User Inclusion during ERP Implementations and its effects on Symbolic Adoption

(1)

Linköping University | Department of Management and Engineering (IEI)

Master Thesis, 30 hp | Industrial Engineering and management Digitization and Management |Strategy and Management Spring term 2021 | LIU-IEI-TEK-A--21/04161--SE

User Inclusion during ERP

Implementations and its effects on

Symbolic Adoption

Henrik Hörnlund (henho110)

Johannes Ålander(johal921)

17/06/2021 Examinator: Alf Westelius Supervisor: Emelie Havemo

(2)

Executive Summary

Having a modern Enterprise Resource Planning (ERP) system is seen as increasingly important if an organisation wants to compete in today’s business environment. The implementation of a new ERP system can result in added automatization and the replacement and/or coupling of legacy systems. The replacement and coupling of legacy systems can help an organisation increase data integrity and accessibility by lowering data fragmentation.

Previous research on the topic of ERP implementations has emphasized user participation as an important factor when aiming to succeed with an implementation. The research is however inconsistent with its use of participation and involvement where involvement is often used synonymously with participation, but sometimes it signifies that the implementation is of personal importance to the user. This study defines user participation as the user partaking in activities relating to the implementation, and user involvement is defined as the user perceiving the project or system as important. This study also introduces a new term: user inclusion, which encompasses both user participation, and if a user is on the receiving end of one-way communication, for example when receiving a newsletter.

This study argues that symbolic adoption, the belief that implementing the system is a good idea, is an important part of system success and uses a combination of factors from the UTAUT model and the I/S success model as antecedents to the dimensions of symbolic adoption that has been presented in previous research.

The purpose of this study is to examine how users are included in ERP implementation processes, and how their inclusion in the process affects users’ symbolic adoption. To address this purpose, a qualitative multi-case study that included two cases was conducted. In order to answer how users are included, this study divides how into three sub-questions; who; when and in what way users were included throughout the implementation.

This study concludes that users’ inclusion increases over time throughout the implementation process, both in terms of influence, and in number of participating users. This study also shows that users included in project initiatives can affect the antecedents to symbolic adoption in various ways, both their own symbolic adoption, and that of their non-participating colleagues. An example of this is how intended users’ expectations can be managed through communication and when they get first-hand experience of the system, their expectations correct themselves so that they are closer to the actual performance and effort required. The antecedents, performance and effort expectancy, are in turn linked to the symbolic adoption dimension effort worthiness.

(3)

Acknowledgements

Over the course of this project, we have had a lot of support, and would therefore like to acknowledge their help in producing this study.

Emelie Havemo, for providing us with direction and guidance. Without her this would have been a much longer journey, and the end result would not have been nearly as good.

Alf Westelius, for his insights at checkpoints. His feedback helped us set a clearer direction for this study by determining what was important and what we needed in order to be able to draw valid conclusions.

One Economy, who assisted us in the search for case companies that were relevant for this study. Without you this study would not have been possible, we are especially thankful to the interviewed consultant who helped us sort through possible case companies and provided us with context for Case 1.

Our opponents, Jakob Linderstam and William Bergmark, for a fruitful cooperation over the last five months. The meetings we have had together this spring has led to several interesting discussions that deepened our understanding and helped us write a report that had an interesting story.

Finally, we would also like to acknowledge our friends and family, who have put up with us during the last five months. This report has taken its toll on all of us, but with your support and understanding we were finally able to complete it.

This thesis marks the end of our time at Linköping University, a time that has been both challenging and rewarding. This thesis has helped us enhance our understanding of ERP implementations and will serve as a springboard when we now start focusing on our future careers. The experience of writing a thesis this extensive has been unlike any other. It has been both fun and interesting, but also exhausting, and now when we are at the end of it, we are glad that we pulled through.

(4)

4.1.1 Early implementation ... 58 4.1.2 Mid implementation ... 61 4.1.3 Late implementation ... 63 4.1.4 Post implementation... 66 4.2 Case two ... 68 4.2.1 Early implementation ... 69 4.2.2 Mid implementation ... 71 4.2.3 Late implementation ... 72 4.2.4 Post Implementation ... 74 5 Analysis ... 76

5.1 How are users included? ... 76

5.1.1 User selection ... 76

5.1.2 Timing and degree of user inclusion ... 78

5.1.3 Summary and answer to RQ1 ... 86

5.2 Effects of user inclusion on antecedents to symbolic adoption ... 87

5.2.1 Performance and Effort Expectancy ... 87

5.2.2 Social influence ... 91

5.2.3 System quality & System support ... 94

5.2.4 Summary and answer to RQ2 ... 98

6 Conclusions ... 99

6.1 Conclusions ... 99

6.2 Scientific Contribution ... 101

6.3 Future Research ... 103

(6)

List of Figures

Figure 1. Project Phase Model (Parr & Shanks, 2000) ... 16 Figure 2. The changing as three steps (CATS) model that was first introduced by

Lewin (1947) and later reworked many times. This version was presented by

Schein (2010). ... 17 Figure 3. DeLone & McLean's (2003) updated model for system success called the “Updated D&M IS Success Model”. Shows different categories of system success measures and their proposed interplay... 23 Figure 4. The Technology Acceptance Model (TAM). Showing proposed causal connections from Perceived Usefulness and Perceived Ease of Use to Attitude, Intent and Actual System Use... 24 Figure 5. Phase A (Klonglan & Coward, 1970)... 26 Figure 6. Phase B (Klonglan & Coward, 1970) ... 26 Figure 7. Karahanna and Agarwals (2006) modified TAM model connecting Symbolic

Adoption to intentions to explore. ... 27 Figure 8. Wang & Hsieh's (2006) model connecting symbolic adoption to extended and

emergent use. ... 28 Figure 9. Venkatesh, Morris, Davis, & Davis’ (2003) Unified Technology Acceptance of

Use and Technology model ... 31 Figure 10. Proposed antecedents to the four dimensions of symbolic adoption. ... 33 Figure 11. Project initiatives that include users will be considered if and only if they happen in the early, mid or late implementation phase. The post implementation phase is interesting only as a final evaluation. ... 38 Figure 12. The analytical model used in this study, showing how timed project initiatives

could lead to symbolic adoption. ... 39 Figure 13. Who; in what way and when are users included in initiatives as three parts of the first research question ... 40 Figure 14. Example of how initiatives fit into the Early, Mid and Late phases. ... 42 Figure 15. This figure illustrates the research process used for this study. ... 46 Figure 16. Length of the implementation in both cases, with approximate length of each phase. 56

(7)

List of tables

Table 1. Critical Success Factors connected to user participation from different studies. CSFs

are colour coded to differentiate between different types of CSFs ... 8

Table 2. Our Version of Adopter Categories, inspired by Rogers (2003) ... 13

Table 3. User Inclusion dimensions inspired by Mumford et al. (Mumford, 1979; Maas, van Fenema, & Soeters, 2016; Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011; ... Ives & Olson, 1984) ... 16

Table 4. Summary of the three steps of CATS and what types of initiatives can have an effect on a particular step. ... 21

Table 5. Table depicting the outcomes of different combinations of Symbolic Adoption and Actual Adoption. If symbolic and actual adoption or rejection matches, all is well, but when there is a mismatch, innovation dissonance occur. ... 26

Table 6. Table comparing different iterations and developments of TAM. The table shows similarities and differences of the different iterations in terms of Antecedent Variables, Main Variables, Context Variables/Moderators and Dependant Variables/Outcome Variables ... 30

Table 7. Table showing how the phases from Parr & Shank’s (2000) PPM model correspond to the proposed phases that are going to be used in this study. ... 37

Table 8. Codification, Degrees of inclusion ... 41

Table 9. Criteria for defining Early, Mid and Late initiatives ... 41

Table 10. Codification criteria for SA antecedents ... 43

Table 11. Interviews with employees at the case companies... 50

Table 12. Interviews with the consultants at UCS One Economy ... 51

Table 13. Systems to be implemented Case 1... 58

Table 14. Interviewee Descriptions, Case 1 ... 58

Table 15. Organisational documents, Case 1 ... 58

Table 16. Summarized description of initiatives that include users, Case 1, Early phase ... 61

Table 17. Summarized description of initiatives that include users, Case 1, Mid phase ... 63

Table 18. Summarized description of initiatives that include users, Case 1, Late phase ... 66

Table 19. Interviewee descriptions, Case 2. ... 69

Table 20. Summarized description of initiatives that include users, Case 2, Early phase ... 71

Table 21. Summarized description of initiatives that include users, Case 2, Mid phase ... 72

Table 22. Summarized description of initiatives that include users, Case 2, Late phase ... 74

Table 23. User selection Criteria ... 77

Table 24. Inclusion degree of users in project initiatives in Case 1. ... 79

Table 25. Inclusion degree of users in project initiatives in Case 2. ... 80

Table 26. Initiatives affecting Performance & Effort Expectancy ... 89

Table 27. Initiatives affecting Social influence. ... 92

(8)

1 Introduction

The introduction consists of 4 sub-chapters. The first sub chapter gives some background explaining what ERP systems are, why the implementation of ERP systems is an interesting research area, how user participation can be a way to achieve a successful implementation, and why symbolic adoption can be an important part of said implementation success. The second sub-chapter defines the purpose of this study, at the same time introducing the term user inclusion, which is a combination of user participation and users being on the receiving end of communication. The third sub-chapter presents how this study will be conducted in order to address its purpose. The fourth sub-chapter presents a brief summary of what this study contributes to current research.

1.1 Background

An increasing number of organisations are coming to the conclusion that they need to have a modern ERP (Enterprise Resource Planning) system in order to keep up with their competitors (Panorama Consulting Group, 2020). An ERP system is a packaged, often modular business software that enables efficient and effective management of resources such as finances, materials and human resources (Nah, Lau, & Kuang, 2001). It is also possible to integrate separate software packages, and in doing so forming a complete solution for an organisations’ information processing needs (Nah, Lau, & Kuang, 2001). The benefits of implementing an ERP system varies, but there are some common benefits that are generally associated with ERP systems (Beheshti, 2006). In general, most ERP implementations replace several legacy systems where fragmented and/or redundant data has been stored in several different systems in order to comply with format or access needs (Beheshti, 2006). By reducing the redundancy of these systems, it is possible to achieve benefits such as higher data integrity and better accessibility. It should also be noted that one of the primary values often associated with the digitization of an organisation is the prospect of automatization this is often true for ERP systems and can be the antecedent that promotes the development of new processes. A new ERP system can therefore lead to higher efficiency and effectiveness, which makes the implementation of ERP systems an attractive target for practitioners to implement and researchers to study.

Due to the technological trend and the potential benefits related to investing in ERP systems it is not surprising that it has been heavily researched over the last decades. Several of these researchers, who have proposed, what they deem to be, Critical Success Factors (CSFs) when conducting ERP implementations (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Aldwani, 2001; Bano & Zowghi, 2013; Kiran & Reddy, 2019). It is debatable whether these CSFs are in fact critical, since most researchers have arrived at different sets of factors, but they can be treated as indications of what might have an influence on the success of ERP implementations. These factors include but are not limited to: creating a balanced team, having effective communication, and having top management support. Another CSF, which is mentioned in management literature and information system (IS) literature alike, is user involvement (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Aldwani, 2001; Bano & Zowghi, 2013; Kiran & Reddy, 2019). The idea that user involvement is a critical factor for system success has prevailed for over half a century (Barki & Hartwick, 1989; Bano & Zowghi, 2013; Markus & Mao, 2004; Swanson, 1974). In a literature review on the topic of user involvement and its effect on system success, Bano and Zowghi (2013) found that about 60 % of studies reported that user

(9)

involvement contributed positively to system success, 23 % were uncertain and 14 % found a negative impact. These results seem to support the idea of a positive relationship between user involvement and system success. The results do however also seem to show that not all kinds of user involvement are necessarily beneficial, or that the context in which this involvement takes place matters. The variance could also in large part be attributed to inconsistencies in the use of the terms user involvement and system success (Bano & Zowghi, 2013). Before any further evaluation of findings connecting user involvement to system success, it is therefore necessary to establish what we mean when referring to these concepts. With user involvement in particular, there are inconsistencies in what is meant by both user and involvement, as well as when and how this involvement takes place, (Bano & Zowghi, 2013).

User involvement and user participation

When it comes to research in the Information systems (IS) field, user involvement is often used synonymously with user participation. In an attempt to separate the two, Barki and Hartwick (1989) defines user participation as enacting a set of behaviours or activities during the implementation process, whereas user involvement is defined as a subjective psychological state where the system is seen as having importance and personal relevance to the user. With this definition, user involvement could come as a consequence of user participation and act as a mediator between user participation and different measures of system success (Barki & Hartwick, 1989). It is also possible for a user to be involved without participating (Barki & Hartwick, 1989). An example of this could be a user who wants to participate but is not allowed to attend meetings or software testing. Unless stated otherwise, the terms user involvement and user participation will henceforth be used in accordance with the definition above, irrespective of which term is used in cited sources.

User participation is theorized to have a number of benefits that might lead to system success, such as: project buy-in, improvement of system quality, and the emergence of relationships between users and developers (Markus & Mao, 2004). Another benefit of user participation could be avoiding user resistance by addressing some of the common reasons for resistance, such as: uncertainty about the future, user input not being considered, users losing control and/or power and lacking communication (Blanton & Klaus, 2010). Users can however participate in different ways with varying degrees of influence (Ives & Olson, 1984), which is likely to affect the effects of participation. User participation can also have unintended and unwanted side-effects, depending on context. The timing of user participation could for example be an important consideration (Wagner & Piccoli, 2007). Although it is a commonly held belief that user participation early on in implementation projects is beneficial, it could be more damaging than helpful (Wagner & Piccoli, 2007). Busy users might have trouble engaging in the project early in the implementation, when it will not have an effect on their daily work processes for the foreseeable future (Wagner & Piccoli, 2007). The adverse effects of this were demonstrated in a case study by Boonstra & Govers (2009). In this case, a lack of feedback from disengaged users was interpreted by managers as approval, which led to lacking system quality on launch and irritated managers who felt that issues should have been brought up earlier (Boonstra & Govers, 2009). This could be interpreted as users that did not feel involved, although they did participate. Other considerations for user participation are: which users to pick (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011), and how to include them (Mumford, 1979; Ives & Olson, 1984).

(10)

If user participation is usually, but not always helpful (Bano & Zowghi, 2013), then a more interesting question than should users participate? might therefore be How, when and under what

circumstances should users participate?

Implementation success and symbolic adoption

There are many ways to defining implementation success. While keeping the implementation within budget and on schedule can be of great interest, a more important consideration is being able to realize the expected performance gains (Kiran & Reddy, 2019). Many scholars choose some variation of system use or user acceptance (often measured by intended or actual use) as a proxy for system success (Bokhari, 2005), which could be reasonable since system use is necessary in order to gain any benefits at all from a system (Askenäs & Westelius, 2000). Possibly gaining

any benefits at all does however seem like a low bar. DeLone & McLean (1992) created the so

called I/S Success model, which was later updated to form the Updated D&M IS Success Model (DeLone & McLean, 2003). This model features seven categories of popular system success measures, as well as how they are connected. These seven categories are: System Quality,

Information Quality, Service quality, Intentions to Use, Use, User Satisfaction, and net benefits.

The Technology Acceptance Model (TAM) is a very popular model for measuring user acceptance. TAM fits reasonably well with some of the constructs from the Updated D&M IS Success Model, namely System Quality, Information Quality, Intentions to Use and Use, but it is lacking in terms of User Satisfaction and Net Benefits. This is especially problematic in mandatory settings, as opposed to voluntary settings where only those that have accepted a certain technology will use it. A solution could be to add a good measurement for User Satisfaction, which, in conjunction with Use, could lead to both Individual and Organisational Impact. Symbolic adoption could be such an addition, which has been attempted by Karahanna & Agarwal (2006) and Wang & Hsieh (2006). In these studies, symbolic adoption was described as a heightened mental state where the user views a concept or technology as worthwhile and was measured on an individual level as a multidimensional construct consisting of heightened enthusiasm, use commitment,

mental acceptance and effort worthiness. The findings from these studies support the idea that

symbolic adoption leads to extra-role behaviours post implementation, in terms of using more functionality or using existing functionality in novel ways. This extra role behaviour is the kind of behaviour that can lead to the net benefits depicted in the Updated D&M IS Success Model. Symbolic adoption could also be useful during the implementation, since the opposite, symbolic rejection, might turn into passive or active resistance from the would-be users (Heidenreich & Talke, 2020). Achieving Symbolic adoption during the implementation, and carrying it over into the post-implementation phase could therefore be an important part of a successful implementation. The connection between user participation and symbolic adoption can in part be explained by the possibility of user participation creating user involvement. User involvement, defined as a feeling that a system is of personal importance and relevance, does however not necessarily mean that users are positive. Highly involved users are likely to develop strong positive or negative attitudes towards a system (Barki & Hartwick, 1989), so managers would do well to involve users with care.

(11)

User participation is however not the only way to achieve involvement or system success. Communication is commonly identified as a CSF (Aldwani, 2001; Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Selvin & Pinto, 1986) and is a very important part of change management (Kotter, 1995; Schein, 2010). Communication can for example build expectations (Kotter, 1995; Schein, 2010), which is a big part of technology acceptance research (Davis & Bagozzi, 1989; Venkatesh & Davis, 2000; Venkatesh, Morris, Davis, & Davis, 2003; Venkatesh, Thong, & Xu, 2012), an is as such also expected to affect symbolic adoption. The term user

inclusion will therefore be introduced, which we define as: users participating in activities, or being

recipients of communication.

1.2 Purpose

The purpose of this study is to examine how users are included in ERP implementation processes, and how their inclusion in the process affects users’ symbolic adoption. In order to conclusively formulate research questions that reflect this correctly, another term is defined: project initiatives. We define project initiatives as: initiatives related to the project that are instigated by members of the project group or decision makers at a higher level in the organisation. With this definition, the following research questions are proposed:

1. How are users included in the implementation process through project initiatives? 2. How does user inclusion in project initiatives affect users’ symbolic adoption?

This study consists of two research questions where the first one is defined by three questions; Who is included; In what way are they included; When are they included. This is followed by the second questions that focuses on how user inclusion in project initiatives affects users’ symbolic adoption.

1.3 Course of action

To answer these questions, this report includes two case studies from different companies, that have recently gone through an ERP implementation process. One of the companies was sourced with the help of the accounting firm UCS One Economy, provider of ERP implementation services. The other case was sourced through one of the researchers personal contacts. The case companies were selected pragmatically, with the criterion that users participated during the implementation process. The empirical data was gathered through interviews with; an implementation consultant from the accounting firm, giving an outsider’s perspective on the process for one of the cases; decision makers; users and project managers from the case companies. Both cases had external consultants who helped with their implementations, this approach poses some limitations on the study. First and foremost, it limits the results to only reflect companies that have used external help. Secondly, the hired consultants may have been significantly better or worse than other consultants that provide the same service. We do however believe that regardless of the skills and aptitude of the consultants, the gathered data should enable us to discern patterns in how users are included in the implementation process since the consultants originated from organisations that have helped countless organisations implement ERP systems. Since the consultants originate from organisations that focus on ERP implementations it is likely that their recommendations, on how to implement, is standardized to some degree.

(12)

The analysis was conducted using an analytical framework, that draws on literature regarding ERP implementation in general, user participation and user involvement in ERP implementations, communicative boundaries, symbolic adoption and the use of management controls systems. Based on this framework appropriate interview questions were formed. The interviews followed a semi-structured approach with open ended questions in order to provide rich elaborate data where the interviewees could fully express their experiences while still following a structure aimed at gathering relevant data. This structure was also applied in order to legitimize the study by lowering the risk of the interviewers’ own biases affecting the responses through confirmation bias (Powell, Hughes-Scholes, & Sharman, 2012). Some of the questions had to be more direct in order to provide a foundation for comparison between the different cases. The gathered empirical data was analysed through a thematic analysis using themes that was inspired by the theoretical concepts in the literature review. The themes were then studied and discussed by the researchers in order to draw final conclusions that served to answer the purpose of this study.

1.4 Contribution

The primary contribution of this study is the expansion of knowledge on how different approaches to user inclusion affects symbolic adoption, and in doing so affecting system success. What will make the contribution especially useful are the distinctions between degrees and of timing of user participation, which is called for by authors such as Bano & Zowghi (2013). This study is also an attempt to further solidify the distinction between user involvement and user participation, which has been called for, for a long time (Barki & Hartwick, 1989), but is still very much lacking (Bano & Zowghi, 2013).

This study also contributes with new antecedents to symbolic adoption that can help explain how users symbolically adopt systems. Previous research by Karahanna & Agarwal (2006) and Wang & Hsieh (2006) have previously completed quantitative studies where they have analysed symbolic adoption as a part of system success. This study takes a qualitative approach where we try and analyse how different initiatives, where users are included, affect users’ symbolic adoption through the proposed antecedents in chapter 2.7.3.

(13)

2 Literature Review

In the following chapter the literature review findings are presented. The findings consist of previously published literature regarding both empirical and theoretical findings. These findings are then synthetized into an analytical model that is used to answer the research questions.

The first sub-chapter establishes the differences and relation between participation and involvement. The second sub-chapter focuses on previous research regarding the user participation as an important factor for the success of a system. The third chapter is used to explain the benefits and risks related to user participation. The fourth chapter presents a framework for how users can be part of project initiatives. The fifth chapter presents a framework for the phases an ERP implementation goes through, which will be used in the analytical framework to generate a model for defining the timing of initiatives. The sixth chapter focuses on change management and the Changing as Three Steps model as a basis for understanding the process of change. The seventh sub-chapter aims to connect previous literature focused on systems success and user acceptance with the concept of symbolic adoption in order to extend the symbolic adoption concept with potential antecedents. The eight chapter analytical framework is used to synthesise previous chapters into a conceptual model that can be used to explain user inclusion during the ERP implementation and how it leads to the proposed antecedents of symbolic adoption.

2.1 User participation and user involvement in an ERP context

There has been a lot of confusion surrounding the term user involvement, which motivates that this is cleared up right away. User involvement has been inconsistently used as a synonym for both participation and engagement (Bano & Zowghi, 2013). These inconsistencies are especially noticeable for IS (Information system) research (Barki & Hartwick, 1989). Other research areas such as marketing, psychology and organizational behaviour have all converged on conceptualisations of user involvement that are used more consistently than in IS research and are very similar to one-another (Barki & Hartwick, 1989). In the name of conceptual consistency across research disciplines, we will therefore follow Barki & Hartwick’s (1989) suggestion that user involvement should be used to refer to a psychological state of an individual and defined as the importance and personal relevance attached to a particular system. It is also possible for a user to place personal importance and relevance onto activities surrounding development or implementation of a system, making the user involved with the implementation process rather than the system (Barki & Hartwick, 1989). User participation, will on the other hand be used to refer to any behaviour and activity that the user performs in the implementation process, as suggested by (Barki & Hartwick, 1989).

With user involvement and user participation differentiated, it is interesting to examine how they are related. User participation will be assumed to be an important antecedent to user involvement. This means that a user could participate without feeling involved at first but their participation could make them involved, i.e. starting to feel that the system is important and personally relevant, without being allowed to participate in its implementation. Barki & Hartwick (1989) proposes a number of predictions based on previous research on user involvement and user participation. (1) Highly involved users are more likely to have strong feelings about a system, positive or negative. (2) Highly involved users are more influenced by strong factual and logical arguments, whereas (3) less involved users are more easily swayed by normative influences. These less involved users might care less about the facts, placing a higher importance on factors such as top management

(14)

support and the attitudes of their peers (Barki & Hartwick, 1989). Another consideration, pointed out by Wagner & Piccoli (2007), is that it might be important to actively involve users that are participating. While user participation is theorized to be an important antecedent to user involvement, involvement is not guaranteed (Wagner & Piccoli, 2007; Barki & Hartwick, 1989). Having users that are not emotionally involved participate in the implementation might have some adverse effects if the participation alone is not enough to make them involved. This is exceedingly likely to happen when users are involved early on in the implementation process, and their daily work processes will not be affected for a long time (Wagner & Piccoli, 2007). An example of this is the previously mentioned case study by Boonstra & Govers (2009), where some users did participate, but without being fully involved in the process. The result was close to catastrophic as the users refused to use the system on launch since it lacked critical functionality, and the managers where disappointed that the participating users had not brought up the issues earlier (Boonstra & Govers, 2009). If user participation in the early stages of the implementation process is still deemed as necessary or preferable, then it should be warranted to take special care that they are actually involved (Wagner & Piccoli, 2007). A simple way around this issue could be to make participation voluntary (Barki & Hartwick, 1989), or picking among users that have announced an interest (Bano & Zowghi, 2013). Other than involvement, selecting the right users for participation includes other considerations such as; individual skills; ability to represent a cross-section of the users and organizational boundaries which will be discussed further in chapter 2.4.

2.2 User participation to achieve ERP implementation success

Enterprise Resource Planning systems, or ERP systems for short, are packages of software aiding companies to track and manage their resources, such as finances, materials, and human resources (Nah, Lau, & Kuang, 2001). ERP systems are often modular, where the different modules are integrated to allow for information to flow effortlessly through the organisation (Beheshti, 2006). A new ERP system often replaces several legacy systems, leading to better data access and a reduction in fragmentation and redundancy in data, are some of the most important benefits of ERP systems (Beheshti, 2006). The potential for gains in efficiency and effectiveness from implementing a new ERP system have spurred a lot of research on the subject. Several of these researchers have ended up with different sets of what they deem to be Critical Success Factors (CSF). It is, as mentioned in the introduction, debatable whether these CSFs are in fact critical, especially since they do not come up with the same set of CSFs. User participation (often referred to as user involvement) is one such CSF, which is mentioned in the literature from both the management and IS literature (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Aldwani, 2001; Bano & Zowghi, 2013; Kiran & Reddy, 2019). Several of the other suggested CSFs can be conceived as interconnected and/or influencing the effectiveness of user participation. Some are directly linked to how user participation is conducted, such as picking a balanced project team (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Parr & Shanks, 2000; Shanks, Parr, Hu, Corbitt, & Thanankit, 2000), or providing training (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Aldwani, 2001), while others are indirectly linked by providing the right kind of conditions for effective user participation, such as promoting the right kind of culture (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016) or communicating effectively (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016; Aldwani, 2001; Selvin & Pinto, 1986). A compilation of CSFs related to user participation that might have a direct or indirect effect on the success of an ERP implementation can be found in Table 1.

(15)

Critical success factors (Altamonoy, Tarhini, Al-Salti, Gharaibeh, & Elyas, 2016) (Aldwani, 2001)

(Selvin & Pinto, 1986) (Parr & Shanks, 2000) (Shanks, Parr, Hu, Corbitt, & Thanankit, 2000) User participation x x Balanced team Team effectiveness Personnel x x Champion x

Best people full time x x

End-user training x x Eternal expertise/consultant support x Minimizing adoption effort x x Change management x Project leadership x x Communication End-user communication Communicating ERP Benefits, Communication ERP general operations Client consultation, Client acceptance, Communication, project mission Culture x Securing support of opinion leaders x Commitment to change x Top management support x x x Organizational Leadership x Empowered decision makers x

Colour coding: Selection of user for participation Training &

Support Change management

Table 1. Critical Success Factors connected to user participation from different studies. CSFs are colour coded to differentiate between different types of CSFs

The CSFs in Table 1 are colour-coded in an attempt to categorise them into themes. It could be argued that a specific CSF could fit into more than one theme, and that they could be divided into additional themes, but the identified themes will only serve to inspire the direction of this study, not define it. The first identified theme is the selection of users for participation, suggesting that

(16)

who is participating is important. Different selection criteria for users will be discussed in chapter

2.4, user inclusion dimensions. Training and support is the second identified theme, which will be discussed as an antecedent to symbolic adoption in chapter 2.7.2. Training and support can also be considered to belong in the theme change management which will be discussed at length in chapter 2.6, change management. A note that should be made regarding Table 1 is that several of the studies found different CSFs. We interpret this as the result of using the CSF concept when looking at users in ERP implementations from several different perspectives. Therefore, we deem that the synthesis of what the studies found can represent a summary of the important factors that include users during ERP implementations.

The evidence showing that not all user participation leads to system success (Bano & Zowghi, 2013; Wagner & Piccoli, 2007), motivated the research questions for this study, of who, how and when users participate in order to promote symbolic adoption. What is meant by symbolic adoption and how it relates to system success will be discussed in chapter 2.7, whereas the who, how and

when will be discussed in chapters 2.3-2.6.

2.3 Potential, Challenges and Risk related to User Participation

Since previous research in general have been positive to user participation in the implementation, it is interesting to figure out why. Markus & Mao (2004) presents three possible explanations to why user participation is theorised to be linked to positive outcomes in the following ways:

• Project buy-in

User participation leads to users feeling as though they are a part of the development and implementation and become committed to using the system. Peoples attitude towards change is positively influenced by participation in the implementation process.

• Improvement of system quality

User participation leads to a higher system quality since the project gains added knowledge in the affected processes and what the users actually prioritize. This in turn leads to a system that can support its users in a better way.

• Emergence of relationships among users and developers

This explanation differs from the above mentioned explanations in that it requires direct cooperation between users and developers while the explanations above talks of user participation in general. The process of developers working directly with the intended users creates a natural environment for users to share their concerns directly with the developers and by sharing it can be possible to implement their ideas into the software. We believe that this relationship is also something that creates a level of trust that may make a participant more receptive to the system.

However, it should be noted that these three explanations all suffer from the fact that they mainly affect the users who are participating in the implementation (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011), it is unlikely that a non-participating user will achieve these benefits to their full extent if their colleagues participate. This due to factors such as the fact that non-participating users do not acquire a direct relationship to the developers and the improvement to the system will be based on the participating users’ needs and ideas which are not necessarily identical to that of a non-participating one.

(17)

Rogers (2003) who researched how ideas spread and the different types of innovation adopters argues that the diffusion of ideas is dependent on four factors: communication channels, a social system, the innovation itself and time. A part of the social system is the opinion leaders (Rogers, 2003). Not all users have the same influence over the behaviour of their fellow colleagues, some are simply more respected than others. Opinion leaders are actors who have great influence over their colleagues and are generally held in high regard (Rogers, 2003). We believe that if the participating users are selected in such a way that they are also opinions leaders, then their experiences and opinions will disperse throughout the organisation and influence their non-participating colleagues. It is therefore not only important to value user participation based on how it affects all intended users, but also how it affects the participating user directly, since there are likely to be indirect benefits that relate to the user group as a whole.

User participation is not a silver bullet (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011). Clement and Van der Bessela (1993) found in their study, where ten projects were reviewed, that user participation does not necessarily lead to users being more positive to the technology. For example, we believe that having users participate in the implementation gives them new insight into the project and with it, it also adjusts their expectations. Users that are not participating in the project, are mainly exposed to formal communication, formal communication which is easier to regulate and control. When they participate, informal communication is unavoidable since upcoming challenges in the project needs to be discussed internally within the project group. Users form their expectations based on what is communicated to them through both formal and informal channels, which is apparent in the case studied by Bano, Zowghi and Rimini (2015), where the users experienced that the promises regarding their participation and the software capability were not realized in the end. There were no formal documents to strengthen that these promises were made, which lead to the end-users getting a feeling of dissatisfaction (Zowghi, Bano, & da Rimini, 2015). This type of failure to realize user expectations can result less extra-role behaviour, lower job satisfaction, lower employee trust (Coyle-Shapiro, 2002) and overall negative attitudes (Porter, Pearce, Tripoli, & Lewis, 1998). Blanton found several determinants in ERP implementations that lead to resistance from users: uncertainty about the future, user input not being considered, users losing control/power and lacking communication (Blanton & Klaus, 2010). It becomes apparent that many of these relate to how decisionmakers’ decisions regarding user participation and communication during the implementation. Looking at user resistance and the effects that contribute towards it becomes relevant since it is likely that users who resist change do not think that the change is a good idea and therefore have not symbolically adopted the concept.

2.4 User Inclusion Dimensions

As mentioned in the previous chapter, communication and user participation can play a large part in convincing users that a project is worthwhile, as well as actually improving the system itself. Being the recipient of communication is not part of the definition of user participation, which entails taking part in activities, but is recognized as a CSF by several researchers (Table 1) and can be used to induce user involvement which might make user participation more effective (see chapter 2.1). The term user inclusion will therefore be used when discussing project initiatives where users either participate or are recipients of communication. There are many ways to include users in the project, and since it can influence the project both positively and negatively, it is important to have a well-thought-out strategy for how to appropriately select users and in what capacity to utilize them. In this chapter we will discuss different ways of selecting users, as well

(18)

as to what degree they can be included, which will end up in four user inclusion dimensions; approach to user participation; type of user; degree of user inclusion; user selection criteria. The idea of including users in the implementation might seem simple on a surface level, but there are many ways to do it. For starters, Mumford (1979) proposed that there are three ways of approaching user participation:

• Consensus

An attempt is made to involve all users in the project group and make them participate. They are not necessarily decision makers with the power to force change, but they can participate through consultation and communication. This is a more direct approach to user participation where users do not have to voice their opinions through proxies.

• Consultative

The project group makes the decisions but are influenced by needs and wishes from the users. A majority of the users are not consulted in the process and there is no clear structure for who should be consulted.

• Representative

All user groups have representatives in the project group that are responsible for voicing their groups concerns and needs. This is an indirect approach where users’ needs and opinions are communicated through proxies.

The approaches vary from direct, where all affected end-users participate, to indirect, where employee representatives act as bridges between the project team and the users (Ives & Olson, 1984). Both indirect and direct approaches give the users a chance to voice their opinion and affect the outcome of the development (Ives & Olson, 1984), as long as their degree of user participation is not purely symbolical. Deciding to include users following Mumford’s approaches for consultative and representative participation divides the user pool into two categories; those who are include and those who are not. This type of division of users is also used by Wu & Wang who argues that there are two main types of users, key-users and end-users. End-users are all users who are intended as users for the finished project. Key-users are users who are utilized in the implementation as experts on business processes, advisors, educators, trainers or change agents and play an important role in connecting experts and dispersing knowledge throughout the organisation (Wu & Wang, 2007). Our definition of key-users is that they are end-users who have been selected to be a part of the project group and are utilized according to Wu & Wang’s (2007) definition of key-user tasks. We would also argue that in the case of participation in line with Mumford’s consultative approach there are no key-users according to our definition since they are not officially part of the project group.

This approach of dividing the user pool into end-users and key-users, will be used in the analytical framework in conjunction with a modified version of Ives and Olson’s (1984) degrees of user participation in order to give a perspective on how different project initiatives include users and how influential the users were.

Ives and Olson (1984, p. 590) defined the degree of user participation as “the amount of influence the user has over the final product”. Ives and Olson (1984, p. 590) extended Mumford’s three

(19)

approaches with six degrees that describe settings where user influence range from user input being completely disregarded to users acting as gatekeepers or product owners.

• No participation

“Users are not willing or not invited to participate” • Symbolic participation

“User input is requested but ignored” • Participation by advice

“User advice is solicited through interviews or questionnaires” • Participation by weak control

“Users have sign-off responsibilities at each stage” • Participation by doing

“A user as a design team member, or as the official liaison with the implementation group” • Participation by strong control

“Users may pay directly for the development out of their own budget, or the users’ overall organizational performance evaluation is dependent on the outcome of the development effort. “

The division of users and deciding on user influence creates a new challenge of deciding on how to select appropriate key-users that can represent a cross-section of their user group, while at the same time contribute with necessary knowledge and a contact network that can potentially provide more effective communication throughout the organisation. Bano & Zowghis (Bano & Zowghi, 2013) research supports this theory as they found that for achieving effective user participation it is very important to select the right people. There have been several different strategies for selecting the right participants. Rasmussen & Christensen et al. (2011) suggests that the key-users should be selected based on 13 criteria where half of the criteria are focused on the individuals’ abilities while the rest are focused on how well they represent the cross-section of all users. This is done in order to balance the key-users role as user advocates, which serves to supply a users’ perspective on the system, against their role as system champions, a role that encourages its organizational implementation (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011). This differs from other authors that emphasize the value of selecting a representative cross-section of the users (Mumford, 1983; Damodaran, 1996; Wilson, Bekker, Johnson, & Johnson, 1997). There are several different dimensions that can be used when devising a way of identifying a cross-section of users, a few examples on these are (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011); Age group; work experience; user group and adopter category. Age group and work experience are fairly self explanatory concepts but user group and adopter category requires further explanation. User group refers to selecting users from different departments or with different user needs. For example, an administrator of an ERP system might use the system differently compared to a business analyst. When Rasmussen et al. (2011) refer to adopter category they talk of Rogers (2003) adopter categories that distinguish peoples’ tendencies towards adopting new technology, by separating them into five different categories; innovators; early adopts; early majority; late majority and laggards. In this report we will use a similar definition of adopter categories that is a simplified but inspired version of Rogers (2003) categories. This simplification is used since it makes the codification of empirical data easier while still keeping the relevance of peoples’

(20)

differences in adopting new technology. Our inspired model with three types of adopters and the criteria that defines each one is presented below in Table 2.

Adopter Category Criteria

Early Adopter • Identifies as eager to adopt new

technology.

• Expected by others to be eager to adopt new technology.

Average Adopter • Does not identify as either an early or

late adopter.

• Expected by others to follow the general group when adopting new technology.

Late Adopter • Identifies as resistant to adopting new

technology.

• Expected by others to be resistant to new technology.

• Identifies as slow to adopt new technology.

• Expected by others to be slow to adopt new technology

Table 2. Our Version of Adopter Categories, inspired by Rogers (2003)

Rasmussen & Christensen et al (2011) hypothesize that the importance of selecting a cross-section of the user group is higher when selecting a large number of users that will participate for a brief time, for example surveys or brainstorming activities, but when the selection is small and for an extended amount of time, the individuals’ abilities are more important since it directly affects their possibility to contribute to the progress of the project. They also gave examples of what they deem to be potential criteria that display relevant individual abilities (Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011):

Knowledge of the work domain. • Knowledge about IT.

• Interest in the project and a desire to contribute to it.

• Ability to empathize with others and understand their needs.

• Ability to pass knowledge and enthusiasm about the system to other users. • Readiness to work with technical issues.

In addition to this we have included organisational boundaries as potential user selection criteria since according to Maas et al. (2016) key-users can play many roles in an implementation process, but their key focus lies in the dispersion of knowledge (Maas, van Fenema, & Soeters, 2016). Previous research that looked at what hinders the dispersion of knowledge, have stressed the role of organizational boundaries (Farjoun, Ansell, & Boin, 2015). The boundaries can appear in organizational charts and similar tangible objects but behind the tangible boundaries there are deeper, hidden boundaries that that appear in the form of cognitive and interest differences between groups on opposite sides of a boundary (Maas, van Fenema, & Soeters, 2016). Following this,

(21)

organizational boundaries are defined as a composite object holding three dimensions Structural, Social and cognitive (Maas, van Fenema, & Soeters, 2016).

Structural boundaries refer not only the organizational infrastructure but also the formal and

informal regulations that shape the interactions within the organization and the environment when executing tasks and utilizing resources (Maas, van Fenema, & Soeters, 2016). The concept of structural boundaries also includes the access to official communication pathways that connect employees throughout the organisation. Structural boundaries create a sense of structure, stability and predictability within the organisation, while at the same time hindering the flow of information between employees through the creation of geographical, functional and responsibility differences within the organization (Sturdy, Clark, Fincham, & Handley, 2009). This concept is in some literature called physical boundaries (Pan & Mao, 2016), but since organisations are acting in a digital age where a lot of employee communication is accomplished digitally the term structural boundaries will hereafter be used since it in a clearer way encompasses, not only holds physical geography but also digital geography.

Social boundaries refer to the differences in employees’ identity and interests, which are largely

influenced by the social bonding between employees. The boundaries arise from individuals identifying with some subgroups and rejecting other subgroups. The social boundaries are likely to be reflected in the social dynamics regarding loyalty, norms, trust and identity within the groups. The members of the social groups tend to evaluate the knowledge of each other more favourably compared to that of others outside the group (Melissa C. Thomas-Hunt, 2003), and knowledge originating from outside the group may be denigrated or even blocked due to “not invented here” syndrome (Katz & Allen, 1982). This is however problematized by Menon & Pfeffer (2003) who argues that internal knowledge is accessible and therefore assessable for flaws and do not bring the same value of uniqueness and possibly legitimacy as external knowledge. It is likely that we have preconceptions about our close employees’ skill sets and therefore may not value their input as high as we would value the input from external sources we deem knowledgeable.

Cognitive boundaries refer to differences in ideas, beliefs and understandings that guide

organizational actions and resembles the concept cognitive dissonance. In simplified terms it is the differences in how we perceive the world. It encompasses differences in how we view problems, opportunities and how we communicate. For example, hired consultants may use a technical language in communication that users do not understand which may result in misunderstandings. If the cognitive distance between two groups is large then it can give rise to barriers of communication and shared understanding, which may hinder the flow of information (Sturdy, Clark, Fincham, & Handley, 2009).

ERP implementations are projects that often affects multiple departments and requires that knowledge regarding business processes and activities to flow between stakeholders in order to be successful (Chan & Rosemann, 2001). Thus, we believe that when selecting a key-user that is supposed to act as a boundary spanner it is important that he/she shares these boundaries with the group that he/she represents, and preferably the project team as well. It is also important to acknowledge that by having users join the project team for an extended period of time, it is possible that they will start to feel alienated and start identifying with the project team instead of their original department (Westelius, 2006; Symon, 1998). This state of alienation hinders their abilities to effectively anchor the project within their department of origin and to represent the needs of

(22)

their department of origin when compromising with the project team (Westelius, 2006). We believe that this is due to a change in the key-users organisational boundaries where they over time drift further and further away from their department of origin.

Thus, we can now identify that the user participation can be understood as a combination of the approach to user participation, the degree of user participation, and the type of users affected. Organizational boundaries are added as User selection criteria since previous research show that key-users are often used to disperse knowledge and overcome organizational boundaries. Seeing as the initiatives being studied in this thesis are ones where users are included, we have removed

No participation and replaced it with recipients of communication, which is defined as users taking

part as recipients in one-way communication related to the implementation. We have also relabelled it as Degrees of user inclusion which will be used henceforth. The degree of user inclusion and user selection criteria will be added to our analytical framework since they in a comprehensive way summarizing users’ participation in a project initiative. The other dimensions will be used as supporting theory in order to analyse the initiatives more in detail, see Table 3.

Dimension Aspect

Approach to user participation • Consultative

• Representative • Consensus

Type of users • End-users

• Key-users

Degrees of user inclusion • Recipients of communication

• Symbolic participation • Participation by advice • Participation by weak control • Participation by doing

• Participation by strong control

User selection criteria Individual skills

• Knowledge of the work domain. • Knowledge about IT.

• Interest in the project and a desire to contribute to it.

• Ability to empathize with others and understand their needs.

• Ability to pass knowledge and enthusiasm about the system to other users.

• Readiness to work with technical issues.

(23)

Cross-selection • Age group, • Work experience, • Adopter category • User groups. Organisational boundaries • Structural • Social • Cognitive

Table 3. User Inclusion dimensions inspired by Mumford et al. (Mumford, 1979; Maas, van Fenema, & Soeters, 2016; Rasmussen, Christensen, Fjeldsted, & Hertzum, 2011; Ives & Olson, 1984)

2.5 Phases of an ERP implementation

In order to be able to discuss timed events and compare them across different cases, a common model defining the steps that an implementation goes through is needed. In 2000 Parr & Shanks (2000) published an article where they presented the project phase model (PPM, see Figure 1), which is a synthesised model of previous implementation process models. Their study is realized through a qualitative cross-case study on two large companies that operate in the oil business. The paper focuses on viewing critical success factors (CSFs) from a perspective of their importance depending on when they are applied during an ERP implementation (Parr & Shanks, 2000). The PPM divides the ERP implementation into three main phases:

• Planning

This phase consists of purpose clarification, ERP selection, project scope, deciding on implementation approach and resource budgeting (Parr & Shanks, 2000).

• Project

The project phase is divided into five sub-phases:

1. Setup

The setup sub-phases consist of assembling the project team(s), integration and reporting processes are established and guiding principles are agreed upon (Parr & Shanks, 2000).

2. Re-engineering

The re-engineering processes consist of, business process analysis in order to define the needs for business process reengineering (BPR) and training the project members (Parr & Shanks, 2000). 3. Design

The design phase starts with high-end design and is then accompanied by detailed design that is subject to end-user acceptance (Parr & Shanks, 2000).

4. Configuration & testing

(24)

This sub-phase consists of the development and/or configuration of the system, this entails activities such as configuring a test environment with real data, testing user interfaces, writing and test reports and system and user testing (Parr & Shanks, 2000).

5. Installation

The Installation sub-phase consist of installing necessary software and hardware, user training and setting up support (Parr & Shanks, 2000).

• Enhancement

The enhancement phase consists of continuous development and maintenance while the system is operating live in the organisation. This can be done through activities such as user support, system repair, system transformation and the extension of the system (Parr & Shanks, 2000).

2.6 Change Management

While previous chapters have focused mainly on ERP theory, this chapter focuses on change management and how an organisation prepares, executes, and institutionalises change. The purpose of this study is to research how user inclusion, i.e. user participation and communication, can be used to promote symbolic adoption, which is a state of mind in which the user thinks of the new ERP system as a worthwhile concept. The theories from change management will be used to understand how communication and participation can sway a user’s perceptions of a change on a cultural level, enabling and institutionalizing not only the change itself, but symbolic adoption as well.

One of the most used and influential models for change management is the CATS model, changing

in three steps. The concept emerged from an article by Kurt Lewin (1947) and is the foundation

of the field of change management. There have been many proposed variations of the model over the years, either with the same three-step structure (Schein, 2010), or with additional steps (Kotter, 1995). Such models can however often be arranged into the same three steps of the classic CATS model, where each smaller step is part of one of the steps from CATS (Cummings, Bridgman, & Brown, 2016). Edgar H. Schein is an avid proponent of CATS who elaborated and refined the basic model no less than 5 times between 1961 and 1965, and then again in 2010 in his book on organizational culture and leadership (Schein, 2010). Schein’s (2010) book is an elaborate and foundational piece in the CATS literature with over 60’000 citations and will as such be the basis for this study. The classic CATS model can be seen in Figure 2. Kotter’s (1995) article introducing the 8 steps of change is well cited in its own right, with close to 9000 citations, and will be used as a complement and as comparison to the more basic three step model as depicted in Figure 2.

Figure 2. The changing as three steps (CATS) model that was first introduced by Lewin (1947) and later reworked many times. This version was presented by Schein (2010).

An underlying assumption of the CATS model (Figure 2) is that humans, and especially humans in groups, are naturally resistant to change (Lewin, 1947). They need to be convinced that a change is necessary (unfreeze), implement the change (change) and then reinforce the change and make it

(25)

the new norm (refreeze) in order not to fall back into old behaviour. Schein (2010, p. 320) argues that all humans seek to maintain an equilibrium and that disturbances to this equilibrium is what causes humans to change, in search of a new equilibrium. This equilibrium can therefore cause resistance to change when it is stable, whereas if it is disturbed, it can be the very motivation for change. Lewin (1947) conceptualized this equilibrium as a force field created by the group standard (values and norms held by the group) and that any deviation from this group standard by an individual would be met with resistance. Lewin (1947) therefore argues that changing the group standard is necessary in order to make a persistent change. He even goes further and suggests that it might also be easier to change an individuals’ behaviour by first forming them into a group and then changing the standards of that group. The CATS model does however have its critics (Bartunek & Woodman, 2015; Kanter, 2003), claiming that the model is too simplistic in its assumptions that an organisation can ever be truly frozen, and that change happens in a linear fashion. Although both are valid points, we still believe that the model can be valuable. We believe that group behaviour can have a frozen quality as per the previous discussion about equilibria of group behaviour and the three steps will also be considered as intertwined, as opposed to being distinct and linear.

Unfreeze

The act of unfreezing might look very different in different cases, but the goal is to create the necessary motivation to change (Lewin, 1947, pp. 34-35). Lewin (1947, pp. 34-35) also suggests that this step is not always necessary, which Schein (2010, p. 320) elaborates on this and argues that unfreezing is necessary if the core cognitive structure is to change in more than in minor incremental ways. Transformative change is change that requires a group or individual to unlearn something as well as learning something new, which almost always requires some amount of culture change (Schein, 2010, pp. 320-321). The implementation of an ERP system should therefore be considered a transformative change for those whose work processes will be changed. The greatest difficulties with such change have to do with the unlearning of embedded routines that might be part of the individuals personal or group identity (Schein, 2010, p. 321). The fear of losing part of our identity or sense of belonging to a group might make some behaviour hard to give up, even if it seems to be dysfunctional (Schein, 2010, pp. 320-321). It is therefore necessary for managers and project managers to enact project initiatives making change seem both necessary and possible before a change can occur. Schein (2010, pp. 320-325) proposes three conditions necessary to achieve the desired unfreezing effect:

1. Enough disconfirming data to cause serious discomfort and disequilibrium.

2. Discomforting data connected to important goals and ideals, causing anxiety and/or guilt. 3. Enough psychological safety to be able to see a possibility to solve the problem without

loss of identity or integrity.

Disconfirming data is any information showing that some goals are not being met, or some processes are not producing the desired outcomes (Schein, 2010, p. 321). Examples of disconfirming data is dwindling sales, higher employee turnover, or an increase in customer complaints. Such disconfirming data is usually only symptomatic, meaning it is not clear what the underlying problem is, but it does create a disequilibrium by showing that something is wrong somewhere (Schein, 2010). This disconfirming data can create two types of anxiety: survival

User Inclusion during ERP Implementations and its effects on Symbolic Adoption