analysis and could either support or refute statements made by the interviewees. Strangely enough it were predominately the men who chose to do the oral interview, while the women gave off the record information and decided to do the interview in written form. Afterwards I asked why some of the participants made this decision and the men stated that they did not truly think about refusing and did not think about other options to aid this research, while the women in general stated that they felt more comfortable about answering the questions on their own and having the opportunity to think about their answers instead of answering instantly.
3.3 Analysis of the data
The analysis of the data follows the conceptual model and theoretical propositions since that shaped the orientation of the data that had to be gathered. The focus of this paper is on individual adaptation, which is part of coping, and the interplay with habits and there for most paragraphs in the results section are dedicated to coping behavior and habits in relation to the other constructs.
Since the coping cycle is a sequential model, the case is also presented in a sequential manner.
Furthermore patterns within the data will be identified and analyzed in order to build the case and to support the interpretations. The theoretical predicted events are compared to the empirically observed events. As a result the overall set up of the analysis is in line with what Yin (2011) describes as the logic model. Recurrent patterns in the entire data set were grouped, coded and then analyzed (see appendix 5: Coding tree).
19 | P a g e
3.4 Validity and Reliability
To ensure construct validity, several sources for obtaining data were used: side-by-side observation, oral interviews, written interviews and internal documents. Internal validity was sought after by looking for patterns, dominant themes and explanations in the material. I tried to avoid bias by asking respondents to comment on my interpretations.
The very fact that I work as a customer service employee at the cooperation that is examined for this case study and use the actual systems that are discussed, may have an advantage since context- dependent knowledge and experience are at the very heart of expert activity and lie at the center of the case study as a research method (Flyvbjerg, 2006). My work at BRP is at same department as the participants of the case study. Besides working at the customer service, I am part of the user group that advices the project leaders of UDT on the development of UDT when it comes to practical implications. This entails that as an employee; I am familiar with the internal terminology of the organization and know which specific questions to ask when I need to delve deeper in to an answer or comment of the interviewees.
Before starting the interviews, the participants agreed that I could observe their behavior regarding system usage. During the observations the participants helped actual customers, which meant that neither the customer’s questions nor the systems that had to be used were staged.
All of the interviews were recorded for transcribing purposes (see digital appendix: Interviews).
Respondents checked the accounts for accuracy and sometimes suggested changes. The interviews were semi-structured and contained a set of 25 fixed questions that were based on the work of Lassila & Brancheau (1999), Moore & Benbasat (1991) and Sun (2012). Additional questions were asked to clarify statements of the interviewees when needed or to delve deeper into the underlying arguments. As for the off the record conversations, no detailed records were kept; only key words and snippets of thoughts were recorded on paper.
Access to internal documents regarding the implementation of UDT was also granted, but due to the sensitive content of these documents they had to be omitted from the research appendixes. But they did contribute to the sense making process about why and how certain decisions were made during the implementation of UDT, but also about the results of the implementation.
20 | P a g e
4. Results
In this chapter the results from the case study will be presented. In chapter 2 the constructs: coping strategies, familiarity pockets, habits and routines, were introduced and related to one another. The conceptual model and propositions that were formed in the second chapter will be used as a guide line during this chapter to tell the story of this particular case. As mentioned, several constructs have been connected within this paper’s conceptual model and for the sake of comprehensibility the interpretations regarding each particular subject will presented straight away instead of one final section dedicated to interpretations.
At the very heart of this case study are the participants, so I will start with a table providing a short introduction of their background. A more detailed background can be found in appendix 4.
Table 1: Overview Participants
4.1 Initial Appraisal followed by Inertia- Coping & Habits
Even though the initial introduction was considered to be too early by some users, the majority did have a positive, yet abiding attitude towards the implementation. Years had gone by and the old systems became outdated, the need for a more effective way of working was recognized by most employees. Due to mergers and expanding activities of the organization, a multitude of systems entered the working life of the employees. Most of those systems were not connected in any kind of way, so in the end users had to learn to how operate up to 15 systems. The depth of these systems made it virtually impossible for a user to use the systems to their full potential. The news that a new all-encompassing system would be introduced that would integrate the essential components of the prior systems was welcomed by both management and the users. Many of the users had seen a lot of new programs over the years so they did have some reservations about whether UDT would be the final solution, but all agreed that if UDT would keep its promise, than it should be a huge improvement. And of course there are always people who embrace change right away. When one of the users, in this case Alice, passionately promoted the use of UDT while I was observing her, I asked her whether she had always been an advocate of the new system. She answered with a full heartedly: “Yes! Absolutely, and it’s only getting better.”
21 | P a g e But not everybody was as jubilant. John for instance was asked about his response to the introduction of UDT and he replied:
“Ik denk dat het wel een goede stap in de goede richting is. Maar ik denk wel dat ik daar meer de functionaliteit van [systeem a] meer terug in zou willen zien, voordat ik denk dat ik echt daar in overga.” / “I do think it’s a step in the right direction. But I think that I want to see more of the functionalities of [system a6] before I truly make a transfer”
As showcased in the quote above, the reason why some considered the implementation to be premature had to do with the limited set of features. The organization chose to release implementations in a gradual way instead of a revolutionary approach. This led to a situation where, at that particular point in time, the legacy systems were considered to be clearly superior to the new one.
Marc: “In het begin was [UDT] log. Het sloot ook niet aan bij mijn wensen. Er was gewoon te weinig informatie of dat het niet realtime was. Dus dan moest je toch in een ander programma de antwoorden gaan zoeken.” / “In the beginning [UDT] was unwieldy. It did not suit my requirements. There was just too little information, or it wasn’t real time. So you just had to go to a different system to look for answers.”
Alice: “In het begin was het wel zo dat je nog heel veel [van de benodigde informatie] niet kon vinden of er heel veel nog niet in stond. En dan moest je vaak nog terug naar [system a] en [system b7]” / “During the startup I couldn’t find much [of the needed information] or it just wasn’t there. And then you often had to switch back to [system a] and [system b].”
As a result the initial enthusiasm faded among the users. Only Alice was persistent among the observed participants and kept using UDT as her main system. Users lacked in-depth information and there were rumors that the information displayed in UDT was incorrect. Even now, almost 4 years after the introduction, these rumors are persistent:
Pete: “En op de een of andere manier, vertrouw ik [UDT] ook niet helemaal ofzo. Als ik het echt zeker moet weten, ga ik toch terug naar [systeem a] of systeem [c] om het te checken. Op de een of andere manier klopt het altijd wel, maar toch wil ik het nakijken in het echte systeem.”
/ “And in some way, I just dont trust [UDT] completely. If I have to be absolutely sure about something, than I always go back to [system a] or system [c]. Somehow it is always correct, but I still want to check it in the real system.”
6 System a is a registration system. All contacts with a customer are logged in this system as well as all sales registrations
7 System b is a back end system where all the customers’ product details are stored and is used to process all transactions and other changes regarding the customer and his/her products.
22 | P a g e Albert: “Soms klopt [UDT] niet helemaal [...] waardoor ik toch automatisch een ander, [system c]
erbij pak.” / “Sometimes [UDT] is incorrect [...] where upon I still automatically will use [system c]”.
The fact that the information displayed in UDT is retrieved directly from the legacy systems, and cannot be different, has been communicated multiple times. Over time, the functionalities were rapidly expanded and trust in UDT grew, but with the introduction of a new feature there are still mixed feelings about its trustworthiness.
The project leaders saw a steep increase in individual users and consider the implementation a huge success. However, the project leaders could only monitor how many times UDT was accessed on a daily basis, but not how the system is used. The depth and the way in which UDT is used might show a completely different story.
In reality, only one of the observed users, Alice, did not –unnecessarily- switch back to one of the legacy systems at all during the observations. Three years after the introduction, with UDT in place and the legacy systems still fully operational, most users did not use UDT at all or used it for an initial overview of the products that a customer had, but then switched back to the legacy systems when they had to answer the customer’s question. Everybody was free to use the legacy systems and no pressure in any form was exerted on switching to UDT. In practice this meant that UDT was consulted, but hardly used. Users showed inertia behavior, used the new system as little as possible and decided to continue using the systems as they had been using it for decades. Having the choice to stick with common practices was welcomed by the employees and many did so. John and Pete explained their reasons for their inertia as follows:
John: “Ik moet heel eerlijk zeggen, ik heb niet veel ervaring met [UDT]. Weinig tot niet. Ik vind het wel overzichtelijk. Dus voor een globaal overzicht, pak ik het er wel vaak bij. Maar qua functionaliteit [gebruik ik UDT] nog niet, of nauwelijks. En dat is voornamelijk omdat ik nog gewend ben dat oude [systeem a] en [systeem b] te gebruiken.” / “I have to be completely honest, I don’t have a lot of experience with UDT. Slim to none. I do think it is easy surveyable.
So for a global overview, I do use it. But when it comes to functionality I don’t or hardly [use UDT]. And that is mostly because I’m still used to using [system a] and [system b].”
Pete: “[Je] hebt nu dus [UDT], dan kun je loggen vanuit [UDT]. Nou dat doe ik gewoon niet omdat ik dat niet gewend ben”/ “Right now you got [UDT], and you can log contacts within [UDT].
Well I just will not do that, because I am not used to that.”
This first quote is a prime example of how many users treated UDT. They used it to get a global overview, but when push came to shove; they went back to the legacy systems to perform most of
