To what extent do institutional pressures impact the intention to increase cloud computing adoption similarly for private and public organizations, and is

(1)

To what extent do institutional pressures impact the intention to increase

cloud computing adoption similarly for private and public organizations, and is

this relationship affected by IT decision makers’ attendance to international

gatherings?

Author: Janiek Touw

Master program: MSc. International Business & Management Student number: S3241564

Email: j.f.a.touw@student.rug.nl Word count: 15002

Faculty of Economics and Business University of Groningen

Duisenberg Building, Nettelbosje 2, 9747 AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands

http://www.rug.nl/feb

(2)

Abstract

Despite the growing interest of the emerging technology Cloud Computing (CC), little is known about how external factors influence the intention to increase CC adoption of private sector organizations similarly as public sector organizations. Based on institutional theory, this thesis develops a theoretical framework that explains to what extent external factors lead to an increase in the cloud adoption in private sector organizations as well as public and if this relationship is affected by the attendance of international gatherings of the IT decision maker. The hypotheses were tested using survey data from 86 private and public sector organizations in the Netherlands. Results from hierarchical regression analyses suggest that coercive customer pressure positively affects the increase in the CC adoption intention in the private sector and mimetic pressures positively affects the increase in the CC adoption intention in the public sector. Surprisingly, no support for the hypothesis of normative pressure influencing the increase of the CC adoption intention has been established. Moreover, no moderated effect of the attendance at international gatherings of IT decision makers was found. The findings provide important managerial implications for cloud providers.

(3)

Acknowledgements

(4)

List of Figures

Figure 1 The conceptual model ... 24

List of Tables

Table 1 Descriptive statistics and correlation matrix Private sector ... 34

Table 2 Descriptive statistics and correlation matrix Public sector... 34

Table 3 Regression results - Private sector ... 36

(7)

List of Acronyms

CC Cloud Computing

CP Coercive Pressure

CTO Chief Technology Officer

E-government Electronic government

ERP Enterprise Resource Planning

ICT Information and Communication Technologies

IS Information Systems

IT Information Technology

KMO Kaiser-Mayer-Olkin

MNE Multinational Enterprises

MP Mimetic Pressure

NP Normative Pressure

PCA Principal Component Analysis

(8)

1. Introduction

For more than a decade, information and communication technologies (ICT) have been undergoing rapid changes, and companies are deciding on adopting new technologies. Cloud computing (CC) is especially one of the most important technological drivers in the digitization of the business environment (Hentschel, Leyh, & Petznick, 2018). In November 2019, Gartner reported that by 2022, up to 60% of organizations will use an external service provider’s cloud managed service offering, which is double the percentage of organizations from 2018 (Costello & Rimol, 2019). A CC environment helps companies to achieve business processes transformation, reduced information technology (IT) expenditures, using real-time on-demand applications, ubiquitous storage, and unlimited computing power (Oliveira, Thomas, & Espadanal, 2014). As a result, both public and private sector organizations have been undergoing a similar trend of the emerging development of CC (Hentschel et al., 2018) and considering IT investment in CC services because of its business efficiencies (Low, Chen, & Wu, 2011). Organizations that already use CC services even decide to increase their level of CC adoption because cloud providers continuously develop and look for new, better-performing solutions for the CC environment (Cearnău, 2018). Even though the cloud environment could provide benefits for public organizations as for private organizations (Diez & SIlva, 2013), it is interesting to note that private sector organizations are embracing the CC relatively quick while public organizations are lagging (Shin, 2013).

(9)

argue that the differences in the external environment of the two sectors have narrowed (Rocheleau & Wu, 2002).

The external environment can have a direct effect on the firm’s decision (Hasty, Schechtman, & Killaly, 2012). To examine in more detail the effect of the external environment on the intention of organizations to increase the level of CC adoption, the institutional theory is used in this thesis. The institutional theory assumes that organizations are motivated to comply with external formal and informal pressures to achieve legitimacy (Dimaggio & Powell, 1983). Legitimacy can be defined as ‘the acceptance of the organization by its environment’ (Kostova & Zaheer, 1999: 64) which is critical for organizations to succeed and survive (Meyer & Rowan, 1977). Three external pressures are identified by DiMaggo & Powell (1983) and used to understand the effect of these pressures on the decision to increase CC adoption: Coercive pressures are exerted either formally or informally by other organizations upon which one is dependent; normative pressures result from norms defined by institutions such as professional networks; mimetic pressures result as organizations respond to uncertainty by copying the actions of other organizations (DiMaggo & Powell, 1983).

Most of the literature concerning cloud computing is focused on factors affecting cloud adoption in just a single industry (Oliveira & Martins, 2011; Maqueira-Marín, Bruque-Cámara, & Minguela-Rata, 2017; Gao & Sunyaev, 2019). However, understanding the external factors impacting the CC adoption intention in both private and public sector organizations could contribute to the existing literature. More specifically, there is no clear consensus in the literature if there is a difference in the effect of external pressures on the IT decision of private and public organizations. On the one hand, there is reason to belief that the external pressures do have a different effect as, for example, Rocheleau and Wu (2002) found that private sector organizations are willing to invest more in IT than public sector organizations due to their competitiveness. On the other hand, it can also be argued that the external pressures are similar for both sectors. Kelly (2007) stated that the behavior of public sectors changed inevitably because more and more public IT services are offered to outsourcing, and therefore competition will increase. This inconsistency makes it interesting to examine external pressures impacting the adoption intention to increase the level of CC services of private and public organizations.

(10)

gatherings are forums for social learning. Temporary gatherings encompass formally organized in-person events lasting between several hours and several days (Fang et al., 2020). In the context of CC, such events are often provided by multinational enterprises (MNE) such as Microsoft, Gartner, and Amazon in the form of IT trainings and IT events (CloudTech, 2020). As such, most of these gatherings are internationally oriented and information is exchanged among participants from different countries. The exposure to different cultures offers managers access to a diverse and larger number of ideas, concepts and inputs that push individual’s existing knowledge (Godart, Maddux, Shipilov, & Galinsky, 2015). Therefore, an IT manager attended an international conference has a larger bundle of CC knowledge and skills. This result in a low uncertainty over CC services and thus the tendency to mimic the actions of others will be less strong. Therefore, mimetic pressure on the cloud adoption intention is negatively moderated by the attendance of an international gathering.

Given the topic at hand, current thesis addresses the following research question: ‘To what extent do institutional pressures impact the intention to increase cloud computing services similarly for public and private organizations, and is this relationship affected by IT decision makers’ attendance to international gatherings?’ It empirically examines the research question by doing a survey study among 86 firms originating from the private and public sector already using CC services pressured by coercive, mimetic and normative factors and moderated by the attendance of an IT decision maker at an international gathering.

This study presents the following contribution: first of all, research about CC is relevant because it is an extremely current subject as it is an emerging technology (Hentschel, et al., 2018). Secondly, there is no clear consensus in the literature about studies of CC in both public and private organizations. A growing number of studies has examined the antecedents of CC adoption, but often focused on only a single sector. This paper addresses this gap and focuses on the influence of institutional factors that increase the adoption intention of CC. Thirdly, this study sheds light on the global nature of CC services by examining external pressures in an international gathering context. Finally, this research could provide useful insights for cloud providers to understand which institutional factor impact the adoption intention of IT decision makers.

(11)

(12)

2. Literature review

2.1 Cloud Computing adoption intention

The adoption process is a sequence of stages that the adopter of an innovation passes through before acceptance of a new product, service or idea (Rogers, 1995). The outcome of the process is a decision, which is made by an entity on a specific object in a particular context (Mohammed, Ibrahim, Nilashi, & Alzurqa, 2017). In the context of this study, the intention to increase future CC usage (the object) by a public sector organization (the entity) and a private sector organization (the entity) are investigated. An initial acceptance or purchase decision is made, an evaluation of the technology is carried out, and a decision to increase future usage of the product or technology can be made (Obal, 2017). As cloud providers are constantly looking for new, better-performing solutions in their cloud environment, organizations have the need to obtain these new solutions which can improve their business (Cearnău, 2018).

IT has become a necessity for many organizations, but the traditional in-house IT model is increasingly being challenged by the CC delivery model (Yigitbasioglu, 2014). A widely used definition of cloud computing is stated by National Institute of Standards and Technology (NIST), “Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned.” (Mell & Grance, 2011: 2). The cloud provides a new framework for organizations to develop their IT competence (Oliveira, Thomas, & Espadanal, 2014). It virtually eliminates the need to own and maintain expensive hardware and software by allowing access to IT resources over the Internet (Yigitbasioglu, 2014).

2.2 Cloud Computing examples

(13)

Royal Dutch Shell, the global leader in oil and gas industry, has operations in more than 70 countries. To meet the world’s growing energy demand, their software development team works hard to keep its specialized software up to date and enhanced with new features. The productivity can be hindered by disparate locations and technologies. A more streamlined, centralized solution to support application development was required. They decided to deploy their IT environment in the cloud. The Chief Technology Officer (CTO) stated that moving to the cloud is the organizations’ strategy to obtain a much more agile, fast-moving environment than they could with a local data warehouse. It resulted in shorter lead cycles and better involvement of business stakeholders so that their solutions are more aligned with business requirements (Microsoft, 2019a).

Moreover, the municipality Hollands Kroon considered digital transformation as a way to improve municipal efficiencies and citizen services through economies of scale, boosting citizen satisfaction, employee morale, and reducing costs. Therefore, they implemented a workplace transformation solution by moving its self-managed municipal IT datacenter model to Azure. They found new and better ways to gain insights from the city’s digital data. Consequently, the municipality can better deliver services to its citizens by having a more mobile work force (Microsoft, 2019b).

2.3 Cloud computing for public and private organizations

(14)

of both sectors is investigated more closely, as it can have a direct effect on the firm’s decision (Majumdar & Venkataraman, 1993).

2.4 External environment private sector & public sector

On the one hand, there are fundamental differences between organizations in the public and private sectors (Kenny, Kenny, Butler, Cray, & Wilson, 1987; Rainey et al, 1976). The environmental factors accentuate the differences between private and public sectors (Caudle, Gorr, & Newcomer, 1991) as public services function within a bureaucracy and the private services are driven by market imperatives (Parker, et al., 2013). It is argued that bureaucracies and the market are completely opposing forms of operating environments. Whilst bureaucracies rely on rigid hierarchies, the market relies on end-user choice (Hughes & O’Neill, 2008). This results in public sector organizations having less market exposure and therefore less scope for explicit incentive mechanisms for productivity and effectiveness enhancements, but at the same time they have more legal and formal constraints (Campbell et al., 2010). Another difference between the two sectors is the competition of the market. In the private sector there are a multitude of private organizations that are competing for customers whilst in the public sector there is a single public organization that offers public services. If private sector companies are not perceived to put customers first and continuously improve their offerings, then their customers will get their services elsewhere. However, this is not the case, in the public sector because public services are often free or represent a monopoly in their field which result in the customer having no alternative (Parker, et al., 2013).

(15)

between private and public organizations are blurring. To further examine to what extent the external environment influences IT adoption decisions of private sector organizations similarly as public sector organizations, the institutional theory is used.

2.5 Institutional theory

In the recent decades, the institutional theory is commonly used to explain the influence of external institutions on organizational decision making and transformation. The institutional theory assumes that organizations are motivated to comply with external formal and informal institutions (Dimaggio & Powell, 1983). Institutions are generally described as ‘humanly devised constraints that structure political, economic, and social interaction’ (North, 1991: 98). According to North (1991), institutions serve as the ‘rules of the game’ through their influence on the transaction costs of business activities. Their effect on a firm’s decision-making process and behaviors is significant (North, 2005). Institutional actors that exert pressures include the state, professional, interest groups and public opinion (Oliver, 1991). The theory claims that firms converge due to isomorphic organizational pressures, stemming from uncertainties or in order to obtain legitimacy, resources, and success (Dimaggio and Powell, 1983; Dhalla & Oliver, 2013). Isomorphism is the process that forces one unit to be similar to other units in the same environmental context (Dimaggio and Powell 1983): firms in the same sector tend to become homologous over time (Oliveira & Martins, 2011). The decision of a focal organization to adopt a technology is thus vulnerable to external influences embedded in the institutional environment (Teo, Wei, & Benbasat, 2003).

2.6 Using the institutional theory to explain cloud computing adoption

(16)

deliver services electronically. The results of the research demonstrate that, with regards to e-government adoption, the institutional pressures of public administration organizations affect their Top Management Commitment which in turn impacts the intention to adopt e-government. In other words, the relation between institutional variables and the adoption of IT have led to significant insights regarding the influential roles of external pressures on both public and private organizations (Zheng et al., 2013; Teo et al., 2003; Liang, Saraf, Hu, & Xue, 2007).

(17)

3. Hypothesis Development

3.1 Coercive pressure

Coercive pressures result from formal or informal pressures exerted on organizations by other organizations upon which they are dependent such as regulatory bodies which are built into exchange relationships (Dimaggio and Powell, 1983). When an organization enters into an exchange relationship that runs counter institutionalized patterns, it would be difficult to maintain the relationship, or it could even become unsustainable (Teo et al., 2003). Thus, organizations that are strongly dependent on the institutionalized environment are more likely to to change its IT and adopt new IT if necessary (Teo et al., 2003; Liang et al., 2007; Krell, Matook, & Rohde, 2016).

In the context of CC adoption intentions for private organizations, there has been argued that coercive pressures stem mainly from: the parent corporation, dominant customers, dominant suppliers, and the government (Teo et al., 2003; Krell et al., 2016).

First of all, DiMaggio and Powell (1983) noted that subsidiaries are required to conform to practices and structures that are compatible with the policies of the parent corporation. Hence, it could be suggested that parent corporations that have adopted certain cloud solutions are likely to exert pressure on subsidiaries to do likewise. Parent corporations with foreign subsidiaries can also exert pressure on these subsidiaries in order to enhance system efficiencies and reduce bank and foreign currency conversion costs (DiMaggio and Powell, 1983; Liang et al., 2007). This is supported by Yigitbasioglu (2014) arguing that organizations are likely to follow the behavior of parent corporations that have adopted cloud.

(18)

solutions to acquire legitimacy or status, or to demonstrate their fitness to do business with these dominant organizations (Teo et al., 2003).

Supplier dependence arises when organizations are unable to switch to alternative suppliers. For much of the purchases of organizations they rely on existing suppliers (Hart and Saunders 1997; Webster 1995). The suppliers, the resource-dominant organizations, that have adopted a certain technology, would attempt to influence their resource-dependent trading partners to the technology so as to increase their own benefits of adoption (Teo et al, 2003). If the supplier requires the focal firm to adopt certain CC services, the focal firm is feeling pressured to do so (Krell et al, 2016).

Finally, policy and regulatory environment is influencing the IT adoption decisions (Zhu, Kraemer, Xu & Dedrick, 2004). Governments can encourage the adoption of certain cloud services with policies (Lian, Yen, & Wang, 2014). For example, when a segment of the market is monopolized, regulators can ease potential integration bottlenecks that can arise if the monopolist creates barriers to discriminate against competitors. In this way, the government can enable other organizations to adopt certain IT services (IS) (Dedrick, Venkatesh, Stanton, Zheng, & Ramnarine-Rieks, 2015). Compliance with laws and government regulations often require firms to make changes to their IS, or even adopt completely new IS (Krell, & Matook, 2009).

A firm depending strongly on institutions in the environment in terms of the parent company, customers, suppliers and regulatory bodies, will results in a stronger coercive pressure, and will therefore be more inclined to adopt IT if necessary (Krell et al, 2016; Liang, et al, 2007; Teo et al, 2003). The present thesis therefore hypothesizes the following:

H1a: Greater coercive pressures in terms of parent firms’, customers’, suppliers’, and regulatory bodies demands leads to the higher likelihood of an increase in the level of cloud adoption for private organizations

(19)

Public sector organizations have less market exposure than private sector organizations (Rainey et al., 1976) which result in less incentive mechanisms for productivity and effectiveness increase (Campbell et al., 2010). However, public sector organizations have more legal and formal constraints compared to private sector organizations (Sethibe, Campbell & McDonald, 2007). They are driven by compliancy, regulatory frameworks and have highly structured internal functional boundaries (Parker et al., 2013). The result is that organizations at lower levels of government are dependent on the resources from higher levels of government in the form of democratic legitimacy, legal mandates, and financial means (Andersen & Jakobsen, 2018). This can give rise to coercive pressure because the wish of the political principals is linked to the organization’s dependency on the resources from the political principals (Andersen & Jakobsen, 2018). The Taiwanese government is an example of a government stimulating the adoption of CC services. Determination of best approach to avoid wasting medical resources is one of the main objectives of the Taiwanese government. The government enforce the promotion of the cross-hospital electronic health record interchange, which is a CC technology. This results in hospitals viewing CC as an enabling technology and a necessary investment (Lian, et al., 2014). Hence, there has been argued that superior organizations can exert pressure on subordinates to do adopt a certain technology (Zheng, et al., 2013).

Moreover, dependence on upstream and downstream organizations arises when the focal organization relies on other organizations for much of its input or output information (Zheng, et al., 2013). Often there is an interconnection of the business of public sectors across units and departments (Ramon Gil-Garcia, Chengalur-Smith, & Duchessi, 2007). Therefore, input information in one organization is often the output of other organizations. Hence, upstream and downstream coordinators that have adopted certain CC services are likely to exert pressure on focal organizations to do likewise (Zheng et al., 2013). Zheng et al. (2013) found that when decision makers feel a larger need of IT usage from other organizations, they are more committed to supporting future adoption.

(20)

services, provide services with high quality, quick responses, connectivity and informativeness (Oliveira & Welch, 2013; Shareef, Dwivedi, Kumar, & Kumar, 2016). Although they are not involved in every decision, even in public sector organizations is the making of major decision heavily influenced by customer considerations (Kenny, et al., 1987). This thinking results in the following hypothesis:

H1b: Greater coercive pressures in terms of superior organizations, upstream and downstream organizations’ and customers’ demand leads to the higher likelihood of an increase in the level of cloud adoption for public organizations

3.2 Normative pressure

Once an innovation like CC becomes available, normative pressures emerge (Yigitbasioglu, 2014). Normative pressures determine the direction of actions by prioritizing conformity towards norms (Scott, Scott, & Meyer, 1994). These norms manifest themselves through professional, trade, business organizations (Powell and DiMaggio, 1991). Organizations are expected to meet to standards of professionalism and to implement certain systems that are considered legitimate by relevant professional groups (Kung, Cegielski, & Kung, 2015). Teo et al (2003) found support for normative pressures influencing technology adoption in the private sector. They argue that norms espoused by the business and professional communities play an important role in influencing organizational decision makers’ intention to adopt the technology. In the context of public sector organizations, the organizations are also largely affected by normative pressures (Zheng et al, 2013). For example, in healthcare where regional networks can exert normative pressures to others within that region (Sherer, Meyerhoefer & Peng, 2016). Key decision makers identify with a particular professional or industry association and subsequently engage in activities to comply with the norms defined by the association (DiMaggio & Powell, 1983; Krell et al, 2016). This results into the following hypothesis: H2a: Greater normative pressures in terms of professional or industry associations will lead to a greater intend to increase the level of cloud adoption for private organizations

(21)

3.3 Mimetic pressure

Mimetic pressures on a focal organization refer to the isomorphic pressures resulting from its peer organizations (DiMaggio & Powell, 1983). To acquire legitimacy, respond to uncertainty, or social fitness, in a wider social structure, an organization usually imitates the choices made by other similar organizations, especially successful organizations (Teo et al, 2003; Liang, et al., 2007; Yang & Hyland, 2012).

In the context of private sector, mimetic pressure might occur when firms face uncertainty regarding adoption of a new unfamiliar CC service (Yigitbasioglu, 2014). Due to this uncertainty, a firm imitates behavior performed by a seemingly successful organization in the firm’s environment. Managers are inclined to mimic the actions of their successful peers since it shields them against potential loss of face and helps to maintain the legitimacy of their decisions (Liang et al., 2007). Generally, if decision makers believe that a behavior was successfully performed by similar institutions, it is easy to imitate those organizations because the chance of success seems higher. Consequently, firms are likely to mimic organizations that operate in similar markets, use similar resources, or sell similar products (Teo et al., 2003). Thus, behaviors performed by similar organizations are perceived to be appropriate for a firm that engages in the imitation of other firms (DiMaggio & Powell, 1983). Resulting in the following hypothesis:

H3a: Greater mimetic pressures in terms of the number of similar organizations adopting cloud computing services will lead to a greater intend to increase the level of cloud adoption for private organizations

(22)

organizations that are generally high in risk aversion and low in entrepreneurial activities (Bozeman & Kingsley, 1998). Therefore, this thesis hypothesizes:

H3b: Greater mimetic pressures in terms of the number of similar organizations adopting cloud computing services will lead to a greater intend to increase the level of cloud adoption for public organizations

3.4 International gatherings

When IT decision makers are not familiar with a technology, they can decide whether or not they want to attend an event to understand the opportunities these technologies can bring for their business. In the case of CC, these events are hosted by CC experts and specialists and are focused on providing attendees with insights and benefits of the technology. Interestingly to note is that these events are often in an international context as CC is a global infrastructure and concerns companies from all around the world (Rittinghouse & Ransome, 2016). IT decision makers from different countries come together at a gathering learning from IT experts about CC. In addition, the participants also learn from each other as attendees observe and influence one another, by exchanging knowledge and sharing experiences on how to adopt and use a technology (Fang, et al., 2019). Especially at these events, when their social learning is stronger due to the uncertain situation (Gaba & Dokko, 2016), the international context could influence the adoption intention of IT decision makers.

(23)

in a difference in the decision-making process of managers across countries due to their national cultural background (Azam, Boari, & Bertolotti, 2018).

At international gatherings, there is a process of communication and social influence between IT decision makers with different cultural backgrounds, as they share knowledge and experiences. International attendees talking to one another, results in meeting different cultural environments (Glazer & Karpati, 2014). Managers exposed to different environments offers them access to a diverse and larger number of ideas, concepts and inputs that push individuals to questioning the status quo and their existing knowledge. Moreover, it allows managers to look for a higher number of approaches of a specific matter (Godart, et al., 2015; Azam, et al., 2018). For example, China has the highest spending on cloud services among countries in the Asia Pacific region and they are still growing rapidly in their cloud usage (Statista, 2020). Moreover, over half of the enterprises in Finland (65%), Sweden (57%) and Denmark (56%) use CC. Whereas enterprises in the Netherlands uses for 47% CC (Smihily & Kaminska, 2018). IT decision makers from China and Finland could improve the understanding of possibilities and benefits of CC services of a Dutch IT decision maker from their own CC experiences. In this way, they improved their CC knowledge by attending an international gathering. Therefore, the sense of uncertainty regarding CC services and the urge to mimic successful others will be lower than an IT decision maker who did not attend an international gathering. To summarize: H4: Mimetic pressures for private and public sector organizations will have a less significant impact on intention to increase the level of cloud adoption when the decision maker attended an international gathering than when s/he did not.

3.5 Conceptual model

(24)

both private and public organizations. While private sector organizations impacted by competitors and public sector organizations impacted by peers, Hypothesis 3a and 3b expects that mimetic pressure leads to a higher intention to increase the level of CC adoption. Hypothesis 4 proposes that H3a and H3b are negatively moderated when the IT decision maker attended an international gathering; attendance improves CC knowledge through exchanging experiences with international others which decreases uncertainty and therefore the need to mimic successful others is lower.

Figure 1 - The conceptual model

(25)

4. Methodology

The previous chapter discussed institutional theory in the context of private and public sector organizations. It outlined how private and public sector organizations can be impacted by the external environment which in turn can impact the cloud adoption intention. In relation to this, hypotheses have been formulated. The current chapter explains the data and

methodology to empirically test the raised hypotheses. 4.1 Design

In order to examine the research model in Figure 1 and the associated hypothesis above, a quantitative data study is set up. The empirical method to collect data was a quantitative survey of customers of a single cloud provider1_{. The survey method was chosen as Fink (2003) stated} that it is one of the most often used techniques of collecting information from people to describe, compare, explain, or predict their knowledge, attitudes, or behaviors (Fink, 2003). A dataset was generated using a survey questionnaire distributed through the software Qualtrics. The software Qualtrics is chosen because it is a widely used software (Mathur & Reichling, 2019). Using an online survey in order to generate data has several advantages compared to other research instruments such as interviews, focus groups, or observation. The most notable advantages are both the speed and the quantity of the data collection process (Fleming & Bowden, 2009), the data accuracy, the possibility of respondents to remain anonymous (Hooley, Wellens, & Marriott, 2012), and the ease of distribution (Phillips, Phillips, & Aaron, 2013).

However, this form of data collection also has disadvantages. For instance, Auger and Devinney (2007) argue that when using online surveys as the method of data collection, respondents are likely to overestimate their actual preferences and intentions, meaning that social desirability bias is likely to occur. This bias entails that participants have a tendency to

1_{This thesis has been produced in terms of an agreement with an outside body governing the supply of confidential material and}

(26)

deny behaviors that they are assumed as ‘socially undesirable’ and report behaviors that they assume to be ‘socially desirable’ (Zerbe & Paulhus, 1987). This means that there could be the tendency to consciously report exaggerated self-descriptions in order to create a socially desirable image in front of an audience (Hart, Ritchie, Hepper, & Gebauer, 2015). This could especially the case in this research as only the customers of a single cloud providers are asked to fill in the survey about the same cloud provider technology. Therefore, in order to control for the social desirability, the participants were ensured that all of the information collected during this study would be kept private. The researcher was being disconnected from the data collection resulting in the avoidance of bias and other noises. An anonymous online survey was also conducted to make sure the researcher could not influence respondents (Firestone, 1987). In this way, ethical issues associated with the use of online surveys were prevented (Hooley et al., 2012).

4.2 Sample and data collection procedures

The survey was administered to IT managers and IT decision makers of the Netherlands which were directly involved in their firms’ cloud computing adoption decisions. In prior research it was demonstrated that IT managers and IT decision makers possess knowledge of success determinants, knowledge of project outcomes and knowledge of adoption motives (Cragg, King, & Hussin, 2002). Therefore, it can be expected that they are competent to assess CC adoption decisions for the purpose of this research.

The questionnaire was sent to participants by email with attached instructions and description of the study. After agreeing to participate they were asked first to indicate details about their role and about the organization such as firm size, IT department size and firm age organization. To ensure face validity the questionnaire was intensively discussed with four different IT practitioners of both the private and the public sector and a cloud specialist was invited to review the questionnaire. Based on these insights and feedback for improvement, the questionnaire was modified and finalized. According to Punch (2003) a questionnaire should not be longer than 20 minutes because a survey that takes too long can be problematic. It may have discouraging effect, and thus influence the response rate. The survey of this study took on average 5 minutes to complete.

(27)

ensure the meanings were retained in the translation. The reverse-translated version of the items in English was then compared with the original items to check if the meanings were preserved.

All of the data were collected in a month, starting on the 18th of November until the 21st of December 2020. Of the 152 questionnaires distributed, in total 119 responses of which 47 responses for the private sector and 72 responses for the public sector. In total, 12 respondents did not adopt some form of CC of which two in the private sector and 10 in the public sector. After gathering responses, 21 respondents were removed due to improper completion of the survey (e.g., missing answers, survey completed too quickly) resulting in 86 questionnaires usable for data analysis of which 39 in the private sector and 47 in the public sector. The total response rate is 56,58%. A nonresponse bias using t-tests tests is assessed to compare the early and late respondents on the dependent variable and demographic characteristics. It assumes that late respondents in a sample are similar to theoretical non-respondents (Armstrong & Overton’s, 1977). For both the private and public sector data gathered in the first two weeks is compared with the data gathered in the last two weeks. No significant differences (p > 0.05) were found in the data of both sectors, and thus non-response bias is not likely to be a concern of this study. The data collected was analyzed by using IBM SPSS 27.0 software.

The data indicates that most of the respondents in the private sector were IT managers, 48.7% of the respondents. In addition, 41.0% of the respondents worked in ‘other sectors’ than the options given. Examples were gambling, services, retail, and several sectors. The majority of the respondents had a firm size of 101 until 500 employees and existed for longer than 25 years: 33.3% and 76.9% respectively. Lastly, 48.7% of the respondents had an IT department size less than 10 employees.

(28)

4.3 Initial measures

Measurement scales were derived from previous studies in order to ensure content validity. Content validity is defined as “the degree to which items in an instrument reflect the content universe to which the instrument will be generalized” (Straub, Boudreau, & Gefen, 2004: 424). For some of the constructs multi-item measures were used which are analyzed using factor analysis discussed in section 4.4. Afterwards, summated scales (factor scales) were created by taking the average score of the variables. For a complete overview of the scales, see Appendix B.

4.3.1 Dependent variable

Intention to increase the level CC. The dependent variable, the intention to increase the CC adoption (AI), is measured using the scale of Gewald and Dibbern, 2009. This scale is also used in previous studies (Benlian & Hess, 2011; Yigitbasioglu 2014) and measures to what extent an organization wants to increase future usage of CC. Some modifications were made to the existing scales to make them more relevant to the context of CC adoption of a single cloud provider. It consists out of two items: “Our organization should increase the existing level of CC services.” “I support further adoption of CC services.”. All items were measured using a 7-point Likert scale (Likert, 1932), ranging from score 1 ‘strongly disagree’ and score 7 referring to ‘strongly agree’.

4.3.2 Independent variables

(29)

consisting out of three items adapted from Liang et al., (2007). All of these items were measured on a 7-point Likert scale (Likert, 1932), ranging from score 1 ‘strongly disagree’ and score 7 referring to ‘strongly agree’.

Coercive pressure in the public sector was operationalized by three sub-variables treated as independent variables: (1) Coercive pressure from superior organizations (CP-SO) consisting out of three items adapted from adapted from Teo et al (2003) and Zheng et al. (2013) (2) coercive pressure from other organizations (CP-O) consisting out of three items adapted from Teo et al (2003) and Zheng et al. (2013) and (3) coercive pressure from customers (CP-CU) consisting out of two items adapted from Teo et al (2003) and self-developed. All items were measured using a 7-point Likert scale (Likert, 1932), ranging from score 1 ‘strongly disagree’ and score 7 referring to ‘strongly agree’.

Normative pressure (NP). Normative pressure was measured for both sectors in the same way. It measures to what extent organizations are impacted by existing norms in a professional association (Teo et al, 2003) in their intention increase their CC adoption. The variable consisted out of one item. It was adapted from Krell et al (2016) e.g. The extent to which my firm’s decision to use the system was affected by promotions by industry, trade, or professional bodies is measured on a 7-point Likert scale, ranging from 1 ‘very low’ to 7 ‘very high’.

(30)

4.3.3 Moderated variable

International gathering (IG). The experience of an IT decision maker whether or not attending an international event is described as ‘international gathering’. One item is used to examine this concept: ‘Did you participate at an international gathering regarding CC services?’. This item is measured as a dummy variable, coded ‘one’ for respondents who did participate at an international gathering, and ‘zero’ for the respondents who did not. In order to make the interpretation of the impact of the interaction effects easier, the interaction effect was mean centered (Cohen & Cohen, 2013).

4.3.4 Control variables

The introduction of control variables to the regression model is an attempt to separate their effects from those of the independent variables of interest and thus, remove their effects on the equation (Bryman, & Buchanan, 2018). Therefore, following best practices in research, firm size, IT department size and firm age are included as control variable in this research.

Firm size. Organization size is measured through number of employees. It has been found to have a positive influence on adoption behavior (Rogers, 1995). Large organizations are more likely to adopt CC services than small organizations because they possess the resources and the skills necessary to assimilate that innovation effectively, and they possess the economies of scale in transactions to leverage their investment in CC (Teo et al., 2003). Hence, firm size is included as a control variable.

IT department size. IT department size represents the technical resources an organization possesses to effectively assimilate an innovation. Technical resources have been found extremely important in adoption of technological innovations because the larger the department size, the broader the technological knowledge base of the organization for introducing and deploying information system innovations (Teo et al., 2003).

(31)

4.4 Measurement validity and reliability analysis

To ensure validity of the variables in the data set factor analyses and Kaiser-Meyer-Olkin (KMO) scores will be assessed. Next, a reliability analysis will be conducted by testing for the Cronbach’s alpha to measure the internal consistency of the final set of scale items.

With the purpose of verifying construct validity, a factor analysis for each variable was conducted utilizing principal component analysis (PCA). Construct validity refers to how well a construct which is a concept, idea, or behavior is transformed into a functioning reality (Taherdoost, 2016). With the research of Hair, Black, Babin & Anderson, 2010, variables with communalities higher than 0.30 is considered to be significant. All of the communalities met the threshold with the exception of one of the mimetic items of public sector ‘The proportion of my firm’s competitors that use similar systems is’ (0.083). It was decided to exclude this variable from the dataset and re-run the analysis without it. All of the items have a high loading, which means that the items in the same group were highly interrelated and are able to measure the corresponding construct (see Appendix C). In addition, the Kaiser-Mayer-Olkin (KMO) and Bartlett’s test of sphericity should have a minimum value of 0.50 (Kaiser, 1974). When the KMO threshold of 0.50 is not met, it indicates that distinct and reliable factors cannot be produced. All of the KMO scores are high, with the exception of ‘the intention to increase CC adoption’ in private and public sector and the ‘coercive customer pressure’ in the public sector. However, the threshold of 0.50 is been met which indicates that the factors are distinct and reliable (see Appendix C). For ‘coercive parent pressure’, both ‘normative pressures’ and the moderator ‘international gathering’, no factor analysis was needed because only one item was used to measure the variable.

After the resulting factor analysis, the reliability analysis was assessed by testing for the Cronbach’s alpha. The Cronbach’s alpha indicates the degree to which the items that are supposed to measure one underlying concept consistent with each other. Reliable results are reached when all the items related to a factor measure the same concept. The items capture a concept satisfactory, if the Cronbach’s alpha for each factor is higher than 0.6 (Malhotra, 2010). All items meet the criteria with the exception of the coercive customer pressure (α = 0.217) of the public sector. This item was remained to ensure content validity (Taherdoost, 2016).

(32)

detailed overview of the results for each item of the scales in detail. Still, some measures are explained, starting with coercive pressure of the private sector.

Private sector: Coercive customer pressure. The KMO measure (0.691) shows that the internal correlation between the variables is high. All of the factor loadings are above the threshold of 0.50 (Hair, et al., 2010) and one factor captures 70.37% of the variance in the original variables. The Cronbach’s Alpha of coercive customer pressure (α =0.786) surpasses the required threshold of 0.6 (Malhotra, 2010) and therefore, confirms the reliability of the scale items.

(33)

5. Results

5.1 Descriptive statistics

The new averaged variables are formed to be used for further analyses. Tables 1 and 2 present an overview of the descriptive statistics of the different variables, including the correlations between the variables.

Private sector. Table 1 shows that the overall adoption intention to increase CC services of the respondents was relatively high (M=5.22). Among coercive pressures the pressure of customers is the highest (M=4.29) whereas the pressure of the government is the lowest (M=1.83). Moreover, the mimetic pressure is strong (M=4.35) whereas the normative pressure is moderate (2.92). Correlation analysis has been done to see the significant interrelationships among the variables used in this study. It shows that correlations of coercive customer pressure (0.684), coercive supplier pressure (0.474) and mimetic pressure (0.629) with the intention to increase CC adoption are significant and relatively strong. Whereas the correlations of coercive parent pressure (-0.033), coercive government pressure (0.163), normative pressure (0.273) and international gatherings (-0.032) with the intention to increase CC adoption are not significant and relatively low.

Public sector. For the public sector was the overall adoption intention to increase CC services also relatively high (M=5.18) as shown in table 2. Among coercive pressures is customer pressure is the highest (M=4.83) whereas the pressure of superior organizations is the lowest (M=3.86) but still relatively high. The normative pressure was the lowest (M=3.53) amongst the pressures and the mimetic pressure was relatively high (M=4.62). In addition, also for the public sector an initial look at the correlations between the dependent and independent values with respect to the hypotheses are analyzed. It shows that correlations of coercive superior pressure (0.361), coercive pressure of other organizations (0.357) and mimetic pressure (0.513) with the intention to increase CC adoption are significant. Whereas the correlations of coercive customer pressure (0.221), normative pressure (0.183) and international gatherings (-0.104) with the intention to increase CC adoption are not significant and low.

(34)

analyses might not produce the results that were expected which will be further discussed in the hierarchical regression part.

Table 1- Descriptive statistics and correlation matrix Private sector

Mean SD 1 2 3 4 5 6 7 8 9 10 11 Dependent variable 1. AI 5.22 1.26 1 Independent variables 2. CP-P 0.38 0.49 -.033 1 3. CP-C 4.92 1.29 .684**_.191 ₁ 4. CP-S 4.78 1.08 .474**_.272 _.619**₁ 5. CP-G 1.83 0.96 .163 .328* .115 .242 1 6. NP 2.92 1.66 .273 -.124 .083 .144 .244 1 7. MP 4.35 1.24 .629**_.090 _.676**_.598**_.324*_.093 ₁ Moderator 8. IG 0.49 0.51 -.032 .031 .254 -.113 -.113 .014 .254 1 Control Variables 9. Firm Size 3.59 1.23 .354 -.404*_{-.001 .045} _.244 _{-.196 -.027 -.220 1} 10. IT department Size 2.51 1.93 .174 -.166 .067 .209 .190 .078 .322 *_.061 _.169 ₁ 11. Firm age 4.54 0.94 .279 -.285 .097 -.043 .260 -.124 .152 -.123 .719**_.235 ₁

Notes: N=39 ** correlation is significant at the 0.01 level * correlation is significant at the 0.05 level** p < .01, * p < .05 (two-tailed).

Table 2- Descriptive statistics and correlation matrix Public sector

Mean SD 1 2 3 4 5 6 7 8 9 10 Dependent variable 1. AI 5.18 0.75 1 Independent variables 2. CP-SO 3.86 1.35 .361* ₁ 3. CP-O 4.05 1.39 .357* _.730**₁ 4. CP-CU 4.83 1.08 .221 .375**_.302* ₁ 5. NP 3.53 1.43 .183 .342* _.545**_.222 ₁ 6. MP 4.62 0.78 .513**_.329* _.336* _.579**_.343*₁ Moderator 7. IG 0.45 0.50 -.104 -.054 .133 .023 .146 -.109 1 Control Variables 8. Firm Size 4.19 0.90 -.182 -.133 -.071 -.033 .122 .117 -.097 1 9. IT department Size 3.11 1.94 -.261 -.256 -.179 -.080 .129 -.102 .106 .150 1 10. Firm age 4.34 0.96 -.118 -.119 -.154 -.121 .150 -.160 -.142 .099 .120 1

(35)

5.2 Hierarchical regression analysis

In order to empirically test the hypotheses, hierarchical regression analysis was performed in three steps. The method is suitable as it allows examination of the effects of the control variables and the independent variables separately (Nasco, Toledo, & Mykytyn, 2008). Each variable was regressed on the dependent variable ‘the intention to increase CC adoption’. Model 1 only includes the control variables, ‘firm size’, ‘firm age’, ‘IT department size’. In order to test hypotheses 1-3, model 2 adds the constructs of the Institutional Theory: coercive pressure, normative pressure, and mimetic pressure as independent variables (i.e., factor scores). In model 3 the potential moderating effect of the mean centered interaction between mimetic pressure and attendance of international gatherings on the dependent variable were added. Prior to conducing a hierarchical multiple regression, the relevant assumptions of this statistical analysis were tested

5.3 Assumptions

(36)

amongst the predictor variables, Variance Inflation Factors (VIF) were calculated and analyzed. All the VIF values are less than 10, meaning that there is no multicollinearity (O’Brien, 2007) (see appendix D). Thus, all testing assumptions were met for the models and regressions that were tested, which shows that the robustness and validity of the regressions and models. Table 3 - Regression results private sector

Model 1 Model 2 Model 3

Private sector _B _β _B _β _B _β Control variables Firm Size -.471 -.461 -.288 -.282 -.241 -.236 IT department Size .099 .152 .037 .057 .050 .076 Firm age _.570 _.427 _.416 _.312 _.263 _.197 Main effects Coercive pressure – Parent company _{-.394 -.155} _-.375 _-.148 Coercive pressure – Customer _.486 _{.499** .440} _.452* Coercive pressure – Supplier _.122 _.105 _.120 _.104 Coercive pressure – Government _.027 _.020 _-.004 _-.003 Normative pressure _.119 _.158 _.095 _.126 Mimetic Pressure _.150 _.148 _.188 _.187 Moderation effects MP * IG _-.045 _-.205 F value 1.932 5.334*** 5.317*** R2 .142 .623 .655 Adjusted R2 .069 .507 .532

(37)

Table 4- Regression results public sector

Model 1 Model 2 Model 3

B β B β B β Control variables Organization Size -.116 -.140 -.175 -.210 -.149 -.179 IT department Size -.089 -.231 -.052 -.134 -.050 -.130 Age of organization -.059 -.076 .021 .027 .029 .037 Main effects Coercive pressure – superior organizations .056 .100 .064 .114 Coercive pressure – other organizations .074 .137 .074 .138 Coercive pressure – Customer -.137 -.198 -.127 -.183 Normative pressure -.022 -.041 -.033 -.062 Mimetic Pressure .554 .578** .515 .537** Moderation effects MP * IG -.334 -.169 F value 1.500 3.168** 3.053** R2 .095 .400 .426 Adjusted R2 .032 .274 .287

(38)

5.4 Regression Results

Examining the results of both regressions in tables 3 and 4 provide the information needed to conclude whether the proposed hypotheses are empirically supported.

Private sector. The three models in table 3 were examined separately. The explained variance of Model 1 was not significant with R2 was 0.142 and F (3,35) =1,932, p>0.05. Model 2 included the control variables and the independent variables to test H1a-3a. The explained variance of Model 2 was significant with an R2 of 0.623, F (9,29) = 5.334, p<0.001. Model 3 was a full model with all variables and the hypothesized interaction effect as independent variable to test H4. Taken together, the independent scales and moderator combined significantly predict adoption intention, R2 = 0.655, F (10,28) = 5.317, p< 0.001. The regression analysis results are shown in table 3. Hypothesis 1a predicted a positive relation between coercive pressure and the intention to increase CC adoption. Only for coercive customer pressure there was found a positive significant relation (β = 0.499, p <0.01). Empirical support for coercive pressures of the parent company (β= -0.155, p> 0.05), suppliers (β= 0.105, p>0.05) and government (β=0.020, p> 0.05) was not found. There was even a negative relation between coercive pressures of the parent company and the intention to increase CC adoption. Hence, hypothesis 1a is partially supported. Second, a positive but not a significant effect of normative pressure on the intention to increase CC services is found (β= 0.158, p> 0.05). Therefore, hypothesis 2a is not supported. Third, empirical support for hypothesis 3a was not found. Mimetic pressure does have a positive effect on the intention to increase CC adoption, but this relation is not significant (β=0.148, p= >0.05). The results of hypothesis 4a in model 3 showed a negative but insignificant moderated effect of the interaction between the international gatherings and mimetic pressure on the intention to increase CC adoption intention (β = -0.205, p >0.05). Therefore, hypothesis 4a is not supported. None of the control constructs showed a significant effect on the dependent variable (p>0.05). Thus, the control variables have no influence on intention to increase the CC services within this study.

(39)

(40)

6. Discussion

The present thesis examined whether higher institutional pressures influence an increase in the CC adoption intention moderated by the attendance of international gatherings. Prior research mainly focused on the CC adoption intention in a single sector (Oliveira & Martins, 2011; Maqueira-Marín, et al., 2017; Gao & Sunyaev, 2019), whereas this study examines both the private sector as the public sector. Moreover, this study aimed to clarify an inconsistency in the literature about the differences in the external environment of private and public sector organizations. In addition, the international nature of CC is taken into account by understanding the impact of IT decision makers’ attendance at CC international gatherings. Therefore, the hypotheses of this study empirically examined to what extent institutional pressures impact the increase in CC adoption intention in the private sector similarly as in the public sector moderated by the attendance of international gatherings. The previous chapters developed a theoretical framework leading to several hypotheses related to this research question. In this chapter, the findings of the research are discussed.

First of all, it was found that some external factors positively affect a higher intention to adopt CC services for the private sector and public sector. The main argumentation for these effects was based on the institutional theory. In this study, it meant that organizations are impacted by institutions in their decision making whether they intend to increase their CC usage. It turned out that some of these external pressures experienced by organizations influenced their likeliness to increase CC adoption. Therefore, it can be concluded that these findings are in line with and partially support the institutional theory of DiMaggio and Powell (1983).

(41)

of customers impacting focal organizations in their IT adoption is a consistent finding with previous studies (Teo et al., 2003; Liang et al., 2007). It can be explained by the strong customer focus of private organizations. If they are not perceived to put customers first and continuously improve their service offerings, then their customers will get their services elsewhere (Parker et al., 2013).

In addition, it was expected that private sector organizations would experience pressure of their parent company to adopt more CC services when the parent company of the focal firm used these services as well. Contrary to the expectations based on theoretical assumptions (Teo et al., 2003), the regression did not find supportive evidence for this. This implies that the parent company do not seem to function as an indicator for the intention to increase CC services. A possible explanation could be that subsidiaries pursue their own interests and are not an instrument of headquarters’ decisions (Mudambi, & Navarra, 2004). This is highlighted by Yigitbasioglu (2015) who stated that the adoption of technologies is not always necessarily driven by a top-down approach. This may be especially significant in the CC context given the accessibility of such solutions and the extent of local experimentation (Yigitbasioglu, 2015). Perhaps the subsidiaries want to experiment with their current cloud environment first, before adopting more CC services their parents company use.

Moreover, also suppliers using CC services did not impact the intention of private organizations to adopt more CC services in this study. Dependence on suppliers arises when organizations are unable to switch to alternative suppliers. However, the results did not significantly support this hypothesis. Relating to Chwelos, Benbasat, and Dexter, (2001) who also did not found support for suppliers as external pressure influencing IT adoption, it could indicate that organizations are not being held hostage by sole suppliers. They might have the possibility to choose between suppliers and therefore not feel pressure to use a certain CC service.

(42)

technology adoption by companies. Moreover, another possible explanation could be the political and administrative system of the Netherlands. Whereas in other countries such as China the government has been coercing firms to use IT systems (Liang et al, 2007), in the Netherlands there is an emphasis on consensual policy making (Nieto, Wittek, Heyse, 2013). Consequently, there is for example no single ministry in charge for e-government policy which results in critiques concerning the direction and control of e-government (Aagesen, van Veenstra, Janssen, & Krogstie, 2011).

For the public sector, none of the coercive pressures, superior organizations, up and downstream organizations and customers, are showing an impact on the intention to increase CC adoption. In this study, it is stated that superior organizations impacting the CC adoption of organizations because organizations at lower levels of government are dependent on the resources from higher levels of government in the form of democratic legitimacy, legal mandates, and financial means (Andersen & Jakobsen, 2018). Therefore, superior organizations have an impact on the technology decisions of public sector organizations (Zheng et al., 2013). Even though this effect is not found in this study, I still strongly believe the government could have a significant effect on the intention to increase CC adoption of public sector organizations. As CC technologies have evolved rapidly, the policies for adoption and use of CC technologies are still trying to catch up (Ali, Soar & Shrestha, 2015). Thus, when policies concerning CC services are clearly established, superior organizations could play a critical role in the adoption of CC of public sector organizations, for example in the form of data protection policies regulated by government (Ali and Osmanaj, 2020).

(43)

organizations are monopolies and they do not experience the risk of losing customers to peer organizations (Nieto, Wittek, & Heyse, 2013). Therefore, it could be the case that public organizations do experience customer pressure, but the urge is not high enough to increase the intention to adopt more CC services because they will not lose their customers to competitors which is the case in the private sector (Rocheleau & Wu, 2002).

The second hypothesis tested if normative pressure would lead to greater intention to adopt CC services for both private and public sector. However, in both sectors, the results did not significantly support these hypotheses. Whereas Oredo, Njihia, & Iraki (2019) found support for normative pressures having a positive relationship on CC adoption in the private industry, in this study this relation is not found. Normative pressure was measured by following Powell and DiMaggio (1991) stating that normative pressure mainly stems from the development of professional networks of organizations. Besides including the participation in professional bodies to measure normative pressure, Oredo et al. (2019) also included two other factors: the extent of CC adoption by its suppliers and by its customers. The two factors were the main source of the normative pressure in their study and thus professional bodies effect on the CC adoption decisions was also limited in the study of Ordeo et al., (2019). A possible explanation could be the willingness and ability to exchange innovation information and experience among the members of a network. This is influenced by the degree to which organizations are linked by interpersonal networks. For example, a lack of mutual trust or a high degree of inter-organizational competition will hinder the exchange of knowledge and experience (Korteland & Bekkers, 2008). As a high degree of inter-organizational competition exist among private organizations, this could explain their unwillingness to share knowledge and experiences among each other.

(44)

professional norms are the strongest determinant of behavior due to the threat of peer sanctions. Perhaps the standardization of norms in the public sector does not exist because rules are not sanctioned within the profession demands.

When considering mimetic pressure, surprisingly, the relation between mimetic pressure and the intention to increase CC adoption was not found in the private sector. A possible explanation for this finding could be as followed: As private organizations are driven by competitive advantage, they constantly trying to outperform their competitors (Tillema, 2010). To do so, they do not want to expose their strategies to gain this advantage. CC is viewed as a key factor for achieving competitive advantage (Low et al, 2011) which might result in organizations not wanting to expose their strategy to competitive others. However, other studies conducted in the private sector did find support for the relation between mimetic pressure and enterprise resource planning (ERP) adoption (Liang et al, 2007; Teo et al, 2003). Perhaps one particular IT system is easier to mimic because organizations can decide to publicize particular successful examples of adoption which can strengthen the effects of mimetic pressures (Oredo, et al., 2019).

There is found support for public organizations experiencing mimetic pressure in their cloud adoption intention. This means that IT managers experience pressure to use more CC services in their organizations due to the perceived success of their peers. The value of the beta coefficient (β = 0.578) is even the highest of all independent variables, suggesting that the effect of mimetic pressure on the intention to adopt more CC services is stronger than each of the other independent variables. In this study it was argued that public sector organizations benchmark themselves against best-in-class to identify best practice and use it as a basis for evaluating agency performance to identify areas for improvement (Gunasekaran, 2005). Moreover, they also often imitate similar others when they perceive a feeling of uncertainty and ambiguousness towards a technology resulting in imitation of other organizations' policies and practices for technological innovation (Currie, 2012). It is interestingly to note that previous study conducted in the context of public organizations did not found support for this relationship (Zheng et al., 2013) and therefore this study extend prior research.

(45)

(46)

7. Conclusion

To conclude, the comparison of private and public organizations can be viewed as changing. As noted earlier, during the past decade public organizations have, in many cases, sought to equal the private sector in many characteristics that have distinguished them in the past. Whereas Shin (2013) argued that private sector organizations are embracing CC faster than public sector organizations, this thesis showed that both sectors experience incentives of the external environment impacting the intention to increase CC adoption. The finding that public sector organizations experience external incentives means that they are responsive to changing demands in their environments, for example in the demands of their peer organizations. Perhaps this is a reason why they try to improve their services and to reduce their costs (Eliassen and Sitter 2008) which result in public organizations going into the direction of private organizations. However, the private sector experience different external pressures than the public sector, namely customer pressure. The profit driven character of the private sector organizations stimulates them to always look for ways to outperform their competitors, optimize processes and deliver the best services to their customers (Parker et al, 2013), where CC is viewed as a key factor for achieving this in the private sector (Low et al, 2011). Due to the differences experienced in the external environment between the private and public sector, I argue that key differences in technology adoption as a result of external pressure between the two sectors remain. Moreover, the findings showed that an international gathering context is not a setting where an IT decision maker is impacted by the mimetic pressure to make a CC adoption decision to increase usage. Whether public organizations will become equally responsive to the same cues of change as their private counterparts remains a relevant question. However, for the time being, it must be concluded that in the last decade a lot of the public sectors’ external environment changed, but when compared to its private counterpart, some concrete differences remain.

7.1 Research contributions

To what extent do institutional pressures impact the intention to increase cloud computing adoption similarly for private and public organizations, and is