How can government initiated efforts influence the motivations of
entrepreneurs to re-use open data?
An explorative case study of open data entrepreneurs
Student: Paul Meeuwissen (10636188) Supervisor: Dr. N. van der Meulen
Study: MSc. Business Administration – Digital Business Date: August 18, 2017 – Final version
2 | P a g e
Abstract
This study examines the influences of government initiated efforts on motivations of entrepreneurs to re-use open data. In contrast to prior research on open government data, this study examines the user’s perspective with a main focus on their motivations to use open data. This has implications for understanding the adoption process of open data and the role of the government as stimulator. Through an inductive, multiple-case study of 11 open data entrepreneurs, this study finds that open data entrepreneurs are motivated by five main drivers: passion, market opportunity, part of job, data or product shortcomings, and knowledge and skill development. Additionally, the research identifies seven influencing factors on these motivations: the supply of data, the funding of open data projects, the communication about open data, the organization of open data events and users’ technical knowledge and available time. These findings make three important contributions to the field of open data. Firstly, this study extends the academic knowledge on open data re-users by providing insight in the field of open data from a re-users’ perspective. Secondly, the findings extend the literature on both open data and entrepreneurship through the identification of motivations of open data entrepreneurs. Thirdly, this study includes practical suggestions that helps policy-makers and civil servants to improve the effectiveness of open data initiatives and thereby increase the impact of open government data.
Keywords: business, entrepreneurship, motivation, open data, open government data
3 | P a g e
Statement of Originality
This document is written by Student Paul Meeuwissen who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that
no sources other than those mentioned in the text and its references have been used
in creating it.
The Faculty of Economics and Business is responsible solely for the supervision
of completion of the work, not for the contents.
4 | P a g e
Table of contents
1. Introduction ... 6 1.1 Theoretical relevance ... 8 1.2 Practical relevance ... 9 2. Theoretical background ... 10 2.1 Definitions ... 10 2.2 Open data entrepreneurs ... 11 2.3 The open data ecosystem ... 13 2.3.1 Data supply ... 15 2.3.2 Data governance ... 17 2.3.3 User characteristics ... 19 2.4 Entrepreneurs motivations to use open data ... 20 2.4.1 Entrepreneurs in general ... 20 2.4.2 Entrepreneurs using open data ... 22 2.5 Motivations, the entrepreneurial process and open data ... 23 3. Methodology ... 25 3.1 Research design ... 25 3.1.1 Research approach ... 25 3.1.2 Research strategy ... 26 3.1.3 Sampling and case selection ... 26 3.2 Data collection ... 32 3.2.1 Method ... 32 3.2.2 Collection procedure ... 32 3.3 Data analysis ... 33 3.4 Quality criteria ... 33 3.4.1 Validity ... 34 3.4.2 Reliability ... 34 4. Findings ... 35 4.1 Motivations ... 35 4.1.1 Personal motivations ... 35 4.1.2 Organizational motivations ... 43 4.2 Influences on motivations ... 45 5. Discussion ... 56 5.1 Main findings ... 56 5.1.1 Re-use motivations of open data entrepreneurs ... 56 5.1.2 Research question ... 59 5.2 Theoretical relevance and practical Implications ... 63 5.2.1 Theoretical relevance ... 63 5.2.2Practical implications ... 64 5.3 Limitations and suggestions for future research ... 65 6. Conclusion ... 66 7. References ... 685 | P a g e 7. Appendices ... 74 7.1 Appendix A: list of abbreviations ... 74 7.3 Appendix B: invitation mail interview ... 74 7.3 Appendix C: Interview Protocol ... 75 7.4 Appendix D: Coding Scheme ... 76
List of figures
Figure 2 – The holistic open data assessment framework ... 14 Figure 3 – The concentric shell model ... 15 Figure 4 – Adjusted holistic open data assessment framework for entrepreneurs ... 24 Figure 5 – interconnectedness between personal motivations and open data use ... 41 Figure 6 – Influences on motivations ... 59List of tables
Table 1 – Case description overview ... 31 Table 2 - Primary motivators of participants to use open data ... 36 Table 3 - Perceived motivators of organizations to start using open data ... 44 Table 4 – Factors that (potentially) influence motivations to use open data ... 46 Table 5 – Comparison of motivations with prior literature ... 57
6 | P a g e
1. Introduction
Open government data is a fast-growing field of practice and research that draws the attention of policy-makers, practitioners and researchers. Accordingly, global organizations such as the World Bank and the United Nations acknowledge the value of open government data and stimulate its release. (Gonzalez-Zapata & Heeks, 2015).
The literature provides three main arguments as justification for the release of government data, it has the potential to increase transparency in government functioning (Bertot, Jaeger, & Grimes, 2010), to create social and commercial value (Chan, 2013), and to foster public participation in government activities (Conradie & Choenmi, 2014; Open Government Data, n.d.; Attard, Orlandi, Scerri & Auer, 2015). First of all, the concept of transparency is based on the idea that in a well-functioning, democratic society citizens and other stakeholders have the right to be informed about how the government is functioning. Open government data not only contributes to the provision of information, but also enables citizens and stakeholders to use, share and re-use government information (Attard et al., 2015). Secondly, governments are among the largest collectors and producers of data, of which a large amount is low or non-sensitive (Alexopoulos, Zuiderwijk, Charapabidis, Loukis & Janssen, 2014). This data includes meteorological data, geographical data, planning data, road traffic information or budget data which provides opportunities for both individuals and businesses (Attard et al., 2015; Chan, 2013). By releasing this data, governments pave the way for innovative initiatives that can deliver economic and social value (Chan, 2013; Jetzek, Avital, & Bjorn-Andersen, 2014). Thirdly, public participation refers to citizens’ participation in government activities, such as decision and policy-making (Attard et al., 2015). The emergence of social media and other online platforms expanded the possibilities for public participation (Sandoval-Almazan & Gil-Garcia, 2012; Attard et al., 2015). At the same time, national and local governments increased the possibilities for public participation through the use of information and communication technologies, including the use of open data (Sandoval-Almazan & Gil-Garcia, 2012). The opportunities for public participation are likely to expand even further due to the continuous pressure on governments to improve open data availability (Zuiderwijk & Janssen, 2014).
Alongside the three arguments to release data, national governments in Europe are also under pressure to open up data. There are two main reasons for this. First, the European Commission is a stimulating force in opening up data since the end of the 1980s (Janssen, 2011). In 2003, this resulted in the adoption of the PSI Directive 2003/98/EC with the main
7 | P a g e
objective to foster the equal treatment of potential re-users (European Parliament and Council, 2003; Carrara, Chan, Fisher & Steenbergen, 2015). To further stimulate the re-use of open government data the PSI Directive is amended by Directive 2013/37/EU in 2013 (Carrara et al., 2015). This legislation encourages states to make as much information available for re-use as possible.
Second, high expectations on the benefits of open data put an additional pressure on governments to reach the full potential of open data. Carrara et al. (2015) estimate that in Europe the open data market size will increase by 36.9% between 2016 and 2020 with a total value of 75.7 billion euro in 2020. Furthermore, estimations show that open data is currently responsible for 75.000 direct jobs, which grows to 100.000 in 2020 (Carrara et al., 2015). This growth correlates with the tremendous increase of available open government data in the past few years (Algemene Rekenkamer, 2016; World Wide Web Foundation, 2015; HM Government, 2016). Not only in Europe, but also in the rest of the world governments are opening up data for access and re-use (Jetzek, Avital, & Bjorn-Andersen, 2014). To date, 75 countries worldwide are part of the Open Government Partnership (OGP) and commit to make data free to use, re-use and redistribute following a set of ‘Open Data Principles’ (Attard et al., 2015; Open Government Partnership, 2015).
While the number of available datasets rapidly increases, the use of datasets is still rather disappointing (Zuiderwijk, Janssen, & Dwivedi, 2015; Bertot, McDermott, & Smith, 2012). Prior research fails to give an explanation, since only limited information is available about the identity of users and their drivers to use open government data (Zuiderwijk et al., 2015). Still, the open data debate mainly focuses on the provision of data (Foulonneau, Martin, & Turki, 2014). This focus on provision is remarkable, since there is agreement among researchers that information about who is using data for what purpose is essential for optimizing the supply of data. Additionally, information about users helps in making investment decisions that meet the needs of open data re-users (Davies, 2010; Loenen, Welle Donker, & Braggaar, 2016; Carrara et al., 2015). Consequently, a shift in the open data debate from the provision to the use of open data is necessary to get better insight in the (re-)use of open data. Therefore, this study aims to explore the relationship between the drivers of open data entrepreneurs and government efforts to stimulate the use of open data. To do this, this research examines the process in which entrepreneurs start with the use of open data and explores the related needs and stimulating forces. Therefore, this study aims to answer the following question:
8 | P a g e
How can motivations of open data entrepreneurs be influenced by government initiated efforts to stimulate data re-use?
Because of the limited research on motivations of open data users, the following sub-question helps to answer the research question: What are the motivations of open data entrepreneurs to
re-use open data?
To answer these questions this research is organized as follows. The study starts with a review on the contemporary literature on the concept of open data and links this to literature on entrepreneurship to find potential overlap. This literature review initially examines the definitions related to open data, and places entrepreneurship in the context of open data. Subsequently, the literature review examines the open data ecosystem and its separate elements and compares this with research on entrepreneurial motivations to identify the research gap. Next, the methodology describes the research design and gives an extensive description on the selected cases. Additionally, the method section includes a description of the data collection and analysis procedure. Chapter four then presents the findings, which are subsequently discussed in the next chapter. The discussion section also includes a discussion on the contributions and limitations of this study, and suggests directions for future research. Finally, this study draws a conclusion that encompasses a summary of the main findings. 1.1 Theoretical relevance
This study expands the academic research on open data on at least four aspects. Firstly, the current state of academic literature on open government data primarily examines the open data ecosystem from a supply-side perspective (Foulonneau et al., 2014). Through a focus on the user, this research expands the research on open data from a user perspective. Secondly, while prior research identifies factors that have the intention to stimulate open data use (Welle Donker & Loenen, 2017), no research provides evidence that these influences have an actual effect. This study addresses this gap and examines the process and influencing factors on the basis of which (potential) re-users start with the use of open data. Thirdly, this study examines ‘open data entrepreneurs’ as a specific user group. In doing so, this study helps to distinguish re-users and identifies specific capabilities based on user characteristics. Fourthly, this research examines the motivations of open data entrepreneurs and thereby expands the research on entrepreneurship.
9 | P a g e
1.2 Practical relevance
From a practical point of view this study is particularly relevant for policy-makers and civil servants. Because of the limited availability of information about open data users and their usage, it is difficult for governments to determine the actual use of data. Let alone to determine the predictors that influence users’ willingness, ability and intention to use open data (Zuiderwijk et al., 2015). Therefore, this study provides insights in the characteristics and motivations of open data entrepreneurs. A better understanding about re-users and their drivers to use open data helps governments to determine how they can target potential re-users more effectively. Moreover, a better understanding of re-re-users also helps civil servants to organize effective open data events, such as workshops, user groups and hackathons.
10 | P a g e
2. Theoretical background
2.1 DefinitionsPrior literature has used different terms to refer to public data resources. The definitions most commonly used are ‘Open Data’, ‘Public Sector Information (PSI)’ and ‘Open Government Data (OGD)’ (Carrara et al., 2015).
Open Data is referred to as “data that are freely accessible online, available without
technical restrictions to re-use, and provided under open access license that allows the data to be re-used without limitation” (Open Knowledge Foundation, 2012; Jetzek et al., 2014). Open data is an subset of ‘Big Data’ (Carrara et al., 2015), which is in turn described as information assets that are high-volume, high-velocity and high-variety and require innovative information processing for enhanced insight, decision making and process automation (Gartner, 2017).
Public Sector Information on the other hand is described as all information collected
by the public sector. It is information that is “generated, created, collected, processed, preserved, maintained, disseminated, or funded by or for the government or public institution” (Carrara et al., 2015; OECD, 2006). In contrast to open data, public sector information does not necessarily involve datasets, but can also be written information.
Open Government Data is the combination of public sector information with open data
as illustrated in figure 1 (Carrara et al., 2015). Open Government Data refers specifically to data that is offered by the government or government controlled entities that are open for use and re-use by public and private agents (Open Knowledge Foundation, 2012; Jetzek et al., 2014). In the context of this study open government data (OGD) is used as main definition unless explicitly indicated otherwise.
11 | P a g e
2.2 Open data entrepreneurs
The potential of data-driven economy lies to a large extend in the development of new services (Hammel, Perricos, Branch & Lewis, 2011). Entrepreneurs as driving force behind innovations and technical change (Schumpeter, 1942) therefore make an interesting target-group to focus a user-perspective based research on. The definition of an entrepreneur however is not self-evident, since there is not agreed upon the definition of entrepreneurship.
The data portal of the Netherlands uses six categories to distinguish users of open data: citizen, developer, entrepreneur, public sector organization, student or scientist and unknown applicants (Dataportaal van de Nederlandse overheid, 2017). Although definitions of the categories are not available, considering the generality of each category it can be assumed that ‘entrepreneurship’ is used in the broadest sense of the word. Interesting to note is that user information is only collected from users and re-users that have submitted a data-request and have thereby classified themselves under a category. Accordingly, it depends on the user or re-user under which category he or she arrays himself. Confusion might occur when someone considers himself in two or more categories, for example when a civil servant considers himself as an entrepreneur while he is part of a public sector organization. In the United Kingdom (UK) more detailed categorizations are used; they distinguish seven organization types: private individual, start-up, Small and medium sized enterprise (SME), Large Company (over 250 employees), Voluntary sector or not-for-profit organization, Public Sector Organization and Academic or Research (data.gov.uk, 2017). Although the entrepreneur is not defined as a separate user type, the data portal of the UK defines more specific types of organizations that could, depending on the used definition, include entrepreneurs. Nevertheless, neither in the Dutch nor the English data portal information is available on the definitions of entrepreneurs.
In literature about entrepreneurship there is a wide variety of definitions of what constitutes an entrepreneur (Walley & Taylor, 2002). Definitions have their origins in different academic disciplines, such as psychology, sociology and economics (Blundel and Smith, 2001). Two basic trends are identified in defining entrepreneurship (Davidsson, Delmar & Wiklund, 2006). In the first view the entrepreneur is seen as the person who creates and develops new organizations (Gartner, 1988). This view has a clear focus, and therewith limits interpretations that might broaden the field of study. Within this perspective still a variety of definitions are used. The narrowest definition of an entrepreneur related to this view describes entrepreneurs as “people starting-up a new business venture which is less than 42 months old” (Williams, 2008). In this first view, if an individual develops a product by using
12 | P a g e
open data and starts a new firm this is regarded as entrepreneurship. However, when the same individual develops the same product within an organization, this would not be regarded as entrepreneurship (Davidsson et al., 2006). Thus, limitations of this first view arise when the creation and development of new products within organizations are considered.
The second view on the other hand let go of the perspective that entrepreneurship only entails the creation of new organizations. Instead it takes a broader definition and regards the entrepreneur as an innovator that changes the economy in some way or another by the creation of new economic activity (Schumpeter, 1942; Śledzik, 2013). New economic activities refer to activities that are new to the firm (or a new organization), and changes the offerings that are available on the market. Entrepreneurship from this perspective can also be for example about an imitative product or service that creates or increases competitiveness and thereby changes the market (Davidsson et al., 2006). This second view is more functionalistic in nature (Bruyat & Julien, 2000), thence entrepreneurs that act within a firm following the above criterion are also regarded as entrepreneurs.
Where the above two views describe entrepreneurship by defining its impact, Venkataraman (1997) instead examines the process of entrepreneurship to formulate a definition. He states that the existence of opportunities is one thing, but the discovery and exploitation is something entirely different (Venkataraman, 1997). Central to this perspective is the belief that “people are different and that these differences matter” and user characteristics therefore influence the discovery and exploitation of opportunities that are offered (Venkataraman, 1997).
As this research has its main focus on the process on which open data opportunities are identified and eventually used (e.g. motivations and capabilities) a process-oriented definition is used as starting point. Therefore, the definition of Shane and Venkataraman (2000) is followed that regards entrepreneurs as ‘people that are responsible for the process by which opportunities to create future goods and services are discovered, evaluated, and exploited’. Thus, when ‘open data entrepreneurs’ are mentioned, the definition of Shane and Venkataraman (2000) is followed with the addition that those entrepreneurs are using open data as their main resource or one of their resources. Thus ‘open data entrepreneurs’ are regarded as people that are responsible of the process by which open data opportunities to
13 | P a g e
2.3 The open data ecosystem
Now the definitions of open government data have been discussed and the adopted concept of entrepreneurs is defined, this section gives a broader view on the open data ecosystem as a whole and its separate elements. In trying to understand and assess the open data ecosystem, several open data assessment frameworks have been developed in the past few years. The Open Data Barometer for example assessed countries worldwide on the readiness, implementation and impacts of their open data initiatives (World Wide Web Foundation, 2016). A framework of the Independent Reporting Mechanism assessed the governance of open data initiatives (Khan & Foti, 2015), and the Open Data Institute in the United Kingdom developed a framework to assess organizations on their ability to publish and consume open data from a data provider perspective (Dodds & Newmann, 2015). Existing frameworks however, often assess open data from the perspective of data suppliers, whereas the user perspective is still absent (Welle Donker & Loenen, 2017).
By combining elements of previous frameworks and by adding the user perspective Welle Donker & Loenen (2017) developed a holistic open data assessment framework (figure
2) with the goal of creating a complete overview of the open data ecosystem and to assess the
effect of open data policy regulation. In their framework Welle Donker & Loenen (2017) distinguish 4 elements: (1) the activity, the action of an organization, such as opening up data or financing initiatives, (2) the output, described as the products or services of an organization, (3) the outcome, considered as the results of an action (4) and the impact, which is described as the contribution of outcomes to the strategic goal of an organization. To illustrate, a government that makes weather data available (activity) allows everyone to use this data (output). If the same government wants people to start developing apps based on this data, they could decide to organize a hackathon (activity), resulting in people working with this data (output) and potential apps (output). A developer that creates a successful app forecasting the weather for a precise location (outcome) thereby helps people to decide when to leave their house to avoid showers (impact).
14 | P a g e
Figure 2 – The holistic open data assessment framework
The question that remains however, is why this developer would actually start developing this app. The framework suggests three indicators as conditions for the successful adoption of open data: data supply, data governance and user characteristics. While data supply and data governance have been a subject of research in multiple studies, research about characteristics of open data users is scarce. Still, although limited defined Welle Donker & Loenen (2017) did acknowledge its importance and included user characteristics as an additional output indicator in their framework. Nevertheless, little can be said about this indicator since only limited information is available about users of open data and the predictors influencing their willingness, ability and intention to use open data (Zuiderwijk et al., 2015). In the next three sections the conditions for open data use are examined. This is done by following the structure of Welle Donker & Loenen (2017) since their framework includes the user perspective and is most complete compared to other frameworks. Accordingly, a distinction is made between data supply, data governance and user characteristics.
15 | P a g e
2.3.1 Data supply
Data supply is discussed in this section using the concentric shell model of Backx (2003). Prior research shows that data supply as facilitating condition is not a good predictor of behavioral intention to use open data (Rana, Williams, Dwivedi, & Williams, 2012; Zuiderwijk et al., 2015). Treating data as a means and taking away barriers for the use of open data is regarded more effective than a focus on the publication of data (Zuiderwijk et al., 2015). To evaluate the data supply of public information from a re-user perspective, Backx (2003) developed a concentric shell model (figure 3). This ‘Data accessibility model’ assumes that a re-user can use public information if three conditions are met:
1. Where and what information is made available is known to the re-user; 2. Information is financially, legally, and practically attainable for the re-user; 3. And information is usable by the re-user.
The model of Backx (2003) therefore consists of three main layers: known, attainable and usable. Every layer consists of additional supply indicators that are discussed below.
Figure 3 – The concentric shell model (Backx, 2003)
Known
The outer layer in the model of Backx (2003) indicates that the existence of information must be known before it can be used. To be known by the user, information needs to be both recognizable and findable. The ‘resource metadata’ of information gives an indication of the
16 | P a g e
recognisability of information (Loenen & Grothe, 2014). Resource metadata refers to titles, abstracts and graphics that help to identify information resources.
The findability of data depends on the familiarity and accessibility of portals where information is offered (Open Knowledge International, 2017). Several countries developed specific data portals to share open government data. Examples can among others be found in the Netherlands (Dataportaal van de Nederlandse overhead, 2017), the United Kingdom (data.gov.uk), and France (data.gouv.fr). Also Europe has a data portal (europeandataportal.eu) functioning as an umbrella portal of the EU countries. Furthermore, the findability of data also depends on search results from search engines (Loenen & Grothe, 2014).
Attainable
Attainability can be seen from three perspectives: financial attainability, legal attainability and practical attainability (Loenen & Grothe, 2014). Financial attainability refers to the costs to require information. While the affordability of access depends on the re-user, it can be assumed that lower fees better meet user needs than higher fees (Loenen & Grothe, 2014).
Legal attainability refers to the availability based on legislation. Legislation applies to both the supplier and re-user of data, such as intellectual property rights, restrictions through licenses and personal data protection.
Practical attainability refers to the way data can be accessed. Optimally data is always available and can be obtained whenever desired; meaning through direct downloads without bureaucratic procedures. Additionally, the physical attainability involves the way in which data is offered.
Usable
To determine the usability of data, multiple indicators can be identified. Loenen et al. (2016) describe six aspects of usability: clarity, manageability, reliability, communication, actuality and permanency. Firstly, the clarity of data refers to the provided resource metadata. Since metadata is also used to improve findability this illustrates the multi-purpose function of metadata and underlines its importance. Conradie & Choenni (2014) even indicate that the adequate use of open data can be hindered through defective documentation of metadata. Secondly, the manageability of a dataset depends on the format and the machine-readability of data. Users might have different format preferences; making data available in multiple formats can improve the manageability of data for different users (Loenen et al., 2016).
17 | P a g e
Thirdly, the reliability is about the completeness, correctness, consistency and validity of information (Loenen & Grothe, 2014). Prior research on big data indicates that the volume and redundancy can be used to crosscheck conflicting cases, validate trustworthiness and compensate for missing data (Jagadish, Gehrke, Labrinidis, Papakonstantinou, Patel, Ramakrishnan, & Shahabi, 2014). Fourthly, communication, involves the availability and the response time of help services, such as help desks. Fifthly, the actuality of open data can be determined by the actuality or the update frequency of data. Applications such as ‘buienradar’ that give real-time information about the weather are dependent on direct available information. Again, a comparison can be made with big data, where timeliness is also mentioned as a challenge (Jagadish et al., 2014). Timeliness of data is in particular important for products or services where time is of the essence, such as travel planners or for identifying fraudulent credit card transactions. Finally, the permanence of data can also influence its usability. If an entrepreneur for example wants to invest in the development of an application that is dependent on actual open data, he wants to know whether the open data will be available in the future to estimate future returns. To determine the permanence of information, legal requirements sometime give guarantees of future availability (Loenen et al., 2016).
2.3.2 Data governance
Next to the supply of open data, data governance can help government entities to foster open data. Data governance can be described as the interaction between entities of the public and private sector with the purpose to realize common goals in the field of open data (Termeer Dewulf, Rijswick, Buuren, Huitema, Meijerink, Rayner & Wiering, 2011). A common goal could be for example to maximize both the social and financial impact of open data. When referring to data governance, this study primarily focusses on government initiated activities to reach these goals. Activities consist of policies, processes and instruments that structure the interaction between public and private sector entities. These activities can affect potential re-users (Shane, Locke & Collins, 2003) and therefore support the re-use of open data. A study assessing the open data maturity in Europe shows that 84% of the national governments in Europe incorporate policies that support the re-use (Carrara, Nieuwenhuis & Vollers, 2016), which means that most national governments in Europe not only ensure the supply of open data but also actively encourage its re-use. Data governance as mentioned in the open data ecosystem framework of Welle Donker & Loenen (2017), distinguishes five elements to determine the functionality of a data infrastructure: vision, leadership, communication,
self-
18 | P a g e
organizing ability, and long-term financing. Because self-organizing ability and communication are directly aimed at re-users, these are further examined below.
Firstly, self-organizing capacity encompasses the efforts to stimulate and promote open data and match its supply and demand. Strategies to stimulate open data could be conference presentations, workshops and hackathons and offering innovation prizes. A study by Zuiderwijk et al. (2015) examined the acceptance and use predictors of open data technologies and found four predictors that significantly influence behavioral intention: performance expectancy, effort expectancy, social influence and voluntariness of use. Based on these results Zuiderwijk et al. (2015) propose some suggestions for policy-makers to stimulate the use of open data technologies such as the integration of open data use in daily activities and decreasing efforts needed to use open data technologies. The study shows that compulsory use of open data increases the behavioral intention to use open data technologies. Additionally, providing training programs and a learning environment can help to increase awareness about the potential of open data which eventually increases the intention to use open data (Zuiderwijk et al., 2015).
Secondly, the aspect of communication involves all communication around open data from a government entity. Therewith it shows some overlap with the previously discussed supply of open data such as the presentation of a data portal or the communication of resource meta-data. However, communication can be seen in a broader context ranging from personal communication during an open data event to the communication of legal issues (Welle Donker & Loenen, 2017). Countries worldwide differ in their approach to communicate about open data. The open data portal of France for example (data.gouv.fr) makes the re-use of data-sets visible and is actively sharing the re-use practices with the public, whereas The UK is frequently organizing events and user groups to stimulate open data use. Based on the study of Zuiderwijk et al. (2015), effective ways to stimulate data use through communication could be showing the benefits of open data technologies, create awareness among current users of their open data use and encourage people to stimulate each other.
Open data policies to encourage open data use can simultaneously involve the self-organizing capacity, communication and other elements of open data governance. Organizing a conference on open data for example can be initially organized to communicate the state of affairs of open government data, but at the same time connects people that could leverage from each other’s knowledge and thereby stimulates the self-organizing capacity. Open data policies which encourage open data use include the organization of events such as hackathons, conferences and user groups. Hackathons primarily focus on the development of
19 | P a g e
applications, whereas conferences are organized to increase awareness about open data (Carrara et al., 2016). To develop applications for the health sector, Switzerland for example organized an “Open Geneva” Hackathon. Another example is Germany that organized the #NRWhackathon to develop educational applications. Conferences are organized across countries in Europe (Carrara et al., 2016). Romania for example organized the ‘Open Government week’ consisting of a week full of debates around elements of open data such as health care and the re-use of open data for smart cities. Next to the organization of events, other data policies could be implemented to encourage data re-use. The Open Data Institute in the UK for example offers an open data startup challenge program to function as an incubator of open data entrepreneurs (Dodds & Newman, 2015).
To effectively communicate and stimulate the self-organizing capacity of open data however, knowledge about the re-users is essential (Loenen et al., 2016). This is endorsed by studies in the field of marketing which demonstrate that particular needs and wants of consumers within an audience can be satisfied when the offering is adapted to a target group (Hoyer, Macinnis, & Pieters, 2016). However, when looking for example at the Netherlands there is little information available about the re-use of open data. Although there are some examples of the re-use of open data, such as rain shower alerts and travel planners, there is no national quantitative or qualitative data available about the use of open data (Loenen et al., 2016). Additionally, a study of Zuiderwijk & Jansen (2014) on characteristics of open data policies shows that none of eight open data policies they examined in the Netherlands has been focusing on a particular target group.
2.3.3 User characteristics
Now two environmental factors have been discussed, in this section the element of ‘user characteristics’ as inherent part of the re-user is examined. To create value with open government data there must be a motivation and the ability to use the data (Attard et al., 2015). To explain the behavior of open data re-users Welle Donkers & Loenen (2017) mention aspects such as technical connectivity, user capabilities, available resources, technical and creative skills without further elaboration.
Shane et al. (2003) on the other hand conducted research to explain ‘human action’ in more general terms without the involvement of open data. In explaining human action, Shane et al. (2003) make a distinction between motivational and cognitive factors, where they further distinguishes cognitive factors into: ability, knowledge, vision and skill. In the next
20 | P a g e
section the element of motivation as an explanatory power of re-using open data is further discussed.
2.4 Entrepreneurs motivations to use open data
No research has been done specifically on the motivations of entrepreneurs that make use of open (government) data. The motivations of entrepreneurs in general however, have been a focus for many studies (Kirkwood & Walton, 2010; Shane et al., 2003).
2.4.1 Entrepreneurs in general
Theories on motivations of entrepreneurship often differentiate motivations between push and pull factors (Amit & Muller, 1995; McClelland et al., 2005; Segal, Borgia & Schoenfeld, 2005; Hughes, 2003). Push factors are personal or external factors such as resignation or a marriage break-up that lead someone to entrepreneurship. Pull factors on the other hand are those aspects that draw people to undertake action, such as identifying a market opportunity or the wish to be independent (Kirkwood & Walton, 2010). Other researchers have used different categorizations to classify motivations. Williams (2009) for example builds on the research of the Global Entrepreneurship Monitor (Frederick & Chittock, 2006) and elaborates on the distinction between ‘necessity-driven’ entrepreneurs who have limited options for work and ‘opportunity-driven’ entrepreneurs who identify business opportunities. This classification can be compared with the push (necessity) and pull (opportunity) distinction. Interestingly, Segal et al. (2005) found that pull factors have generally been more predominant than push factors. This is especially interesting when considering that entrepreneurs that are pulled into entrepreneurship are more likely to be successful (Amit and Muller, 1995; Kirkwood & Walton, 2010).
Although widely used, the push and pull dualism has been criticized by several researchers. Williams (2009) for example argues that either/or dichotomy is too simplistic to describe motivations because motivations commonly involve both push and pull factors and can change over time. In his study on ‘off-the-book entrepreneurs’ he finds that 84% of the 70 respondents are both opportunity as necessity driven (Williams, 2009). Additionally, other studies show that the same motivation can be a push factor for the one individual and a pull factor for another (Hughes, 2003; Giacomin, Janssen, Guyot & Lohest, 2011). Accordingly, Aidis, Welter, Smallbone & Isakova (2007) questioned 378 entrepreneurs in Ukraine asking them to give three reasons why they had started their own business. The study showed that motivations included both push and pull factors and gave some evidence that motivations change over time (Aidis et al., 2007).
21 | P a g e
Related the context of push and pull factors or necessity and opportunity factors, other studies also made alternative classifications. Carter, Gartner, Shaver & Gatewood (2003) conducted a study to identify the motivations of entrepreneurs while simultaneously comparing them with career reasons of non-entrepreneurs that were seeking for a job. They distinguished six motivations: (1) innovation, the intention to accomplish something new; (2) financial success, the intention to earn more money and achieve financial security; (3) independence, the desire for freedom, control and flexibility in time; (4) recognition, the intention to have status and approval (from family, friends); (5) roles, the desire to follow family traditions or another his example; and (6) self-realization, the pursuit of self-directed goals. Interestingly, the study showed that motivations of nascent entrepreneurs were very similar to career reasons of non-entrepreneurs on financial success, independence, innovation and self-realization. Additionally, motivations concerning roles and recognition were rated lower by nascent entrepreneurs then non-entrepreneurs (Carter et al., 2003). Another research focused on a specific type of entrepreneurs: ‘ecopreneurs’, entrepreneurs with green values selling green products or services (Kirkwood & Walton, 2010). They identified five main drivers for starting a business: green values, earning a living, passion (for environment or product/service), being their own boss and seeing a market gap. The study showed that ecopreneurs had similar motivations as entrepreneurs in general, except for financial motivations which were less prevalent in the case of ecopreneurs (Kirkwoord & Walton, 2010). Moreover, Williams (2009) sticks close to the context of necessity and opportunity by adding two intermediate categories. While examining ‘off-the-books entrepreneurs’ he initially identifies five main motives: generate sufficient income to live/survive, generate additional income, desire to have won business, fill a gap in the market and independence. Subsequently he makes the distinction between: solely necessity entrepreneurship; mostly necessity entrepreneurship; mostly opportunity entrepreneurship and solely opportunity entrepreneurship. Finally, Shane et al. (2003) reviewed prior research to identify entrepreneurial motivations and makes a distinction between eight motivations: (1) need for achievement (nAch); (2) locus of control, the believe that actions or characteristics affect outcomes; (3) vision; (4) desire for independence; (5) passion, passionate, selfish love of the work; (6) drive, the willingness to put ideas into reality; (7) goal setting and (8) self-efficacy, the belief in one’s ability to attain a certain level of achievement on a task (Shane et al., 2003). Contrary to most of the other studies discussed, the eight motivations identified by Shane et al. (2003) focus solely on the opportunity (pull) factor, letting aside motivations that arise out of necessity.
22 | P a g e
2.4.2 Entrepreneurs using open data
As mentioned earlier, no research has been done specifically on motivations of entrepreneurs that use open data. Davies (2010) however examined open data users in general by questioning 72 respondents. He identified six overlapping motivational clusters:
1. Government focused; the desire to better understand the government and promote efficiency and accountability.
2. Technology and innovation focused; interest in the creation of new platforms, semantic web and linked-data.
3. Reward focused; looking for recognition or profit.
4. Digitizing government focused, focus on technologically driven efficiency and improvement.
5. Problem solving, driven by meeting particular challenges and willing to learn new skills and engage with open data to do so.
6. Social or public sector entrepreneurialism, focus on the provision of products and services that are based on open data.
The above mentioned motivations do not solely focus on open data entrepreneurs, and also include other types of users such as students, researchers or policy makers. Nevertheless, the sixth motivation explicitly mentions entrepreneurialism as a relatively broadly defined motivation. The other categories are more specified and show some overlap with motivations of entrepreneurs in general. The third category, ‘reward focused’ for example shows overlap with the generating of additional income (Williams, 2009) and need for achievement (Shane et al., 2003). A limitation of the research of Davies (2010) is that he used prior literature and exploratory research to define categories beforehand, while prior literature on motivations of open data users is scarce and the study fails to describe the process of how the exploratory analysis is conducted. As a result it can be questioned whether the six classifications used are representative. Additionally, Davies (2010) also acknowledges that the sample size is too small to draw any strong conclusions.
Moreover, a study conducted by Chan (2013) shortly mentions some motivational aspects to use open government data. Chan (2013) emphasizes the financial aspect, but also mentions other motivations that could influence the use of open data such as political or social agendas, working with data as a hobby, or pure altruism to advance the public good.
Currently, it is hard to compare motivations of open data entrepreneurs with entrepreneurs in general. Accordingly, researchers have multiple times called for more
23 | P a g e
research on the perspective of users regarding open data (Janssen, Charalabidis & Zuiderwijk, 2012; Carrara et al., 2015).
2.5 Motivations, the entrepreneurial process and open data
Entrepreneurship as a process and motivations of entrepreneurs have now been discussed separately. This part concludes the literature review by combining these aspects and explore relations that can be of importance for this study. As discussed earlier entrepreneurship can be seen as a process (Venkataraman, 1997; Shane & Venkataraman 2000). Shane et al. (2003) elaborate on this view and refer to the entrepreneurial process in three stages described as: opportunity recognition, idea development and execution. Similar to entrepreneurship in general, open data entrepreneurship can be viewed as a process instead of viewing open data use as one defined factor. A case study into data-driven innovation through open government data by Jetzek et al. (2014) shows strong similarities with the entrepreneurial process of Shane et al. (2003). To explore the process of innovation through open data they use the innovation value chain of Hansen & Birkinshaw (2007) who divided the innovation process in three phases: the generation, conversion and diffusion of an idea. These phases show similarities with the recognition of opportunities, development of ideas and execution phase of the entrepreneurial process (Shane et al., 2003). That entrepreneurship and innovation are interwoven is no surprise as prior research regularly regard entrepreneurs as innovators (Baumol, 2005).
Furthermore, it is argued that filtering takes place within the process of open data entrepreneurship (Shane et al., 2003). If a lot of people for example see opportunities in the use of open data, it does not mean that all those people will actually start using the data. Causes could be the inability to concretize and execute an idea, or the lack of diffusion capabilities (Jetzek et al., 2014). When a selection takes place within the entrepreneurial process or innovation mechanism it can be argued that influencing factors such as user characteristics and data governance not only have an effect on entrepreneurship as a whole, but can have a different impact on each step in the entrepreneurial process (Shane et al., 2003). This would suggest that when open data entrepreneurship is viewed as a process, motivations could vary or have a different impact depending on the step in the entrepreneurial process. For example, someone that is motivated by the need of achievement might find it easier to acquire money from investors than someone driven by the desire to be independent.
When assuming that the entrepreneurial process happens in different stages, it is most likely that the influence of environmental factors also varies per stage (Shane et al., 2003). In
24 | P a g e
the context of this study it is therefore plausible that the state of data supply and the way in which data governance is arranged, have a different impact on the recognition of open data opportunities, the development of ideas based on these opportunities, and the execution hereof.
To illustrate the context that is researched in this study an adjusted theoretical framework is proposed (see figure 4), based on the previously discussed literature. The framework is based on the holistic framework of Welle Donker & Loenen (2017) and is supplemented with the process of entrepreneurship and elements of the user characteristics proposed by Shane et al. (2003). The primary focus of this study is on the relationship between de government initiated activities to stimulate data use (illustrated by the circles on the left), and motivations to start using open data and become an open data entrepreneur (illustrated under user characteristics). This yet unexplored process is indicated with the blue arrows. Additionally, since Shane et al. (2003) assumes the interconnectedness between entrepreneurial motivations and the process of entrepreneurship, the process of entrepreneurship falls within the scope of this study.
25 | P a g e
3. Methodology
The previous chapter examined the open government data ecosystem, based on which an adjusted model is proposed. This chapter describes the methodology of this study. It starts with a description of the research design adopted for this study. Subsequently, as part of the research design the research approach, research strategy and the sampling process is described, followed by a description of each of the cases. Next, the data collection process and data analysis is explained. Finally, this chapter concludes with a discussion about the quality of the research design.
3.1 Research design
This study examines the central question: ‘How can motivations of open data entrepreneurs
be influenced by government initiated efforts to stimulate data re-use?’. The literature review
shows that prior studies have examined the open data ecosystem and identified potential factors that influence the re-use of open data. However, these have been identified from the perspective of the data supplier, whereas no research has been conducted to examine the influences that are experienced by the re-user. So far, there is no academic evidence that the suggested influences are actually influencing (potential) re-users. Answering the central question of this study will help to fill this research gap, and will extend the research field of open data by adding valuable information about motivations and influencing factors from a user perspective.
3.1.1 Research approach
Often a distinction is made between and inductive and a deductive research approach. The inductive research approach involves the exploration of new theories from data, whereas the deductive research approach refers to the use of data to test existing theories (Saunders, Lewis, Thornhill, Booij & Verckens, 2011). Given the explorative character of this study, an inductive research was adopted. The inductive research is regarded a suitable approach as it underlines and takes into account the context of situations (Saunders & Lewis, 2012). The relationship examined in this study and its separate elements have not been discussed in prior research. Although entrepreneurial motivations as a separate phenomenon is previously examined, motivations of open data re-users is yet an unexplored area. Additionally, government initiated efforts to stimulate data re-use is a relative new area of research from which only elements have been a subject of discussion with a primary focus on data supply
26 | P a g e
rather than data governance. As a result, an inductive approach was considered most appropriate for this study.
3.1.2 Research strategy
Due to the explorative nature of this study, a multiple case study was considered most suitable for this research (Eisenhardt, 1989). A case study is specifically useful to understand a relationship within a presented setting (Eisenhardt, 1989). The setting that was examined in this study involved the process in which open data is found, adopted and used while examining the influence of the data supplier (i.e. government entities) on the one hand, and the re-user of open data on the other hand. Additionally, a case study enables a researcher to get a deep understanding about the context of a phenomenon which was important for this study because of the novelty of the research setting (Saunders & Lewis, 2012). Moreover, the examination of multiple cases allowed the comparison between cases to check whether results were consistent (Eisenhardt, 1989).
3.1.3 Sampling and case selection Research context
Since countries substantially differ in their progress of opening data and the fast developments in the field of open data, it is important to take into consideration that this study is conducted in the Netherlands and data is gathered in May and June 2017. Compared to other countries, the Netherlands is rated relatively high in benchmarks that assess the state of open data of countries worldwide. In 2016, the Netherlands was ranked at place 7 by the ‘Open Data Barometer’, an initiative that ranks governments worldwide on how they publish and use open data for accountability, innovation and social impact (World Wide Web Foundation, 2016). Another global benchmark ranked the Netherlands on place 20 based on their openness (Open Knowledge Foundation, 2016).
Sampling
Cases are selected carefully, as a complete understanding of the motivations of open data users depends on choosing the right cases (Miles and Huberman, 1994). The sample is based on entrepreneurs that make use, are about to use, or have recently used open government data in the Netherlands. The definition of open data entrepreneurs followed is based on the description of entrepreneurs by Shane & Venkataraman (2000) and is formulated earlier in this study as: ‘people that are responsible of the process by which open data opportunities to
27 | P a g e
create future goods and services are discovered, evaluated, and exploited.’ Participants have
been interviewed when the following criteria were met: (a) the participant has in some way used open government data in the past 12 months or is about to use open data (b) the open data is offered or arranged by a local or the national government in the Netherlands; irrespective of which portal is used. (c) The participant is an open data entrepreneur according to the definition in this study or is about to become one. By applying these criteria, it is attempted to enable better comparison between cases and to make this study applicable for saying something about the current state of open data in the Netherlands. To select participants that meet these criteria purposeful sampling is used (Groenewald, 2004). Purposeful sampling allows the selection of specific participants to add extra value to a research (Groenewald, 2004).
Since an extensive overview of open government data users is not publicly available, participants were looked for in four ways. First, the department of the Dutch national government that is responsible for the Dutch national data portal (data.overheid.nl) was contacted. This department receives data request from open data users or re-users who are looking for specific datasets. Through a data requests re-users have to classify themselves as an entrepreneur or other type of user. During a meeting that was set up with two contacts of the department, they indicated to be interested in this study and were willing to support the research where necessary. Subsequently this department approached some people that had recently submitted a data requests and who had classified themselves as an entrepreneur. Four people responded indicating they were willing to participate in a research, who were subsequently sent an e-mail (see appendix B) by the researcher to arrange an interview. Three of them responded and were then interviewed.
Secondly, the network of the researcher was addressed through LinkedIn and Twitter. This led to seven references to potential open data users. All of the references were contacted through a private message on LinkedIn or by e-mail. Of the seven potential participants three were finally interviewed. Two of the other participants were excluded because they did not meet the criteria and the other two did not respond in the end. The limitation of this way of reaching participants is that it is more sensitive for selection bias, as participants are acquired through the network of the researcher.
Thirdly, snowball sampling was used to find additional participants. To do this, participants that already participated were asked if they knew other open data users. This resulted in two direct references who were subsequently contacted are were both willing to participate in an interview.
28 | P a g e
Lastly, the final three participants were found by browsing the internet for open data related projects. Two of the participants were found through the website of data.overheid.nl where applications of open data are presented, and the last participant was found through the website of UtrechtInc; an initiative that supports start-ups.
Case descriptions
In total 11 participants were interviewed. Given the diversity of the cases, it is important to understand each of the cases context (Saunders & Lewis, 2012). Brief descriptions of each of the participants and their work relating to open data are described below and summarized in a
table 1. A distinction is made in the way open data is used between open data as a ‘core’ or
main resource and open data as an ‘addition’ or supporting resource. The first refers to open data as an essential element or condition for the existence of a business or project, whereas open data as ‘addition’ refers to the use of open data to enrich a product or service. Participant six and eleven preferred to remain anonyms and are therefore given pseudonyms.
(1) Wij Verdienen Beter is a platform founded by Peter van der Wijngaart in 2012. Through his website he presents his own research about the transparency and accessibility of the Dutch government and municipalities. Open government data is used in some of his research as a supplementary resource.
(2) Geert Wirken developed and founded the platform Rijden de Treinen.nl. This platform presents travel and background information about the Dutch railways through a website and app attracting over 100.000 visitors a month. Among other things, his product indicates whether there are activities or malfunctions in public transport that might cause delays, it includes a travel planner and it enables comparison between trajectories on their amount of failures by using historic information. Rijden de Treinen.nl is run solely by Geert Wirken who also has a full-time job. Public transport information is used as main resource for the platform.
(3) Casper van Schuppen is one of the founders of Your Next Concepts, a start-up with 5 employees specialized in making smart and innovative (web)applications and conducting data-analysis in the field of education. Where applicable they use open government data as an additional resource to combine with and enrich other data. Casper van Schuppen works full-time at Your Next Concepts.
(4) Lars Nieuwenhoff is co-founder and owner of GoOV, a social enterprise that helps elderly and people with a disability to travel independently at lower costs than
29 | P a g e
current travel solutions. In doing so the platform makes use of currently available technologies with open government data as one of their main resources. Lars Nieuwenhoff initially worked as director innovation at Siza where GoOV initially started as a side project. Later GoOV was separated and continued as an independent organization with Lars Nieuwenhoff, Siza and a technology company as co-founders. Lars Nieuwenhoff now works full-time at GoOV.
(5) Arnold van ‘t Veld is an employee at Nelen & Schuurmans, an organization with 50+ employees that develops products, delivers services and provides consultancy in the field of water management. In doing so Nelen & Schuurmans make active use of open government data to gather and process information. Arnold van ‘t Veld started working at Nelen & Schuurmans through an internship and now works as consultant using open government data as an additional resource for his work. (6) The sixth participant (from now on referred to as Michael) is an ICT developer at
an online toy store. Through his work he gained experience with datasets and got interested in open government data. After exploring publicly available datasets he requested a specific dataset for the development of an app.
(7) Stefan de Konink is involved with open data in different ways. Initially he got in touch with the lack of available data through his study, which eventually resulted in founding Stichting OpenGeo to address these issues. Stichting OpenGeo, a Dutch non-profit foundation stimulates the availability of geospatial information and aerial photography. As part of the foundation Stefan pursued several open data projects including OpenStreetMap and openkvk. His most recent projects are
openOV and ndovloket.nl, which provide real-time travel information for data
re-users.
(8) Jelle Kamsma is data journalist, founder and owner of LocalFocus. LocalFocus is a news platform with six employees that provides visualization tools and publishes newsworthy data for (local) newspapers. In doing so they request and use open government data. Before he started LocalFocus, Jelle Kamsma worked at a national news site as data journalist.
(9) Since 2009, Irma Miedema is senior researcher and advisor at Sardes, a research and consultancy agency that supports innovation processes in the field of education by connecting research, policy and practice. Within Sardes Irma Miedema actively promoted and initiated the use of open data as an additional service of consultancy projects.
30 | P a g e
(10) Arjen de Jong is co-owner of Huispedia, a house platform that provides an easily accessible, transparent housing market for (potential) house buyers and sellers. The platform enables people to sell and buy their house themselves. This is done by providing house profiles that include the estimated market value and public available characteristics of each house. By providing information of houses, open government data is used as main resource of the platform. Arjen de Jong got involved with Huispedia through his interest in and knowledge of the housing sector, whereas his colleague is the founder of the platform and has a more technological background.
(11) Since January 2017, participant 11 (from now on referred to as Jessica) is environment advisor at BMD Advies, a company with 75+ employees that supports companies in their operational management in terms of environment, health, safety, quality and management. Open government data is used to provide companies with information about the environment. Jessica uses available open data as part of the consultancy work and requests additional data when necessary.
31 | P a g e Table 1 – Case description overview F ir m /i n it iat ive In d u st ry C u st om er s, t ar ge t gr ou p O p en d at a-us e R es ou rc e typ e Ro le p ar ti ci p an t P 1 Wi j V er di ene n B et er P ubl ic inf or m at ion se rvi ce C it iz ens & m uni ci pa li ti es A ddi ti on F ounde r, ow ne r P 2 R ij de n de T re ine n. nl T ra ve li ng in for m at ion se rvi ce C it iz ens C or e F ounde r, ow ne r P 3 Y our N ext C onc ept s D eve lopi ng and da ta -ana lys is B us ine ss & e duc at ion A ddi ti on Co -f ounde r, c o-ow ne r P 4 G oO V T ra ve li ng in for m at ion se rvi ce P eopl e w it h a di sa bi li ty A ddi ti on/ cor e E m pl oye e, c o-founde r, ow ne r P 5 N el en & S chuur m ans W at er m ana ge m ent cons ul ta nc y G ove rnm ent a ge nc ie s A ddi ti on E m pl oye e P 6 D eve lopi ng app P ubl ic inf or m at ion se rvi ce C it iz ens C or e Ini ti at or P 7 S ti cht ing O pe nG eo; O pe nS tr ee tM ap; ope nkvk; ope nO V ; NDOV -l oke t T ra ve li ng in for m at ion se rvi ce s, ge os pa ti al inf or m at ion, bus ine ss inf or m at ion et c. O pe n da ta ( re -) us er s C or e F ounde r, ow ne r P 8 L oc al F oc us M edi a and jour na li sm N ew s age nc ie s A ddi ti on/ cor e F ounde r, c o-ow ne r P 9 S ar de s E duc at ion E duc at iona l ins ti tut ions A ddi ti on E m pl oye e, ini ti at or P 10 H ui spe di a Re al e st at e H om e se ll er s and buye rs C or e Co -ow ne r P 11 B M D A dvi es C ons ul ta nc y B us ine ss es A ddi ti on E m pl oye e
32 | P a g e
3.2 Data collection
To ensure the rich rigor of this study, a detailed description of the data collection process is provided (Tracy, 2010). The following section describes the collection method and procedure that has been used to gather data.
3.2.1 Method
In-depth semi-structured interviews were considered the most appropriate for this study because with semi-structured interviews much more detailed information can be gathered compared to other data collection methods (Yin, 2013). Furthermore, qualitative research is particularly useful in areas that need further exploration (Boeije, 2005). As motivations to start re-using open data might strongly vary, semi-structured interviews allow the examination of the context for explaining these differences. The previous discussed literature on entrepreneurial motivations, open data supply and open data governance have formed the foundation of the semi-structured interviews. Together with other relevant aspects that were encountered in the literature review a topic list was composed that served as a solid and similar protocol for every interview. A topic list is generally described as a list with accessible topics that are relevant for answering a research question (Boeije, 2005). The topic list that was composed for this study is included in the appendix (see appendix C).
The interviews were carried out with entrepreneurs throughout The Netherlands. Interviews were held at a location of the interviewees choice, such as work offices (P2; P3; P5; P7; P8; P9), public spaces (P10), at a participant’s home (P1; P6), in a car (P4) and in one case through skype (P11).
3.2.2 Collection procedure
Primary data is gathered through semi-structured structured interviews with open-end questions. The topic list (see appendix C) is used to guide the interview. To avoid that participants would withhold information, confidentially was guaranteed when desired (Saunders & Lewis, 2012). Additionally, the set-up of the semi-structured interview allowed the researcher for asking additional questions to gain a complete overview of the context of each case (Saunders et al., 2011). To prevent bias and ensure the reliability of the study, open questions were carefully formulated. Additionally, secondary data, such as information from websites, is used to gain knowledge about the organizations involved and the participant’s backgrounds. According to Saunders et al. (2011), prior knowledge about the participant helps to increase reliability.
