Contribution of normative stakeholder theory to the improvement of factors affecting data warehousing implementation success

(1)

Contribution of normative stakeholder theory to the

improvement of factors affecting data warehousing

implementation success

Master Thesis

David Szorenyi

Student number: 10629882 Information Studies

Business Information Systems Supervisor: dr. Gábor Kismihók

(2)

Abstract

With exponentially growing data quantity, the importance of data warehousing implementation success significantly increased. High level of system qua lity is more associated with succeeding in organizational- and project implementation than in technical implementation, therefore this thesis aims to find a way to improve factors affecting organizational- and project implementation success. Majority of them is strongly people-related, but still, there is a gap in literature on how these factors should be improved. Many researchers in the discipline of stakeholder theory are engaged with stakeholder identification and classification. This research investigates how normative stakeholder theory can contribute to the improvement of these people-related factors by conducting an exploratory case study. Stakeholders of a project at the University of Amsterdam are identified and classified in accordance with a significant classification model, furthermore, presence of the previously defined factors is measured within the project. As a result, the thesis provides several recommendations on the improvement of these factors by linking them with the identified and classified stakeholder groups.

(3)

As quantity of data exponentially increased in the last decades, collecting data manually from several sources was more time-consuming and the chance for errors became relatively big. Furthermore, data had to be ‘cleaned up’ before analysis and new policy implementations into databases were often impossible (Hoekstra, 2014). Therefore the importance of data warehousing (DW) projects noticeably increased in the last couple of years (Rome, 2004; Cuzzocrea & Dayal, 2011). Understanding the technical requirements of such projects was never my personal interest, but rather the management of people and communication channels between them. My personal experience shows that the difficulties of handling conflicting interests, involving potential stakeholders and many others are present issues in IT projects that need to be better handled. Ineffective stakeholder management can easily lead to mistakes, delays and misinterpretation or misunderstanding of information, which can have crucial consequences, even the failure of the project (Nelson, 2007; Cerpa & Verner, 2009). Many scholars in the discipline of stakeholder theory highlighted the significance of identifying, classifying, involving and engaging stakeholders and their claims (e.g. Mitchell et al., 1997; Agle et al., 1999; Espinosa-Orias & Sharratt, 2011; Parent & Deephouse, 2007). Most of the literature examines the previously mentioned issues in an organizational environment. But what happens if stakeholder identification and classification are examined the context of an IT project-, more specifically a data warehousing project? How could stakeholder theory help in a better understanding of DW implementation success?

There are several factors in the data warehousing literature that are proved to influence data warehousing success, substantial part of them can be closely related to the management of stakeholders, for instance adequate user participation, proper management support, high level of team skills etc. (Wixom & Watson, 2001; Shin, 2002; Yeoh & Koronios, 2010). Using Mitchell et al.’s (1997) normative stakeholder theory as the backbone of the theoretical framework, stakeholder groups can be identified and classified by assigning the attributes of power, legitimacy and urgency to them. This way, better identification and positioning of stakeholder groups are assumed to help improving the factors affecting DW organizational- and project implementation success.

In this research, the following question is aimed to be answered: How can normative stakeholder theory contribute to the improvement of factors that affect data warehousing implementation success? To do so, an exploratory case study is conducted by examining the UvAInform Learning Analytics project that consists of a data warehousing project and several pilot projects based on application of the data warehouse at the University of Amsterdam (UvA). The reason behind choosing the UvAInform project is that during gathering preliminary knowledge about the project it turned out that the complexity of stakeholder relations and unclear organizational structure of the project causing many difficulties in making progress towards successful implementation. The recognition of the need to resolve this practical issue encouraged me to analyze UvAInform and come up with suggestions on improving implementation success factors by underpinning the problem with both stakeholder and data warehousing theory.

The thesis purports to contribute to both stakeholder-, and data warehousing project management theory by offering solutions to the following issues. In my case study, a detailed stakeholder map is created that also contains the relations between stakeholders. This visualization helps see the bigger picture and can also point at possible deficiencies, missing, conflicting or unnecessary roles and relations in an early phase of the project. With aligning identified and classified stakeholder groups with the factors affecting DW organizational- and

(5)

2

project implementation success, recommendations can be provided on the improvement of activities related to the success factors for each relevant stakeholder group.

The thesis is divided into seven chapters. After the introduction, in Chapter 2, DW project environment is described by defining DWs and the UvAInform project, followed by the presentation of factors affecting organizational- and project implementation success of DWs. In Chapter 3, basics of stakeholder theory with a special focus on definitions, categorization and identification is introduced, followed by the detailed explanation of Mitchell et al.’s (1997) normative stakeholder theory. In Chapter 4, the thesis’ research approach is presented that consists of the explanation of the theoretical framework, the research questions and the research design. In Chapter 5, results are shown containing the detailed stakeholder map, classification of stakeholders and presence or absence of data warehousing success factors. In the beginning of each subchapter, particular research methods are presented. In Chapter 6, recommendations are provided on the improvement of the previously defined success factors, also linking them with the classified stakeholder groups. Finally, conclusions of the thesis are drawn in Chapter 7, where the limitations of the research are also recognized followed by suggestions for further research.

2. Data warehousing

Data warehousing has become one of the most significant development areas in the field of information systems already in the late 1990s (Wixom & Watson, 2001). In 1996, 95% of the Fortune 1000 companies either did have a data warehouse or they planned to develop one (META Group, 1996). Since then, especially in the previous couple of years, quantity of data that organizations have to deal with, increased exponentially. In 2008, there were already 12 gigabytes of information processed by the world’s servers per worker per day, which means 9.57 zettabytes in aggregate.(Short et al., 2011)

A data warehouse is a repository of integrated data and information that is meant for analysis and querying most commonly to support decision making (Wixon & Watson, 2001; Inmon, 1992a; Inmon, 1992b). One of the most often quoted definitions in the literature by Kimball (1996) is the following: “a copy of transaction data specially structured for query and analysis” (p. 310). The enormous advantage of data warehouses is that all relevant information is already integrated from different heterogeneous systems, quickly and directly available from one source, bypassing the mediated way of data and information integration. Many years ago, physical copy of data was named as one of the potential disadvantages of data warehousing (Hammer et al., 1995), but decreasing storage price trends were already seen on the horizon. Given the facts that data is filtered before warehousing in order to avoid redundancy, physical storage prices drop from month to month and cloud computing offers lashings of new opportunities in storing, this is definitely not considered as a disadvantage anymore. Because of the inconsistency between the source systems and the data warehouse, latter might be outdated from time to time according to Hammer et al. (1995). Real-time integration of the source and the data warehouse provides a solution for the problem, so urgent client needs will not meet barriers.

2.1. Introducing Learning Analytics and the UvaInform Project

There are many ways of exploiting the advantages of data warehousing, but analyzing individual customer behavior is one of the most profitable options, especially in the fields of sales and marketing (Wixom &Watson, 2001). Considering the same principle in the world of

(6)

3

education, there are lots of opportunities in tracking and analyzing individual student behavior and performance during the whole period of studying. Learning Analytics (LA) collects, measures, analyzes and reports findings on the basis of “digital breadcrumbs” that learners leave in different computer systems with the main purpose of comparing and predicting student performance, discovering social interactions and optimizing learning outcomes and learning environments (Educause, 2011; SOLAR, 2010).

In this research, the UvaInform Learning Analytics project is examined. This project launched in 2013, is running at the moment and financed by the UvA. The objective of UvAInform is to deliver a community sourced, secure, scalable repository for the use of learning analytics within the UvA. Learning Record Store (LRS) subproject is focusing on building a repository of student activity. The LRS will reliably store and retrieve data from Blackboard, Student Information System, MijnUvA and potentially 60-65 other systems from the UvA. The LRS is designed to work at scales above 100 billion records and will enable collecting student activity streams, querying and administration. The LRS is planned to be the basis for several pilot projects focusing on applications and data visualization for potential users. There are several pilot project proposals and developments running or will run in the following weeks, for instance, one of them aims to reduce dropouts in Bachelor programs. Another pilot is planning to mirror traditional (exams/assignments) and non-traditional study performance. Furthermore, in the frames of the Pervasive Personalized Goal-Setting Platform with Mirroring pilot project, an application being developed. This built-in web application will enable students to set goals, monitor their progress, set deadlines and receive dashboards about their activities with a progress bar. Users can share their goals and monitor others’ shared goals as well, this way mirroring their performance. Teachers can also create profiles and provide feedback, recommendations and reminders for students who shared their goals. This pilot will be separately used in some interviews in this research, as this is the one which will probably be introduced first.

2.2. DW success factors

Wide range of DW implementation success factors can be found in the literature (e.g. Wixom & Watson, 2001; Shin, 2002; Yeoh & Koronios, 2010; Adamala & Cidrin, 2011). Implementation success of data warehouses depends on organizational-, project- and technical implementation success. As the assumption that technical implementation success is associated with high level of system- or data quality was not supported (Wixom & Watson, 2001), in my research, stakeholder-related success factors are in focus, consequently technical factors, like the quality of data sources and development technology are excluded.

Image 1 illustrates the factors affecting organizational and project implementation success, building on the framework of Wixom and Watson (2001) and additional sub-factors from Shin (2002 – marked with red frames on the image), Yeoh and Koronios (2010 – marked with blue frames on the image). Yeoh & Koronios (2010) investigated critical success factors of business intelligence systems based on the infrastructural foundation of data warehouses, therefore their findings were used to complement the theory of Wixom & Watson (2001) and Shin (2002) and to build up a complex framework for this thesis. Wixom and Watson’s (2001) research validated their hypotheses that data– and system quality is associated with a high level of perceived net benefits. Data quality refers to data accuracy, completeness and consistency (Lyon, 1998; Shanks & Darke, 1998), while system qua lity is commonly measured by flexibility, integration, response time and reliability (DeLone & McLean, 1992).

(7)

4 Image 1 – Factors affecting organizational- and project implementation success in case of data warehousing

(Wixon & Watson, 2001, p. 20-34.; Shin, 2002; Yeoh & Koronios, 2010)

High level of system quality can be reached by high level of organization– and project implementation success. Building and implementing a data warehouse can cause lots of organizational changes that potential users have to accept and integrate into their work or study processes. It particularly has effects on data ownership (becomes centralized), data access and business processes. Data access responsibilities shift from IT personnel to end-users, whose performance will also be affected by having access to the DW (Wixom & Watson, 2001). Furthermore, in my assumption, these issues would require strong legal and ethical regulation and support around data usage, especially in case the DW stores big quantity of private data (UvAInform is dealing with enormous quantity of private student data). It is also assumed that insufficient regulation of the legal and ethical issues can lead to serious consequences that can result in ending the project. The project implementation is successful if the team is able to finish the project in time, on budget and with the right functionality (Constantine, 1993; Waldrop, 1984). In this research, as an additional factor, high level of legal/ethical support is assumed to affect organizational- and project implementation success of UvAInform project.

Organizational- and project implementation success are both associated with high level of resources and user participation. Money, people and time that are required to successfully deliver a project are called resources (Ein-Dor & Segev, 1978). These resources are especially required in data warehousing projects, because they are expensive, time-consuming and require carefully selected people (Wixon & Watson, 2001). Overlapping scope with other projects is a major risk factor as it can cause resource allocation problems among the projects (Shin, 2002). According to Hartwick and Barki (1994), if users are assigned to project roles and tasks, their requirements are more likely to be met and therefore the implementation of the data warehouse will also be more successful. User participation helps managing user expectation and satisfying user needs (Edelstein & Barquin, 1997; Watson & Haley, 1997). It is especially true when requirements of a DW are initially unclear, hence it is extremely important to involve users when developing a DW as the systems will only be well-usable, if potential users are involved in time and have influence on the development process. Shin (2002) added that lack of clearly defined users is a risk as their feedback is certainly needed for successful organizational- and project implementation. According to one participant of the

(8)

5

research of Yeoh and Koronios (2010), without user participation, it might not even be possible to deliver the adequate system: “Users should be an important partner in building and delivering the right system. Without their consistent input, we technicians cannot deliver the right system” (p. 28).

High level of management support affects organizational implementation success (Wixom & Watson, 2001). This factor means the motivating power and the managerial sponsorship in helping the project be accepted by the whole organization (Curtis & Joshi, 1998; Watson et al., 1998), furthermore political resistance and weak participation can be reduced as well (Markus, 1983). It is not only true for DW projects but generally for IT implementations, for instance decision support systems (Guimares et al., 1992; Igbaria et al., 1997). Watson et al. (2004) emphasized the importance of active management participation, especially from the senior members.

Clear vision and a well-established business case are also required for successful organizational implementation success (Yeoh & Koronios, 2010; Adamala & Cidrin, 2011). However, when applying the framework of success factors in this research, examining business cases will be excluded as UvAInform is an educational project and commercial use of the project outcomes is still not part of the vision. Although the research of Wixon and Watson (2001) did not find relation between the presence of a champion (someone who can bring the project to the next level and “possess the skills and clout needed to overcome resistance that may arise within the organization”; Wixom & Watson, 2001, p. 23) and organizational implementation success, Yeoh and Koronios’ (2010) study emphasized the significance of it as one of their interviewee mentioned: “The team needs a champion. By a champion, I do not mean someone who knows the tools. I mean someone who understands the business and the technology and is able to translate the business requirements into a (high-level) BI architecture for the system” (p. 27).

The previously mentioned, implementation-led organizational changes can lead to resistance from managers, data suppliers and also from potential end users (Markus, 1983; Wixom & Watson, 2001; Shin, 2002), therefore adequate change management within the project organization is necessary and will affect project implementation success (Shin, 2002; Yeoh & Koronios; 2010). High level of team skills supports high level of project implementation success (Wixom & Watson, 2001). According to Edelstein and Barquin (1997), DW developing teams’ skills directly influences the project outcome. As the importance of user participation was emphasized previously, a project team that owns strong technical and interpersonal skills is able to manage tasks and interact with users in a more effective way (Constantine, 1993; Finlay & Mitchell 1994). Other factors either do not affect data warehousing success, or they are technology-related, consequentially they are out of the scope of this research.

2.3. Conclusion

Definition of data warehousing was previously introduced in this chapter in order to define the context of this research. The framework for factors affecting DW organizational-, and project implementation success were presented in order to examine whether they could be improved with better stakeholder identification and classification, these will be covered in Chapter 3. The rationale behind this assumption is that each factor can be related to stakeholder positioning, behavior, skills and/or performance, some tightly, others slightly. Management support largely depends on how manager(s) within the project can motivate and sponsor the team in order to minimize organizational resistance and empower employee

(9)

6

participation in major decisions. Part of that, managers have to get the vision written down at the right time of the project and make it sure that it is clearly communicated towards each relevant stakeholder group. Change management is also mostly dependent of managerial behavior, performance and the way of managerial communication against organizational resistance. The champion factor itself represents a stakeholder whose performance greatly influences DW implementation success, just as team skills themselves mean how project team members are able to perform (technical skills) and effectively communicate with each other (interpersonal skills). Resources represent the human resources (also stakeholders), time (e.g. shareholder’s responsibility to set strict deadlines to make pressure on the team) and money (resource allocation skills). User participation as one of the most important factors requires right stakeholder involvement and adequate communication towards potential users. High level of legal and ethical support can only be provided by related organizations that are adequately identified and involved in the project.

It is assumed that these factors can be improved better if related stakeholder groups’ nature (including stakeholder position, behavior, performance etc.) is better known. As a first step towards reaching this goal, stakeholders have to be identified and classified in a proper way. In my assumption, normative stakeholder theory can cover this task and can contribute to the improvement of the previously mentioned factors, especially with focusing on the managerial body of the organization as most of the factors are related to it.

3. Stakeholder theory

In this section, relevant stakeholder literature is reviewed, definitions of stakeholders, furthermore identification and classification aspects are provided. Mitchell et al.’s (1997) normative stakeholder theory is explicitly explained in Chapter 3.3, as its contribution to the improvement of data warehousing success factors is particularly investigated in this research. 3.1. Who are the stakeholders?

Modern management literature takes the concept of stakeholders into consideration since Freeman (1984) published his significant book: Strategic Management: A Stakeholder Approach. He aimed to enable managers to understand and adequately and effectively manage stakeholders. He defined stakeholders as “any group or individual who can affect or is affected by the achievement of the organization’s objectives” (Freeman, 1984, p. 46). This definition is one of the most cited ones and gives the basis for latter research papers in the topic. According to Achterkamp & Vos (2008), the definition takes a “landmark” position in stakeholder theory, although it has been criticized many times for being too broad. For instance, Donaldson and Preston (1995) said that taking Freeman’s (1984) definition, the number of stakeholders can be unlimited and it could include competitors and the media as well. Many other authors reformulated the definition of stakeholders, for instance Savage et al. (1991) state that they "have an interest in the actions of an organization and ... the ability to influence it" (p. 61). Clarkson (1995) defined stakeholders as those who “have, or claim, ownership, rights, or interests in a corporation and its activities” (p. 106). According to Donaldson and Preston (1995), stakeholders are “persons or groups with legitimate interests in procedural and/or substantive aspects of corporate activity” (p. 85).

(10)

7

3.2. Stakeholder identification

Stakeholders have to be identified in order to manage their claims and to be aware of their influence, (multiple) roles or even their existence. Many scholars identified stakeholders by categorizing them from different perspectives

According to Savage et al. (1991), there are two attributes that have to be considered when identifying stakeholders: their claims and their ability to influence the organization. Scholars provide several ways to identify and classify stakeholders, most of them are fine-grained categorizations from Freeman’s (1984) broad definition. Clarkson (1995) differentiate primary and secondary stakeholders, while according to Blair and Whitehead (1988) there are stakeholders who have “potential for collaboration” and there are stakeholders who have “potential for threatening”. Goodpaster (1991) differentiates fiduciary and non-fiduciary stakeholders, according to Clarkson (1995) there are voluntary and involuntary stakeholders. Vos and Achterkamp (2006) state that stakeholders can be differentiated as actively and passively involved. However, it is not easy to create general predefined categories of stakeholders as every organization is different, but particular categories often appear in several organizations. According to Clarkson (1995) who gathered data from 70 field studies, the following stakeholder groups get attention commonly from the firm: employees, customers, shareholders, suppliers, governments and the local community.

Vos & Achterkamp (2006) proposed a stakeholder identification model in the context of innovation projects (UvAInform is considered an innovation project as LA is a relatively new research area). After defining the goal of the project, people who understand the project have brainstorming session(s) where participants individually collect all possible stakeholders who could be involved in the project. This methodology can fit in the research approach of my thesis as there are many potential stakeholders to be identified.

This thesis project applies Mitchell et al.’s (1997) theory for classifying stakeholders as it fits the requirements that are set in literature (Currie et al., 2009): it takes site-specific influences into consideration (Simpson, 2001), its typology is finer-grained than other generic groups (Harrison & Freeman, 1999). The theory of Mitchell et al. (1997) is separated and explained in detail in the next chapter as it functions as part of the backbone of my theoretical framework. By combining these two methods, it is assumed that all stakeholders within the project will be clearly identified and properly classified and this way it will contribute to the improvement of factors defined in Chapter 2.

3.3. Normative stakeholder theory

According to Mitchell et al. (1997), Freeman’s (1984) definition is a very broad one based on the “empirical reality that companies can indeed be vitally affected by, or can vitally affect, almost anyone” (p. 857). Setting boundaries thus is necessary when it comes to stakeholder identification and classification. Also, effective prioritization is needed as treating stakeholders equally will be cost ineffective and “will potentially conclude in a stalemate with opposing positions” (Currie et al., 2009).

In the significant article of Mitchell et al. (1997), “Toward a theory of stakeholder identification and salience: defining the principle of who and what really counts”, the basis of normative stakeholder theory was formed by researching the potential positive influence between stakeholders’ possession of three attributes (power, legitimacy and urgency) and stakeholder salience (”the degree to which managers give priority to competing stakeholder claims” (Mitchell et al., 1997; p. 869)).

(11)

8

In the following subchapters, first the attributes of power, legitimacy and urgency are elaborated, then the classification model is described. Measuring salience and validating Mitchell et al.’s (1997) proposition is excluded as they do not fit in the scope of this research.. By applying Mitchell et al.’s (1997) framework in this thesis, power relations will be clearer in the organizational stakeholder structure. Investigating whether potential stakeholders possess the three attributes results in the classification model is assumed to help better understand the nature of stakeholder groups, this way it is hypothesized that the previously introduced DW success factors can be improved.

3.3.1. Power

Mitchell et al. (1997) took Pfeffer’s (1981) definition of power as follows: “A relationship among social actors in which one social actor, A, can get another social actor, B, to do something that B would not have otherwise done” (p. 869). This, and other current definitions of power mostly originated from Weber’s (1947) idea, who stated that power is the "probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability exists " (p. 52). Mitchell et al. (1997) share the opinion of Pfeffer and Salancik (1974) who said that it might be difficult to define power, but it is much easier to perceive that it is “the ability of those who possess power to bring about the outcomes they desire” (p. 3).

Mitchell et al. (1997) explain the bases of power by using the typology of Etzioni (1964) who differentiated three types of power in organizational settings based on the type of resource that is being used during exercising power. The first one is coercive power, which is based on physical resources of force, violence and restraint. Etzioni (1964) exemplifies coercive power with the usage of a gun, a whip or a lock as they are both usable for physical sanction. Coercive power is excluded from my research as applying physical sanction is against the law in the UvAInform project environment. The second type of power is called utilitarian power and is based on material or financial resources. This is the ability to control material resources like services, goods and money. The last type of power is called normative power and is based on symbolic resources. Etzioni (1964) claims that normative power includes normative symbols such as prestige and esteem; and social symbols like love and acceptance.

3.3.2. Legitimacy

Mitchell et al. (1997) argue that the notion of legitimacy is often mentioned implicitly together with power, for instance, Davis (1973) differentiates legitimate and illegitimate use of power. Although many scholars link the two attributes together, it is not necessarily the case that they exist together (Mitchell et al., 1997): legitimate stakeholders are not always powerful (e.g. members of a project executing board without voting rights) and powerful stakeholders are not always legitimate (e.g. influential people directing projects from the background, unofficially).

Weber (1947) proposed that the combination of the two attributes is called authority (more accurately the legitimate use of power), so a distinction and a separate attention to legitimacy in stakeholder theory are needed. Based on Weber’s (1947) findings, Suchman (1995) came up with a broad, sociologically based definition: “a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, definitions” (p. 574). This definition was accepted by Mitchell et al. (1997) as they found it as an adequate description to use when it comes to identifying and classifying stakeholders.

(12)

9

3.3.3. Urgency

Mitchell et al. (1997) proposed that urgency, as the third stakeholder attribute, is meant to represent the dynamism of stakeholder-manager interactions. Mitchell et al. (1997) were the first ones who raised awareness on looking at stakeholder relationships considering time as an influencing factor and also the criticality of stakeholder claims, thus they argue that urgency is based on the following two sub-attributes:

“(1) time sensitivity - the degree to which managerial delay in attending to the claim or relationship is unacceptable to the stakeholder

(2) criticality - the importance of the claim or the relationship to the stakeholder.” (p. 867). Based on the definition of ‘urgency’ in the Merriam-Webster Dictionary and the two sub-attributes, urgency is defined as follows: “the degree to which stakeholder claims call for immediate attention” (p. 867).

Parent and Deephouse (2007) found proof for the existence of urgency, but stated that power and legitimacy is more likely to be present. Their research found that activist groups are the stakeholders who most rely on urgency (p. 18). It is very important to note that while power and legitimacy are attributes that are mainly connected to the stakeholders themselves, urgency always refers to stakeholder claims (Mitchell et al., 1997).

3.3.4. Classification model

Based on how many attributes stakeholders possess according to managers, Mitchell et al. (1997) identified seven types of stakeholders, as can be seen on Image 2. Each stakeholder type has different characteristics that are described under the image.

Image 2 – Stakeholder classification (Mitchell et al., 1997, p. 874)

According to the model, if a person or group does not possess any of the attributes then it is not considered as stakeholder. Stakeholders possessing one attribute are called latent stakeholders (marked with 1, 2 and 3 on Image 2). These stakeholders are often not even recognized by managers. Dormant stakeholders possess only the power attribute which can be coercive (force), utilitarian (money) and symbolic (media). Although they have usually little or no interaction to the organization, they can be very dangerous by acquiring another attribute. Stakeholders possessing only the legitimacy attribute are considered as

(13)

10

discretionary stakeholders. There is no pressure on managers to make their relationships more active, a typical example could be the non-profit organizations. Stakeholders, whose relationships with managers are considered urgent, but the attributes of power and legitimacy are not possessed by them, are defined as demanding stakeholders. Mitchell et al. (1997) describe them as “mosquitoes buzzing in the ears of managers”, they can be bothersome, but they are not potentially dangerous.

According to Mitchell et al. (1997), stakeholders who possess two attributes are the ones who enter the ‘active’ zone from the ‘passive’. This category was named expectant stakeholders (marked with 4,5 and 6 on Image 2). Those stakeholders who are powerful and legitimate, meaning they are able to act on their claims, are classified as dominant stakeholders. They are ‘the stakeholders’ for many scholars, they expect and receive much attention from managers. Usually, they have some formal mechanisms, good examples can be the boards of directors or the human resources departments. Stakeholders with urgent legitimate claims are characterized by Mitchell et al. (1997) as dependent stakeholders as they are dependent upon someone for the power needed to succeed with their wills. For instance, local communities can have urgent and legitimate claims but they obviously miss power. Sometimes dominant stakeholders (i.e. state government) adopt dependent ones, moving towards to the most salient stakeholder category. As dormant stakeholders were described as potentially dangerous, it is suggested by Mitchell et al. (1997) that when they acquire the urgency attribute, they become coercive and possibly violent, consequentially, dangerous stakeholders. Examples for this category can be employee strikes and sabotage or terrorism. The actions stakeholders of this class take are not only dangerous for stakeholder-manager relationships but also for the involved individuals and entities.

When all three attributes are present at individuals or groups, they are called definitive stakeholders (marked with 7 on Image 2). Any expectant stakeholder can become definitive by gathering the missing attribute. Definitive stakeholders should always get the highest attention, necessity of acknowledgement and action should always be recognized by these groups. A good example is when stockholders of a company (dominant stakeholders) feel that they legitimate interests are not being served properly by managers and take action (their claims become urgent). If managers respond inadequately to the action, they can easily be removed from their position.

4. Research approach

4.1. Research question

Having personal experience in working in IT projects with many people with many different backgrounds and opposite interests turned my personal attendance towards researching how stakeholders and their claims could be adequately managed in IT projects in order to reach success. After getting to know the UvAInform project, I decided to narrow down my research scope for data warehousing projects (LRS and the related pilot projects) as the settings of UvAInform provided an ideal environment for the research. Preparing the preliminary literature review for the thesis and having initial meetings with managers of the UvAInform project shed light on the chaotic stakeholder management situation including non-identified stakeholders, undefined or multiple roles, lack of resource allocation skills and many other people-related issues causing daily problems and difficulties in the project work. Technical requirements can be more accurately defined for data warehousing projects, but properly

(14)

11

managing the people towards a successful project implementation is never easy, mostly because human behavior is much more unpredictable.

In order to cover this problem, this research aims to identify and classify project stakeholders in a way that it helps in the improvement of factors affecting data warehousing organizational- and project implementation success. Using basic stakeholder definitions, identification methods and the normative classification model of Mitchell et al. (1997), a clear, detailed map with all potential stakeholders and their relations is created, furthermore these identified stakeholder groups are classified. By being aware of the project’s accurate organizational structure and stakeholder relations, it is assumed that this knowledge will help in the improvement of factors affecting DW organizational- and project implementation success. The underpinning rationale behind this assumption is that all of these factors can be strongly related with the nature of classified stakeholder groups, so proper identification, classification and management of these stakeholders and their claims can contribute to the improvement of the factors and this way it can help data warehousing projects reach implementation success (as can be seen on Image 3 below).

Image 3 – Theoreticalframework

Examining the relevant literature, I haven’t find evidence of scholars aiming to research the relation between the discipline of normative stakeholder theory and data warehousing success factors, even the usage of Mitchell et al.’s (1997) stakeholder theory is quite rare in IT project management literature (e.g. Boonstra, 2005; Tan et al., 2005) and it was not applied in data warehousing projects at all. For these reasons, it is suggested that there is a gap in the literature and therefore the following question is aimed to be answered in this thesis:

How can normative stakeholder theory contribute to the improvement of factors affecting data warehousing implementation success?

In order to extensively investigate the question, the following sub-research questions were formed:

1. Which factors affecting DW organizational- and project implementation success? 2. Who are the pot ential stakeholders in the project?

(15)

12

4. Are these factors present in the UvAInform project? 5. In what ways could these factors be improved?

6. How can the knowledge gathered about the nature of previously classified stakeholder groups influence the improvement of these factors?

4.2. Research design

In order to answer the previously listed research question and sub-research questions, an exploratory case study is conducted. Connecting normative stakeholder theory and data warehousing theory is a new research approach, just as Learning Analytics is an entirely new research area. Therefore, describing one case in detail could be the best option to design such research. Case studies can be especially valuable as they are able to describe the events of a specific case (the UvAInform project in this case) in an in-depth manner (Parent & Deephouse, 2007, p. 4). According to Yin (2003), case studies help answer the questions “how” and “why” in a case that the researcher has not much control over. This fits in this research as the goal is to find out in which ways should stakeholders be managed in order to reach data warehousing project success.

The research consists of four steps in the frames of a case study: (1) building theoretical framework, (2) stakeholder identification and classification, (3) examining the presence of factors affecting organizational- and project implementation success and (4) recommendations on the improvement of formerly defined factors, investigating also the relation with the classified stakeholder groups, these steps are shown on Image 4. As different methods or the same methods but with different contents are used in different steps of the research, they cannot be described in one general section. Therefore in this section, only the research design is introduced briefly, the individual methodologies are covered in the beginning of each ‘Results’ section.

As a first step, the theoretical framework was built up, starting with defining data warehousing. Afterwards, in order to answer the first sub–research question, a framework was created using several authors’ findings (e.g. Wixom & Watson, 2001; Shin, 2002; Yeoh & Koronios, 2010) (Chapter 2). In Chapter 3, stakeholder definitions were introduced and literature was reviewed on possible stakeholder identification approaches. Mitchell et al.’s (1997) stakeholder classification framework serves as the underpinning theory for classifying the previously identified stakeholders in my research. The potential contribution of this stakeholder identification and classification on improving factors affecting DW implementation success will be examined in this thesis.

In order to answer the second and third sub-research questions, stakeholders groups were identified and classified in the frames of a qualitative research. By applying refined stakeholder identification method of Vos & Achterkamp (2006), all stakeholder groups, positions and relations were identified in the project. As a next step, ten persons were selected for semi-structured interviews (backed up by a survey) in order to measure stakeholders’ possession of Mitchell et al.’s (1997 ) attributes. Concluding from the interviews, stakeholder groups were classified in accordance with the possession of the stakeholder attributes of power, legitimacy and urgency (Mitchell et al., 1997). The thorough methodology and the results are shown in Chapter 5.2.

(16)

13

In order to answer the fourth sub-research question, presence of factors defined in Chapter 2.2 were examined. The second part of the same semi-structured interviews investigated how interviewees experience the presence of these factors. In order to get different perspectives and opinions, eight additional interviews were also conducted with students, teachers of UvA and members of two organizations to be involved in the project. Questions during the interviews were built around the factors that were defined in Chapter 2.2, also seeking suggestions for resolving problems and difficulties caused by their absence. The particular methods and findings are detailed in Chapter 5.3.

Chapter 6 provides recommendations on the improvement of the factors based on the core interviewee’s answers, also by linking the knowledge gathered about nature of classified stakeholder groups.

5. Results

Chapter 5 covers the second (stakeholder identification and classification) and third step (investigating presence of previously defined factors) of my research design. Each subchapter consists of the research methodologies that were used in that part of the research and the results.

Case study

Image 4 – Research approach

1. Building theoretical framework

•Data warehousing definitions

•Framework for factors affecting DW organizational- and project implementation success •Stakeholder definitions

•Normative stakeholder theory (classification modell of Mitchell et al. (1997))

•Assumption about the potential influence between the norm. stakeholder theory and DW success factors

2. Stakeholder identification and

classification

•Stakeholder identification and stakeholder map development (qualitative method: brainstorming sessions and 2 in-depth interviews)

•Stakeholder classification (qualitative method: 10 semi-structured interviews measuring stakeholder attributes supported by a survey to be filled in beforehand)

3. Examining factors affecting

DW implementation

success

•Investigating the presence of previously identified DW success factors in the project by conducting interviews the same 10 participants about each. Eight additonal interviews helped in a better understanding of facors

4. Recommen-dations

•Recommendations on the improvement of previously defined factors, also examining the possible contribution of the knowledge acquired from stakeholder classification

(17)

14

5.1. Stakeholderidentification

5.1.1. Methodology and process

This chapter discusses how each stakeholder group was identified in the second step of my research. After building up the theoretical framework, a qualitative research has been conducted starting with weekly brainstorming sessions for a month (four appointments) with two project managers (one member of the LRS Steering Board and one faculty representative and pilot project manager, both of them aware of the stakeholders of the project). According to Vos & Achterkamp (2006), all potential stakeholder groups can be identified by individual brainstorming methods. In my opinion, there has to be discussion about the identification of stakeholders, so I applied their method to a small group of three. During these brainstorming sessions, all potential stakeholder groups were debated and were put on a list after arguing pro and contra in accordance with Freeman’s (1984) stakeholder definition(“any group or individual who can affect or is affected by the achievement of the organization’s objectives”, p. 46). Besides its great significance in stakeholder theory, this broad definition can be exceedingly useful in the case of the UvAInform project as more potential groups of stakeholders can be identified, involved and engaged than with using narrower definitions. Before each session, I asked the participant to think of potential stakeholder groups in accordance with Freeman’s (1984) definition. In the sessions, it was discussed whether each potential stakeholder group can affect or can be affected by UvAInform’s achievements. If at least two out of the three participants agreed that a stakeholder can affect or be affected by the achievement of the project than it was put on the initial stakeholder map (it was drawn on a blackboard). If a stakeholder group was put on the map, and if it was related to another group that was already on the map, the relation was illustrated between them with a line or an arrow. This way, the initial map was complete by the fourth session.

This initial stakeholder map helped visualize the organizational structure of UvAInform project. Two key persons were identified who were assumed to be able to clear up the remaining concerns about non-identified groups or stakeholders whose position could not be accurately defined. Therefore a semi-structured in-depth interview was conducted with both of them. Semi-structured interviews allow the interviewer to gain in-depth understanding of stakeholders’ organizational structure by having an interview guide as a checklist, a default wording and an order for the questions with the possibility that the wording and the order can change during the interview and additional questions can also be asked in order to adapt to the answers of the interviewee and to adequately cover all the topics (Robson, 2011). Having the initial stakeholder map as the guide for the interview and the expectation to have additional questions about different stakeholder groups motivated the researcher to use this type of method.

Both interviewees were asked at the beginning of the interviews to introduce themselves, their role(s) and responsibilities in the project. Due to privacy reasons, names of all interviewees in this research are not documented in this research. The first interviewee (Person X) was identified as the coordinator of the UvAInform project and was therefore asked firstly to identify all stakeholders in the project, secondly to compare his list to the initial stakeholder map. The second interviewee (Person Y) was identified as the person being responsible for the communication bridge between the UvA and external expert organizations. Therefore he was asked to explain which external organizations in which ways can affect or can be affected by the achievement of the UvAInform project.

(18)

15

All interviews were audio recorded with the permission of the interviewee. Afterwards, a summary of the interviews were sent to the interviewees to be reviewed and approved. This way, the interviewees could make sure that the researcher interpreted their answers correctly. The summaries of these interviews can be found in Appendix 2. Both interviews were held at the interviewee’s working environment. The objective of the interviews was to complement and finalize the earlier prepared stakeholder map in order to use it in the next phase of the research.

5.1.2. Stakeholder map

Resulting from the brainstorming sessions and the two interviews, the final stakeholder map was drawn including all potential stakeholders, their relations and hierarchical structure. This map can be seen on Image 5.This sub-chapter introduce all stakeholder groups of UvAInform project in accordance with the answers of Person X and Person Y.

The original proposal of UvAInform project was submitted by the UvA ICT Services (ICTS) to the UvA Expertise Group Education with the purpose of allocating a fix budget for Learning Analytics. In this proposal it was stated that different parties from different faculties could hand in LA project proposals (it was akin to a tender procedure). Afterwards it had been realized that a centralized decision making body was needed to coordinate the development of the LRS and the pilot projects, therefore the LA Focus Group (LAFG) was established. The LAFG is directing the project on an operational level and prepares strategic decisions for organizations on a higher hierarchical level. This organization is responsible for steering and managing the project towards the right direction. It consists of LA experts, researchers, teachers, a PhD student and representatives from all faculties of the UvA and from the ICTS. The LAFG votes on every major decision they make (e.g. approval of pilot projects), but only faculty and ICTS representatives have voting rights. Managers of pilot projects are also selected by the LAFG. In order to evaluate the development progress of the LRS and to monitor progress towards the project targets, a separate body was established, namely the Learning Record Store Steering Board (LRSSB). Developers and maintainers of the LRS and the applications and dashboards in the pilots are either members of the ICTS or external computer scientists, but they represent a separate role.

The previously mentioned ICTS is the central ICT unit of the UvA, supporting all IT processes within the university and also contributing to IT-related research projects like UvAInform. One level above the LAFG, the formerly referred Expertise Group Education (EGE) operates. This central body of the UvA receives reports from the LAFG and has the right to overrule their decisions as the project was officially submitted to them.

The EGE also decides on budget allocation among the LAFG and other Focus Groups. Person X is the chairman of LAFG, represents it on the meetings of EGE, furthermore he is also a member of the LRSSB. The EGE advising the UvA Steering Group ICT regarding decision making on ICT and innovation projects in the discipline of education, while the Steering Board ICT allocates budget between the EGE and other expertise groups. The UvA/HvA Board (College van Bestuur, CvB) is operating on the highest hierarchical level of UvA. The Steering Group ICT has the responsibility to report to and advise the CvB regarding new ICT projects. Like on lower hierarchical levels, Steering Group ICT has to compete with other organizations for the budget, but these are on a way too high hierarchical level to consider them as potential stakeholders.

Person X also has to report to the UvA/HvA Portfolio Management. This organization helps manage all projects being carried out at the UvA/ HvA and reports to the CvB, but does

(19)

16

(20)

17

not have the right to cut budget or have a voice in changing the projects’ strategic direction. Person Y is a representative of the ICTS in the LAFG, but besides that, he operates as a communication bridge between the UvAInform project and the external LA expertise groups. On an international level, he is a member of the community of LA initiatives within the Apereo Foundation, which is an open-source community supported by university experts, is building learning management- and portal systems for students. Apereo is cooperating with SOLAR (Society for LA), that is organizing global conferences (Learning Analytics & Knowledge, LAK) and summer institutes (LA Summer Institute, LASI). On a national level, Person Y is a member of SURF’s (collaborative organization for ICT in Dutch higher education and research) Special Interest Group LA where a group of experts disseminate learning analytics on a national level. On a regional level, Person Y is the member of the Regional LA Focus Group where in cooperation with other experts from other universities, they aimed to build up a common infrastructure to use for LA purpose, but the focus group stopped working recently more likely because of busy schedules. Person Y concluded:

“It is my responsibility to keep the LAFG informed about latest research results, directions, methods and ideas to implement updates, protocols and policies from all three levels”.

Potential users of the outcome products of the project are defined as incoming and current UvA students, teachers, lecturers and professors, directors of educational programs and student advisors. UvAInform project also has a long term goal that requires defining HR agencies ad future employers as potential users too. In frames of a PhD project, labor market-oriented learning trajectories are aimed to be built in order to monitor if it is feasible to match data from high school, UvA and labor market data. This way it would be possible to follow students without organizational boundaries and to use data set to run predictive models against particular labor market outcome. Therefore, recommendations could be provided for current students on how to facilitate their decision makings. Related PhD student plays a significant role in accelerating the Pervasive Personalized Goal-Setting Platform with Mirroring pilot project.

Person X emphasized the importance of legal and ethical support regulation in UvAInform:

“As data privacy is a critical aspect in the UvAInform project, legal and ethical support and approval has to be properly regulated within the project”.

At the UvA, Legal Affairs is responsible for advising the CvB, management, staff and students in legal matters if they request it. In ethical questions, the UvA Ethics Committee (AIEC) advises he CvB on guidelines related to ethical issues, but it does not deal with individual projects. According to Person X, in order for the UvAInform project to have the proper legal and ethical support, a centralized Data Governance Ethics/Legal Committee (Centre of Data Governance, CoDG) is needed. Negotiations have already started about establishing this central body as both Person X and Person Y stated that it was not only necessary for UvAInform, but also for other projects at the UvA dealing with serious data privacy issues.

There are two potential organizations to be involved in the project. Firstly the UvA Strategy & Information (S&I), that is responsible for strategic management and supplying information for strategic decision-making for the CvB. The S&I is also responsible for managing and developing Uvadata, which is defined as the data warehouse providing the legitimate source for decision making for UvA’s executive boards and managers in the domains of education, research, personnel, finances and accommodation (Hoekstra, 2014). As Uvadata and UvaInform are likely to have some common goals, cooperation would definitely help UvAInform accelerate its processes, according to Person X. The other potential

(21)

18

organization to be involved is the UvA Academic Affairs (AA) which develops educational strategic policies university-wise. The outcome and achievement of UvAInform might influence AA’s policy making in the future.

The chapter provided answer to “Who are the potential stakeholders in the project?”, which was the second sub-research question in this thesis. All of the potential stakeholder groups were identified, making it possible to classify them in the next step of the research.

5.2. Stakeholder classification

5.2.1. Methodology

In order to classify the stakeholders, ten persons were selected for semi-structured interviews in a way that all of them should be aware of the existence of most of the stakeholder groups and with the aim of representing as many core stakeholder groups with management functions as possible. Stakeholder groups outside the core bodies were not involved in this part of the research as most of them are not aware of many stakeholder groups and they would not have been able to judge whether particular stakeholder groups possess the attributes of power, urgency and legitimacy. Roles of the interviewees together with their given IDs are highlighted in Table 1.

ID Role

Interviewee#1 Representative of UvA ICTS in the LAFG (voter), Learning Analytics expert, involved in the national and international LA communities and keeps the LAFG updated on LA standards, findings and directions (Person Y onImage 5)

Interviewee#2 ICTS functional support, responsible for ISs for education, member of the LAFG

Interviewee#3 Coordinator of Information Managers in the ICTS, supporting the Steering Group ICT and expertise groups, also helping Portfolio Management

Interviewee#4 Faculty representative in the LA Focus Group, pilot project manager

Interviewee#5 Project management assistant, LAFG member

Interviewee#6 ICTS manager, board member of EGE, member of the LRS Steering Board

Interviewee#7 Project manager of the LRS, member of ICTS

Interviewee#8 LA researcher, representative of LAFG in EGE, chairman of LAFG, member of LRS Steering Board (Person X on Image 5)

Interviewee#9 PhD student, plays significant role in a pilot project, member of the Focus Group

Interviewee#10 Consultant, working between ICTS and different faculties, participating in projects concerning ICT in education, passive member of LAFG

Table 1 – Roles of core interviewees

Before the interviews were conducted, participants were asked to fill in an online survey. The survey aimed to support the interviews as limitation of time did not allow the interviewer to ask about every stakeholder group separately. Filling out the survey also helped interviewees think of all potential stakeholder groups before the interview and this way they could be more prepared. Consequentially, the survey was not meant to support the research with quantitative methods and therefore it was not analyzed statistically. In the survey, interviewees could agree or disagree on a 9-point Likert scale with statements regarding all stakeholder groups’ conditional possession of stakeholder attributes (power, legitimacy, urgency) (Mitchell et al., 1997). In order to create statements expressing the possession of stakeholder attributes in the most adequate way, definitions from Mitchell et al.’s (1997) paper were used (introduced in Chapter 3.3). Additionally, these statements were also built on the work of Agle et al. (1999) who analyzed the construct validity of the attributes. The survey can be found in Appendix 1.

(22)

19

During the first part of the interviews, rationales behind participant’s survey answers were detected in order to find out why they think that particular stakeholder groups are more or less powerful, legitimate than others and why different stakeholder groups’ claims are more urgent than others’. The second part of the interviews is introduced in Chapter 5.3.1 as it covered another topic. Summaries of these interviews can be found in Appendix 3.

Interview protocols were the same as by the previous ones (detailed in Chapter 5.1.1) with the exception that one interviewee was interviewed via Skype as he was abroad by that time. Summary of an interview was still not approved as Interviewee#3 was still on holidays by the time the final version of this thesis was handed in. Interviewee#1 did not have the chance to fill in the survey before the interview, so he was asked to send his answers via e-mail. When analyzing the interviews, majority’s opinion on the possession of stakeholder attributes was taken into consideration, it determined the classification of stakeholders.

5.2.2. Results

This chapter describes how stakeholder groups were classified by applying Mitchell et al.’s (1997) model. Based on the interview results, every stakeholder group was categorized in accordance with the number of attributes it possesses, results are shown on Image 6.

Most of the stakeholders were recognized as definitive stakeholders meaning they possess every attribute. Although all of these groups should get the highest attention, there are some differences within the groups. Board of the UvA/HvA (1) and the Steering Group ICT (2) are on the highest level of the internal hierarchical structure of the UvA, huge majority of the interviewees declared them as one of the most definitive stakeholders, because formally they have the final words on every major decision. Lots of interviewees emphasized that every organization within the UvA is legitimate, but hierarchical levels influence the level of legitimacy.

According to almost all interviewees, the two most important stakeholder groups were the EGE (10) and the LAFG (12). LAFG is the actual steering group of UvAInform project, they are deciding on every decision, however EGE is even more definitive as they can overrule decisions of LAFG anytime. They have the real power to cut budgets, allocate resources as bodies above them in the hierarchy make decisions mostly based on their advice. Every interviewee agreed on that every organization that is directly supporting the operational work of the project, possesses all attributes. These are the UvA ICTS (3), the LRS Steering Board (14), the pilot project managers (15) and the developers (16). ICTS and developers are especially important according to some of the participants, for instance, Interviewee#5 said that

“The project is hugely dependent on them technically, ICTS has a large number of experts and they are the ones deciding on technical feasibility of pilot projects”.

(23)

20

Every participant declared that data privacy regulation and support has significant role in the project, therefore UvA Legal Affairs (7) and UvA Ethics Committee (8) have the power and legitimacy to influence the project (even if they don’t use it at the moment, so they are rather ‘phantom’ definitive stakeholders) and of course their claims would be very urgently handled once they are more involved.

There were four groups classified as definitive stakeholders, but it was not that simple to put them in the ‘box’ as some interviewees either did not have enough information on the stakeholder group or there were different and opposite opinions about their attribute possession. UvA Academic Affairs (6) is an organization on a high hierarchical level within the university. According to Interviewee#10, they are definitive, because

“UvAInform has to fit in the academic direction of UvA, so Academic Affairs surely does the power to influence the outcome of the project, UvAInform has to take Academic Affairs’ claims seriously”.

On the other hand, few people said that they are too far away in the organizational structure to have serious influence on the project. Academic Affairs is also a phantom stakeholder group as it does not use its power. Most of the participants did not have enough information on how a potential cooperation with Uvadata (5) would be fruitful, but the ones who did, emphasized that involving managers of Uvadata could help accelerate the project. Consequentially, although they are not willing to cooperate at the moment, Uvadata does have power (even if they do not use it, they are ‘phatom’ stakeholders as well), legitimacy and their claims would be handled urgently. UvA teachers, professors (19) and UvA study advisors (20) are one part of pot ential end-users of the project. They are not fully involved so far, but according to the majority of interviewees, maybe not strongly, but they do possess all attributes, for instance Interviewee#10 said that

“Faculty representatives in LAFG sometimes talk with professors and study advisors, and this way maybe just very weakly, but they can indirectly influence decisions”.

Two groups were recognized as dependent stakeholders, which means they do not possess power, but they are legitimate and their claims are urgent for the management. UvA Portfolio Management (4) is high-level central body at the UvA, but

“They don’t have real power to influence the strategic direction of the project as they only have administrative functions”

according to Interviewee#4. As they are reporting to the UvA Board, they are legitimate and majority of participants said that their claims have to be managed urgently in order to get the project properly documented. Current UvA students (17) are dependent stakeholders as they had power in the project if they would be involved. Participants had different opinions on how urgently their claims would be managed, but more interviewees stated that they possess urgency even if not as most of the definitive stakeholders do. As they are the primary end-users of UvAInform, almost every interviewee would delegate power to them and this way make them definitive in the project. The difference between current UvA students and the ‘phantom’ definitive stakeholder groups is that while Uvadata, Legal Affairs etc. do have a the power, but at the moment they are not willing to use it, current student do not have the power, therefore they are dependent on the management of UvAInform, who could give it to them.

Three groups possess only the attribute of legitimacy (discretionary stakeholders). External expertise groups (9) do not have power to influence the project and their claims are also not being managed urgently enough according to the majority of participants. Although

(24)

21

they do not have as high legitimacy as internal organizations, they are proper stakeholders of the project according to most interviewees. Interviewee#9 stated:

“External expertise groups don’t really have claims in this project, however, management of UvAInform should consider their guidelines, research directions and findings much more often and in a more urgent way in order to contribute to LA research in a proper way. More than one person should be connected with them in order to better communicate the external community standards”.

Two other potential end-users were also recognized as discretionary stakeholders: UvA incoming students (18) and UvA educational program directors (21). As incoming students are not part of the organization of the university yet and educational program directors are only meant to be side users of the project, they only possess legitimacy.

Other expertise groups (9) and focus groups (11) within the UvA are non-stakeholders. They were potential stakeholders because they are competing for budget against EGE and LAFG, but this competition is not a real one, and they cannot influence the project, according to almost all of the interviewees. Future employers of UvA students and HR agencies (22) are still not real pot ential end-users, the will not be involved in the near future, said majority of the interviewees.

This chapter provided answer to the third sub-research question, which was “How are stakeholder groups classified in the project?”. Although interviewees agreed on the classification of some stakeholder groups more than on others, researcher was provided with enough information to classify all stakeholder groups in accordance with respondents’ answers.

5.3. Presence of success factors

5.3.1. Methodology

There were seven main factors defined in Chapter 2.2 that influence DW organizational- and project implementation success, namely management support, change management (managing organizational resistance), clear vision, champion, user participation, resources and team skills. Additionally, it was assumed that working with lots of privacy-sensitive student data will require proper legal and ethical support and regulation. Presence of these factors in the UvAInform project was measured in the second part of the core interviews described in Chapter 5.2.1. In the second part of the semi-structured interviews, participants were asked about the presence of each factor in the project and about the necessity of the factors in order to reach implementation success. They were also asked to provide suggestions on the improvement of factors within the project (these suggestions are detailed in Chapter 6). As previously mentioned, summaries of the interviews can be found in Appendix 3. Data was processed by comparing interviewees’ answers per factors and emphasizing majority’s opinion in the results section. Different, individual answers were also taken into consideration if they provided a new perspective that could contribute to the improvement of factors.

As the previously described interviews were conducted with stakeholders, all of them trying to suggest ways to get legal approval, an additional interview was conducted with a member of UvA Legal Affairs who is in charge for giving legal advice and approval in order to see the other side of the coin of legal aspects as well. The interview aimed to acquire more knowledge on the process of legal approval of UvA projects and to get another perspective why Legal Affairs is still not involved in UvAInform. An interview with the project manager of Uvadata was organized with the intention to clear up the relation between Uvadata and the UvAInform project and to investigate why negotiating did not work out so far. As it was

Contribution of normative stakeholder theory to the improvement of factors affecting data warehousing implementation success