DIGITAL INFRASTRUCTURES:
OPPORTUNITIES, NECESSITY AND PITFALLS
Researching the dependency of start-ups on established digital media
companies
In partial fulfilment of:
MA New Media & Digital Culture
University of Amsterdam
Submitted by:
Tijs Koops
Date: 03 September 2018 Supervisor: dr. J.A.A. Simons Second reader: dr. A. Gekker
Tijs Koops (10559922) Tijs.koops@live.nl
University of Amsterdam Faculty of Humanities Master New Media & Digital Culture
28 June 2018
Special thanks to my supervisor dr. Jan A.A. Simons for his guidance and support throughout the entire writing process. I would also express my gratitude to the people who were willing to contribute
Abstract
Our society has entered a golden age of digital innovation because of a relatively cheap and increasingly easy to use worldwide digital infrastructure. There is however a problem with the monopoly position of the companies who provide these digital infrastructures. Only a few big players
dominate the market where every company is dependent on nowadays. This is particularly problematic for businesses still in the start-up phase. It is this insecure but important sector, which receives a lot of attention from companies such as Google and Amazon. With this trend, start-ups lose
their independency. That is why in this research digital infrastructures were analysed from two perspectives: the website startup.google.com and companies in the start-up scene of Amsterdam. Both
cases represent a different side of the start-up ecosystem. The analysis of ‘Start-up with Google’ provided a two sided answer: being dependent on digital media companies is problematic because of
(1) privacy issues and the data integrity and (2) because these companies completely manage and control the digital infrastructures and therefore influence the way in which start-ups have to run their businesses. This research also provides an insight in the decisions entrepreneurs make relating to the digital infrastructures they use for their businesses. The main finding is the ease with which start-ups
make use of these services. They do not recognized a dependency problem and experience this tendency as a matter of course. Therefore, the findings ask for more awareness on this topic.
Keywords:
Index
1. Introduction ... 8
2. Theoretical framework ... 12
2.1. Defining digital infrastructure ... 12
2.2. Demarcate the start-up scene ... 15
2.3. Digital entrepreneurship ... 18
2.3.1. Hosting ... 20
2.3.2. Distribution ... 22
2.3.3. Marketing and advertising ... 24
2.3.4. API’s ... 25
3. Problematizing digital infrastructure ... 26
3.1. Cases ... 26
3.2. Methods ... 29
4. The case: Startup with Google ... 30
4.1. Hosting with Google ... 31
4.1.1. Google Cloud Storage ... 31
4.1.2. Google Cloud Web Hosting ... 34
4.2. Distribute your product with Google ... 36
4.2.1. Google Play Store ... 36
4.3. Marketing and Advertising with Google ... 38
4.3.1. Google AdWords ... 38
4.3.2. Google AdSense ... 39
4.4. Google’s specific API’s ... 41
4.4.1. Google Maps ... 41
5. The Start-up scene in Amsterdam ... 44
5.1. “It’s a no brainer… unfortunately” ... 44
5.1.1. Why? Because it is easy to use! ... 44
5.1.2. Digital infrastructures are inexpensive ... 45
5.1.3. Expertise in the ecosystem ... 47
5.2. A new perspective ... 49
6. Conclusion ... 51
7. References ... 55
7.1. Webpages used for this research ... 61
1. Introduction
“For a major part we are dependent on other digital media companies. We are dependent on Google for search and Facebook for advertisement (…). We are also dependent on the app-stores, so Apple and Google again. Of course we are dependent on Amazon for infrastructure (…). So, Google, Apple, Facebook and Amazon, if you look at where the traffic comes from, we can’t do without them.”
~ CTO Peerby (Interview 3)
The business structure outlined in the quote above is common for many companies today. Our society has entered a golden age of digital innovation, because of a relatively cheap and increasingly easy-to-use worldwide infrastructure. This development increased the emergence of social media, big data analytics, cloud computing, automated systems and machine learning (Fichman 330).
In his article The internet as infrastructure Sandvig argues that infrastructure traditionally has been defined as ‘the subordinate parts of an undertaking’ and that it is often used in a similar sense as public utility (90). He argues that an infrastructure traditionally refers to ordinary public services like gas, water, building roads and electricity. Digital media practices, as listed in the quote above, have become ordinary products and services in our everyday lives as well. In this research these products and services are referred to as digital infrastructure:
Computational environments owned by digital media companies offered for sale in the form of products or services to private or professional entities.
A digital infrastructure – explained in more detail in the next chapter – is build on a Global Information Infrastructure (GII), defined as “the worldwide information and communication systems that process and transport data inside and outside national boundaries” (Bowker et al. 98). Digital media companies like Google, Amazon and Facebook own these infrastructures and other companies and institutions depend on products and services offered by such companies for communication, distribution and marketing.
This tendency can be seen as problematic. In his book Media, Society, World: Social Theory and Digital Media Practices Couldry describes digital media practices as a “taken-for-granted part of our infrastructure” (8). He argues that this taken-for-granted status makes understanding the consequences of these practices difficult. This same argument can be found in Deuze’s concept media life. Deuze argues that our society is not using media, but is living in media (23). Peters also supports this position in his recent book The Marvelous Clouds: Toward a Philosophy of Elemental Media. Peters uses the following metaphor to make sense of the question how the ‘taken-for-granted’ gets constructed: “How did the water ever become invisible to the fish?” (35). In addition, he argues that infrastructure is not only invisible by nature, but also camouflaged by design. By using examples of
infrastructure buried in the ground or designed to blend in with scenery Peters refers to the physical invisibility of infrastructures. In addition, in his book he also refers to a more abstract invisibility: “Because of their vast technical complexity and costs, infrastructures are often cloaked from public scrutiny, their enormous risks and unintended consequences shielded from open debate” (31). Seen from this perspective, infrastructures are often black-boxed: “A system whose workings are mysterious” (Pasquale 3).
This black boxing concept can also refer to practices of companies owning digital infrastructures. An example is the conflict between Facebook and the company Cambridge Analytica. Cambridge Analytica was able to collect the data of millions of Facebook users. Facebook’s CEO Mark Zuckerberg called it a “breach of trust” (Chaykowski), but the problem remains that users are unaware that companies are selling their data.
However, this black box mechanism is not always intentionally, especially in the context of digital infrastructures. The digital infrastructure is computational and this makes it hard to grasp according to Manovich. In his book The language of New Media Manovich explains that, according to a computer, the ontology of the world consists of only two things: data structures and algorithms, which makes the infrastructure of the digital world very abstract (223). Kyong Chun also shares this notion in her book Programmed Visions: “The computer has a notion of rendering everything transparent, but their hardware and software are anything but transparent” (17). The complexity and black boxing – intentional or not – are central themes in this research and can be seen as two of the major problems of digital infrastructures. To create a short, but complete image of the problem of digital infrastructures, one more aspect of the contemporary digital infrastructures is also important to mention.
In addition to the problem of the ease with which companies use the infrastructure of companies like Google, Amazon and Facebook and the black boxing problem of these public utilities, another often-critiqued aspect is the abuse of power of the companies who own these infrastructures. The European Commission confirms this problem and even fined some Silicon Valley companies multiple times in the past few years. Last year Google had to pay a €2,4 billion fine, because of manipulating search results to favour their own shopping service (Boffey) and this month (June 2018) the European Commission alleged that Google forced licensing conditions for Android that favoured products like Google Play and Chrome (Toplensky). Another well-known example is the European Commission case against Microsoft. In 2013 Microsoft has been fined €500 million because they dropped the browser choice pop-up screen and just promoted Internet Explorer, their own Internet browser (“Microsoft fined by European Commission over web browser”). Deservedly the on-going debate in politics is about whether to regulate companies like Google, Amazon and Facebook as public utilities or not.
The debate about regulations is also being held in the academic field. In his newest book How to fix the future Keen researched these measures taken by the European Commission. He
acknowledges that the EU’s warnings and the threat of fines are “enormously valuable” (n. pag.). He also confirms the importance of a new law that will come into force this year, the General Data Protection Regulation (GDPR). This law needs to ensure that privacy will become the norm and tackles the big companies from Silicon Valley active in Europe. However, Keen questions whether this EU regulation approach is beneficial or unfavourable for Europe, its economy and especially for innovative EU tech innovators (n. pag). In his book Keen quotes the Financial Times: “Will it [EU regulations] help or hurt European start-ups?” Finding an answer to this question is not the aim of this research. However, this discussion is of value because it exposes the complicated relation between Silicon Valley-based digital media companies like Google, Amazon and Facebook and innovative start-ups.
The literature above shows the potential problems of the often taken-for-granted digital practices. Digital infrastructures should not be experienced as normal public utilities. In the following chapters digital infrastructure will be critically analysed for a better understanding of digital practices. Beside this in-depth analysis of digital infrastructure this research also focuses on a group that depends on digital infrastructure professionally: entrepreneurs with their own start-up.
Analysing from another perspective – that of the users of digital infrastructure – can help understand why these digital practices are taken-for-granted. In addition, it is interesting to see if users are aware of the potential problems of these digital infrastructures. The reason for using start-ups as the object of study is twofold.
The first reason is the vulnerability of these particular users in the context of digital infrastructures. Start-ups depend on digital infrastructure in two ways: (1) they make use of the emergent technologies accessible via digital infrastructure and (2) they rely on digital channels to grow fast on a worldwide scale. This will be further explained in the next chapter.
The second reason is the contribution of the start-up industry for the development of a country. The start-up sector is the industry that comes up with innovations that may transform the knowledge base and social, cultural and economic practices in a country (Simons 14). It is an important sector for countries that drive on a knowledge and innovation based economy, like The Netherlands (The Global Innovation Index 2017 ed. Dutta, Lanvin and Wunsch-Vincent).
Meanwhile, the start-up industry is also an unstable sector. Ninety per cent of the start-ups do not make it through the first year (Vervest). It is this insecure, but influential sector that receives a lot of attention from companies such as Google and Amazon. Due to the trend described above, the independent status of innovative start-ups can be jeopardized. Support of Google and Amazon could lead to a deceleration of the creative and innovative process of start-ups, because the support comes with guidelines, rules and obligations (Jansen and Bloemendal 195). There are many examples of established digital companies that have locked-in start-ups from the start by offering these start-ups
their digital infrastructure. An example of this mechanism is a new assistant ecosystem platform Google launched in May 2018:
“Google will work with start-ups directly to provide them with mentorship, and advice from engineers, product managers and design experts. The start-ups will also get early access to new features and tools, as well as access to the Google Cloud Platforms and promotional support” (Lardinois, “Google will start investing in early-stage startups that use the Assistant”).
With this platform, Google supports start-ups and offers them services and helpful products. Beside this lock-in mechanism, there are also examples of start-up take-overs. A study from Luckerson in 2015 showed the number of acquisitions Google has conducted: 221 start-up founders joined Google between 2006 and 2014. Luckerson argues: “The big businesses want to ensure they do not miss out on the next big thing”. Other examples are Facebook buying WhatsApp for $22 billion, Amazon buying the streaming site Twitch for $970 million and Apple who bought over thirty companies in 2014 (Luckerson).
In conclusion, the goal for this research is to investigate the relationship between start-ups and digital infrastructures of established digital companies like Google, Amazon and Facebook. This research will provide an answer to the following questions:
- What are the risks for start-ups when they are dependent on established digital media companies?
- Are entrepreneurs with a start-up aware of the dependency on established digital companies and do they identify this as an issue?
To answer these questions, two different case studies will be conducted. The first case is an analysis of the platform Start-up with Google, the place where Google presents all the products and services that entrepreneurs may need for starting a new business. The purpose for this case study is to get an insight in how established digital media companies present their services and products. In addition, an in-depth analysis of these products can help understand the advantages and disadvantages of using these digital infrastructures.
The second case study will be an analysis of conducted interviews with entrepreneurs in the start-up scene of Amsterdam. This sector is diverse and shows a variety of companies (StartupDelta). Not all these companies are necessarily selling digital products, although start-ups in the agricultural technology, biotechnology and medical technology are dependent on digital infrastructures as well. Thus, in this research the focus will be less on the products the interviewed start-ups have to offer and more on the digital infrastructures start-ups use for support in their business model. With these case
studies and the already existing literature around start-ups and (digital) infrastructures this research contributes to the academic field with more practical findings about the relation between start-ups and established digital companies.
2. Theoretical framework
Before the two research questions can be answered, it is first important to analyse to what extend start-ups are dependent on companies like Google, Amazon and Facebook. How is the digital infrastructure of the start-ups organized? What is the scope of this problem? A literature review on the different aspects of start-up infrastructures can help to delineate an overview in a structuralized way. The focus will be on the literature around some of the core business parts of start-ups today: hosting, distribution, marketing and advertising.
Before the digital infrastructure of start-ups will be researched, first the two key concepts in this research (1) digital infrastructures and (2) start-ups will be demarcated. The next paragraph will start with general theory on infrastructure. When the digital infrastructure is defined, paragraph 2.2 will answer the question: when is a company a start-up? In the broadest sense a just opened bakery in a small village can also be seen as a start-up. However, the core business of a bakery is not dependent on digital infrastructures and in that sense not interesting for this research.
2.1. Defining digital infrastructure
Traditionally, infrastructures refer to physical objects. In Peters’ words: “Infrastructures are generally thought to be bulky and boring systems that are hard to carry, such as airports, highways, electrical grids, or aqueducts” (30). This connotation of infrastructure fits the trend of researchers focussing on the non-digital part of contemporary digital businesses and physical infrastructure. An example is Parks research on the physical objects of media specific infrastructure. This is also an example of a recent turn toward environmental ethics in media studies (Sandvig 100). Parks defines these infrastructures as media infrastructures: “The material sites and objects involved in the local, national, and/or global distribution of audio-visual signals and data (…) such as broadcast transmitters, transoceanic cables, satellite earth stations, mobile telephone towers, and Internet data centres” (356). According to Parks it is important to visit the sites and objects to fully appreciate media infrastructures (370). By exploring these locations a new critical methodology for analysing the significance of infrastructure can be developed. In addition, Parks argues that this opens possibilities for a “renewed focus upon critical issues of materialism, distribution, territoriality, and conceptual visualisation” (356).
Analysing media infrastructures as Parks does is one example of a focus on the non-digital part of infrastructure. Another example is the focus on (im-)material1 digital labour. In Is there a global
digital labour culture? Casili examines the dynamics of social exploitation through outsourcing of online tasks to non-western countries (1). He criticizes the way anonymous users have to do de-humanized labour as micro-paid click farmers or content moderators to circulate our commodified information every day (31). The same focus on (im-)material labour is also noticeable in Finns recent book What algorithms want. Finn uses the term algorithmic labour to describe the precarious human labour that is essential to make digital services work:
“Cloud warehouses and factories transpose computational logic to physical spaces; spanning the gap between the abstract and the implemented entails making them terrible places for humans to work” (134).
In the context of these findings, the tendency of start-ups experiencing digital services and products as public utilities is even more problematic, because of hidden labor and issues of materialism. To understand the relationship between start-ups and companies like Google, Amazon and Facebook the focus needs to be on the part of those infrastructures that are not hidden: digital infrastructures. To understand the concept of digital infrastructures, it is important to notice that infrastructures not always refer to physical objects. In Peters’ definition, infrastructures are more than just physical objects:
“Infrastructures can be defined as large, force-amplifying systems that connect people and institutions across large scales of space and time or big, durable, well-functioning systems and services” (31).
Thus, infrastructures can also denote more abstract objects like systems or services. Later in his book Peters argues that infrastructures can also refer to software, websites and protocols, also called soft infrastructure (32). In other studies, soft infrastructure only denotes “the everyday habits of the human operator” (Jewet and Kling quoted in Sandvig 92). Jewet and Kling argue that it is important to examine the human infrastructure to understand the hard physical infrastructure of computers. However, for this research soft infrastructure will also refer to software and interfaces, not only including habits of the operator and protocols. Therefore, in this research digital infrastructure is defined as follows:
Computational environments owned by digital media companies offered for sale in the form of products or services to private or professional entities.
The concept of digital infrastructure goes beyond the traditional physical – and at the same time invisible – tubes and wires. The digital infrastructures researched in this thesis are the online
environments, servers and software located in the public sphere. The same perspective on digital infrastructures can be found in the article of Bowker et al. about information infrastructure. Bowker et al. refer to information infrastructure as the digital facilities and services usually associated with the Internet and added – to this traditional way of thinking – the “social and organizational dimensions of infrastructure” (99). According to Bowker et al. infrastructures are no longer something exclusive:
“The internet and its associated information infrastructure is reaching beyond the immediacy of the physical and technical, extending to individual and community realms as well as into organizational and cultural structures and expectations” (Bowker et al. 106).
Information infrastructures are now accessible to the public, experienced as a public utility. This development and the scope of these new information infrastructures also carry new responsibilities. Bowker et al. warn for the taken-for-granted status of this mediation. They perceived issues of data sharing, data spaces, community standards and resource sharing and argue for a renewed focus on researching infrastructure (112). An in-depth analysis of digital infrastructures that are sold as products or services can contribute to infrastructure research with a more practical perspective.
There is one more aspect of digital infrastructure that has not been touched on yet. That is the technical complexity of the computation. Kitchin describes the complexity of cyber-infrastructure. The cyber-infrastructure is the infrastructure that makes our data-based economy possible:
“It consists of a suite of dedicated and integrated hardware and networked technologies (…) interoperable software and middleware services and tools, shared services (relating to data management and processing), analysis tools such as data visualisations, and shared policies (concerning access, use, IP, etc.), which enable data to be distributed, linked together and analysed (Kitchin 34).
It is because of this complex infrastructure that data as a product of cyber-infrastructure is critically revered to as ‘cooked’ instead of ‘raw’ (Gitelman 2). Data and data infrastructures are never objective, they are generated. In the work of Manovich and in the work of Kyung Chun, mentioned in the introduction, both researchers also noticed the complexity of these computational environments. They criticise the opaque character of computational environments in our contemporary society. This potential complexity and consequentially opaqueness can also be applied to the digital infrastructures in the remaining chapters of this research. This issue of digital infrastructures, in the introduction defined as a black-boxing problem, will be a central topic in the analysis of digital infrastructures this research.
2.2. Demarcate the start-up scene
The contemporary work environment is called “the golden age of entrepreneurship” by Chatterji and Robinson in their article on Tech Crunch (“The Golden Age of American entrepreneurship”). The up industry in Amsterdam is an example of this golden age, because statistics show that the start-up scene in the capital city of The Netherlands grew with 22 per cent in 2016 (“Amsterdam’s tech and startup scene continues to thrive”). The popularity of start-ups can also be found in the abundance of online articles on entrepreneurial life and start-ups2.
However, a well-founded definition of a start-up is despite the popularity of the term hard to find. The complexity of the term is demonstrated in the online articles on how to set up a successful start-up by giving multiple different definitions of a start-up. For example, a start-up is referred to as (1) a state of mind, (2) a fledging business enterprise, (3) a company working to solve a problem where the solution is not obvious and success is not guaranteed, or (4) a business or undertaking that has recently begun operation (Robehmed). All the above-mentioned definitions are partly true, some more abstract than others. However, a more precise definition can be found in the work of Paul Graham. Due to his experiences as successful entrepreneur, founder of Viaweb3, venture capitalist and
owner of the accelerator company Y Combinator Graham can speak with some authority. In his article Startup = growth Graham argues that the focus on growth is the only characteristic that differentiates start-ups from normal companies. In Graham’s words:
“Being newly founded does not in itself make a company a start-up. Nor is it necessary for a start-up to work on technology, or take venture funding, or have some sort of exit. The only essential thing is growth” (“Startup = Growth”).
The example Graham uses to support this claim is the one of a barbershop. He explains how a barbershop is not designed to grow fast, because the product cannot be sold to a big market. A barbershop can never be a start-up according to Graham, because the product barbers offer, is not scalable. Barbershops offer a service many people want to make use of, but few will travel far for a haircut. Hence, a scalable product is the first characteristic for defining a company as a start-up. As a result, to become a start-up with scalable products it is necessary to have access to the Global Information Infrastructure (GII). Therefore, Graham argues in his article that start-ups are successful especially because of the use of scalable software to reach and serve people all over the world. This
2
Some examples: 3 Things you need for a successful up (Forbes 2018); Want to build as successful start-up? Adopt these 3 mindsets immediately (Medium 2018); The formula behind San Francisco’s start-up success (Tech Crunch 2018)
3
Viaweb was one of the first web-based applications that allowed users to build and host online stores. The company was sold to Yahoo for $49 million (The New York Times).
already touches the idea of start-ups being dependent on digital infrastructure, because they need the right infrastructure to reach this online market.
The second characteristic for defining a company as a start-up follows from the notion of bohemian idealism in the new media work field. In his book No Collar: The Humane Workplace And Its Hidden Costs Ross defines the new media life as a bohemian life and calls the rise of new media start-ups the industrialization of bohemia (10). Ross argues that this new industry provides opportunities for people to reinvent themselves with “spontaneity and self-direction owed to the spirit of the bohemian culture” (18). That is why the work environment in new media productions is embraced with openness, cooperation and self-management.
This notion of bohemian idealism in the new media work field can also be found in Turner’s work about the cultural infrastructure of the new media industry. Turner found a joint celebration of creativity in his research object The Burning Man Festival. He argues that elements of this festival also shape collaborative manufacturing processes within the digital industry (73). Turner defines this as a “dramatic socialization of technical labor” (74). This development was made possible because of emergent technologies, especially the rise of the Internet. He uses the term ‘commons-based peer production’ to explain these new computer networks (76).
In addition, Kennedy focuses in her book Net Work: Ehtics and Values in Web Design on collaboration in a digital environment too. In her study on web standards – a form of self-regulation as argued in her book – she found that cultural work is “characterized by the voluntary sharing of knowledge” (89). She shows that this form of collaboration in the digital environment is challenging the idea of cultural work as an individualized practice. The mentioned collaborative nature based on a self-managing ideology can be seen as the second characteristic of start-ups and entrepreneurialism.
The last characteristic found in the literature above – self-management and (digital-)collaboration – fit perfectly in the ideology central to entrepreneurialism and start-ups: cyberlibertarianism. Liu explains cyberlibertarianism as “the belief that the technology and social covenants of networked information are a new form of politics” (240). Some characteristics central to this new concept are freedom from government, freedom of information, freedom from big businesses and – most of value for this research – freedom of the individual and entrepreneurial freedom. In this ideology the priority of self-management is clear. In the new digital age, the idea of starting your own company without interference is stimulated. This idea can also be found in Liu’s book when he quotes Barbook and Cameron:
“Cyberlibertarianism is the opportunity to become a successful hi-tech entrepreneur (…) Big government should stay off the backs of resourceful entrepreneurs who are the only people cool and courageous enough to take risks” (Liu 249).
Interestingly, the cyberlibertarian ideology is also against ‘big businesses’, according to Liu. Defending consumer privacy against these businesses is one of the main pillars of Internet freedom (Liu 244). A discussion about this topic, that took place in the late nineties, is still a hot topic today. The interviews later in this research will point out how entrepreneurs today think about issues such as privacy and with in mind the powerful position of companies like Google, Amazon and Facebook (originally build on a cyberlibertarian ideology).
Cyberlibertarians believe that all these new ideals like individual and entrepreneurial freedom can lead to more creativity and innovation. In this way start-ups can be seen as companies involved in solving problems for which there are not yet clearly defined solutions (they come with innovations for these problems). The solution is often technology-related by using emergent technologies, but this belief in technology as a solution is also motivation for criticism. Morozov criticizes the way the tech industry recast all complex (social) situations as neatly defined problems with computable solutions, if only the right algorithms are in place (16). Morozov calls this concept solutionism and for him this is central to companies in the innovation industry today.
A second point of criticism on the new media work environment is the precariousness of the entrepreneurial work and life. The industry is environ with a premise of “friction free” labor (Brophy 619), exiting jobs (Neff, Wissinger and Zukin 307), openness and transparency (Marwick 110), a celebration of creativity (Turner 75) and democratic politics (Ross 20). However, the researchers above show that the life of an entrepreneur is not always as cool as promised. They focus on the willingness of workers in this field to take the precariousness of the work for granted. Precarity refers to the “material and existential insecurity suffered by workers as a result of flexible employment arrangement” (Brophy 619). This precarity results in a more explicit, individualized, profit-oriented risk in a winner-take-all environment (Neff, Wissinger and Zukin 310). The precarity of the work is not the topic of this research, but it shows the willingness of entrepreneurs to make their start-up a success despite the high risks of failing.
What can be concluded of the literature above, is that the start-up industry has a history of an entrepreneurial lifestyle that follows a cyberlibertarian ideology. Because of this development, start-ups can be seen as companies with owners who come up with creative and innovative products to solve problems. New technological possibilities and individualized freedom of networked information give these entrepreneurs the possibility to grow their scalable products fast. The importance of self-management paradoxically goes hand in hand with the sharing of knowledge and collaboration made possible by the opportunities of the Internet. At the same time the start-up scene is an industry with high risks and precariousness work, what makes the industry insecure with a high change of failing.
To summarize, the following definition of a start-up will be used in the remaining part of this research: a start-up is a self-managing company that tries to find innovative solutions for contemporary problems and is able to grow fast with the help of emergent technologies and the sharing of knowledge in a digital environment.
2.3. Digital entrepreneurship
Due to the digital opportunities of our contemporary society as explained before, start-ups are successful. The remaining part of this chapter is a literature overview of the main digital practices start-ups generally rely on. This new type and successful way of starting a business is also called digital entrepreneurship. A quote from Davidson and Vaast can be used to define the term digital entrepreneurship:
“We refer to digital entrepreneurship as the pursuit of opportunities based on the use of digital media and other information and communication technologies. Digital entrepreneurs rely upon the characteristics of digital media and IT to pursue opportunities” (2).
To refer to digital entrepreneurship as the pursuit of opportunities based on the use of digital media as Davidson and Vaast do, seems obvious, but the significant role of digital media should not be downplayed. This position of digital entrepreneurship can exist due to fact that digital services and products have become so evolved, cheap and ubiquitous. Therefore, digital services can be easily combined and recombined (“A Cambrian moment”).
In his article Digital entrepreneurship: Toward a digital technology perspective of entrepreneurship Nambisan acknowledges the popularity of extern digital platforms: “The potential for new ventures to deepen their specialization while offsetting their production, marketing, and distribution capabilities explain the attractiveness of digital platforms as a venue for entrepreneurship” (1032). Because of this new potential, entrepreneurial businesses are less bounded and less predefined. Flexibility is a goal for every start-up and outsourcing infrastructures can help to achieve this (Mikkola and Skjoet-Larsen 38). Outsourcing of infrastructure can be a positive development, but Nambisan stresses that it helps address the question ‘who gets to play’ in distributed entrepreneurial initiatives on digital platforms (1036).
Deuze shares this less positive view on emergent technologies and states in his article Managing media workers the following: “Arguably the most powerful factor to be considered in managerial strategy for (media) firms is the role of technology” (332). With this statement, Deuze not only stresses the role of the media industry to ‘supercharge’ technological innovation. He also refers to the impact of technology for the media industry ecosystem itself. Deuze problematizes the way “disruptive potential of increasingly ubiquitous and pervasive information and communication technologies, wresting control over all aspects of the media value chain” (336). A quote from the article A Cambrian moment shows the scale of this potential problem:
“[API’s] allow one service to use another, for instance voice calls (Twilio), maps (Google) and payments (Paypal). The most important are ‘platforms’ – services that can host start-ups’ offerings (Amazon’s cloud computing), distribute them (Apple’s App Store) and market them (Facebook, Twitter). And then there is the Internet, the mother of all platforms, which is now fast, universal and wireless”.
However, there are researchers who approach this tendency with less criticism. Isenberg emphasizes the importance of infrastructure for the success of entrepreneurship is his model of six domains of the entrepreneurship ecosystem. Isenberg defines in his paper The Entrepreneurship Ecosystem Strategy as a New Paradigm for Economic Policy infrastructure as one of the six essential domains for entrepreneurs, beside policy, markets, finance, culture and human capital. In his overview he names infrastructure ‘supports’ (Figure 1).
Figure 1 - Isenberg's six domains of the entrepreneurship ecosystem
Source – Isenberg (2011)
In the supports-domain Isenberg, distinguishes infrastructures, support professionals and non-governmental institutions. Infrastructure in Isenberg’s article does not specifically refer to digital infrastructures. The focus is more on general infrastructures like telecommunications, transportation and energy. The connotation of infrastructure as a ‘support’ is closely related to the notion of infrastructures as ‘taken-for-granted’ public utilities in the introduction of this research. Looking at digital infrastructures in this way is a perspective that will be further examined in the second part of this research. The question is if the interviewed entrepreneurs share this point of view and indeed experience digital infrastructures as public utilities.
To analyse this support-domain - in particular the digital products and services – of the entrepreneurship ecosystem in more detail, digital infrastructures are first divided into four domains: hosting, distribution, marketing/advertising and API’s. There are many digital infrastructure products
and services to choose from today. This categorization structuralizes the complexity of digital infrastructures. These domains were found as the main digital practices on digital infrastructure in the literature written about digital marketing, infrastructures and business planning. The main source for this categorization is the book Software Business: from physical products to software service and solutions edited by Herzwurm and Margaria. This book contains articles from various researchers who published about software and digital technology as the key area of innovation (VI). According to Herzwurm and Margaria, the digital market is a challenging domain for research, because “software companies have to depend on one another to deliver a unique value propositions to their customers” (VII). Dependency is one of the themes in their book and three of the four categorized domains for this research are central topics: hosting infrastructures, digitalization of distribution channels and open source software (API’s).
2.3.1. Hosting
The first domain, hosting, is a popular topic in the bundle of Herzwurm and Margaria (Laatikainen, Ojala and Oleksiy; Schütz, Kude and Popp; Phaphoom, Wang and Abrahamsson) and this domain can also be found in other books and articles (Wu et al; Vaquero et al; Velden and Kruk). Hosting servers are often reviewed as a key resource for companies nowadays (“Using Business Model Canvas to Launch a Technology Startup or Improve Established Operating Model”). Hosting is the process of the storage of a website or other electronic data on a computer connected to the Internet. The data is stored on a location outside the company, instead of running the programs on an individual local server. It is often brought in relation with the term cloud storage or cloud computing. This new technology and way of organising data gives a company access to their applications and data from anywhere in the world (Wu et al. 380). This certainly applies to start-ups that run an interactive platform and this is the reason why especially cloud computing is an essential part of digital infrastructure, according to recent studies. Vaquero et al. define clouds as follows:
“Clouds are a large pool of easily usable and accessible virtualized resources. These resources can be dynamically reconfigured to adjust to a variable load, allowing also for an optimum resource utilization” (51).
According to Wu et al. data storage is one of the primary uses of cloud computing (380). The data is stored on a third-party server like Amazon Web Services or Google Cloud Storage. Beside data storage, cloud computing also offers a new market for ‘infrastructure as a service’, ‘platform as a service’ and ‘software as a service’. Therefore, this new way of organising data for companies results in new interesting business models for providers who own these servers. Companies like Amazon, Google, and Microsoft started to establish new datacentres for hosting these cloud services. The
following example shows how extensive this business is. Amazon Web Services – the cloud-computing services from Amazon – posted $4.57 billion in revenue last quarter 2017 (Miller).
The question remains why companies outsource these services to clouds. Wu et al. argue for the following reasons. First it is the ease of management that makes it interesting for companies to use cloud services. The cloud service often comes with an easy-to-use interface that is applicable to every company. As Wu et al. state: “often all that is needed to manage your storage implementation is a simple web browser leaving the headaches to the service provider” (383).
Second, outsourcing to these services is cost effective, because the maintenance of the software and hardware is now the problem of the service provider. Start-ups do not have the expensive costs to maintain these kinds of servers. Beside that, Wu et al. argue that the economies of scale of these data centres cannot be matched (383). Only the largest organizations have the ability to make it a profitable business. That is why there are only a few big players on the market. The third and fourth reason to make use of cloud services is (3) the lower impact outages and upgrades and (4) disaster preparedness. These cloud-computing services often provide a 24/7 service. The service provider is responsible for updates and the data stored on a cloud-computing server can also be seen as a back up in case of disasters. This reduces the change of a failure.
Finally the use of cloud computing enables a start-up to be flexible. Wu et al. state: “Cloud storage solutions free the IT manager from detailed capacity planning. Cloud-bases solutions are flexible and provide storage as needed” (383). The flexibility in combination with the cost effectiveness of these cloud-computing services seems to make them a rational choice for every start-up. For Phaphoom, Wang and Abrahamsson the biggest advantage of cloud computing and outsourcing this infrastructure is the opportunity to focus on “core competencies and to run business without technical employments” (148).
However, Wu et al. also acknowledge some difficulties with these cloud-computing services. They name security, power, costs, reliability and the most problematic one: data integrity. Data integrity and privacy issues are recurring themes in the contemporary digital environment. Data integrity is problematic from two perspectives, both not mentioned in Wu’s article. The first question is: what does a company like Amazon with all the data they collect? As mentioned before, companies like Amazon, Google and Facebook can be judged as black boxes to a certain point. Users do not exactly know what these companies do with the data they collect. The policies, practices and performances of a data infrastructure of a company have to be audited and measured according to Kitchin, but the complexity of these platforms makes this task difficult (33).
The improved terms and conditions helped these companies to become more transparent to a certain point, but mistakes like the data breach with Cambridge Analytica can always happen. The technology these companies work with is complex and difficult to grasp, even for the tech-companies themselves. That is why Pasquala argues the following in her book The black box society:
“I will repeatedly make the point that transparency is not enough, particularly in the finance sector. When companies parry with complexity too great to monitor or understand, disclosure becomes and empty gesture” (16).
So, privacy issues are always a problem. For example, if something goes wrong with Amazon Web Services 47,1% of the public cloud market and 41,5% of the market using IAAS have a problem (Coles).
All this data in one place is also problematic due to another reason, namely the legal obligations of the platforms providing these services. Amazon for instance is an American company settled in Seattle (Washington) and this gives the US-government the right to claim all possible data of every company that uses AWS. In their article Captives of the Cloud Van der Velden and Kruk acknowledge and criticise this tendency. They explain how this cloud service infrastructure enables the US-government to sue and claim non-US data outside the jurisdiction of the USA Patriot Act. This means that the US-government has access to the entire cloud infrastructure worldwide, because almost all the big companies providing these services are established in the USA: Apple, Facebook, Twitter, Google Amazon, Dropbox and Microsoft. Van der Velden and Kruk deservedly use the following quote in their article: “we are the voluntary prisoners of the cloud; we are being watched over by governments we did not elect”. This argument will also be analysed in the platform analysis of the Start-up with Google platform.
2.3.2. Distribution
The second domain, distribution, is another key resource and also an important topic in many of the by Herzwurm and Margaria bundled publications (Novelli; Jansen and Bloemendal; Zittrain). These researchers focus on the digitalization of distribution channels and the implications of this trend. Their research mainly focuses on application stores and websites as the main distribution channels for companies today.
For every start-up it is important to get the product to the right costumer and to find the right distribution channels to do this. Start-ups are particular dependent on a digital infrastructure for distribution, because they rely on these channels to reach an audience on a worldwide scale. This can be illustrated by looking at developers. Potential customers can only be reached through app-stores like Android Play Store and Apple App Store. There are a few other app app-stores available, but Android and Apple are by far the biggest in this market as shown in the figure below.
Figure 2 - Number of apps available in app stores as of 1st quarter 2018
Source – statista.com
Jansen and Bloemendal define an app-store as follows: “An online curated marketplace that allows developers to sell and distribute their products to actors within one or more multisided software platform ecosystems” (195). The use of the word ‘curated’ is on purpose, because the online marketplace is not open to everyone. This has to do with a so-called whitelist system. To understand this mechanism first the idea of a blacklist system should be explained. In his article Law and Technology: The End of the Generative Internet Zittrain explains this concept when he describes a new ‘gated-community’.
A blacklist allows through all elements except those on the ‘list’. This can help companies to deny extern software on their network. According to Zittrain, the new gated-communities work with a whitelist system and this changed the relation between app-developers and distribution channels like app-stores (19). With this new system only the programs on the whitelist are allowed. The providers of these app-stores make the decision whether an app is allowed or not. As an app-developer you first need approval to access this whitelist.
This is already an extra step and even if a developer makes it to the whitelist, the future of the app is not certain. If Apple for example decides that their own product or application works better, they always have to possibility to remove apps from their whitelist. Zittrain argues: “when Facebook kills an app, the app is naturally not only unavailable to new users, but disabled for current ones, too. So Superwall or Secret Crush can go from millions of users to zero in a heartbeat” (20).
This particular digital infrastructure is not freely accessible for everyone. The question is, if this is fair for start-ups without awareness. In a regular market without an intermediary the user decides whether a start-up will succeed or not, in the example above Apple can make a decision before potential customers have access to the product. Zittrain argues the following:
“People do not know what they are missing (…) who would have invested in Wikipedia at the beginning? And if Wikipedia required an incumbent gatekeeper’s approval or permission to get started, it might have failed to receive it – or languished at the bottom of a to-do list among hundreds of other apps and services awaiting review (20).
The Google appstore is definitely one of the services that ask for further research in the following chapters, because for this aspect of the business model start-ups do not have a choice. There are only two big companies to reach an audience and they control the market entirely because of the whitelisting systems and strict guidelines of their platforms. Before diving into this more specifically, the next chapter will first focus on another major infrastructure needed for running a successful start-up: the channels for marketing and advertising.
2.3.3. Marketing and advertising
Just like hosting and distribution, marketing mechanisms of start-ups also run on a digital infrastructure. Using Google and Facebook to reach customers is not specific for digital entrepreneurs and start-ups; every company can use this marketing strategy. However, as Simons argues, start-ups do need this strategy in the contemporary digital era. Pushing products to the consumer is no longer enough. According to Simons “a producer must ‘pull in’ consumers by engaging them to participate in and interact with the process of production” (18). That is why social media is such an important medium for many start-ups. The interaction between producers and consumers is important and the platforms like Google and Facebook can offer this communication.
The core business of a company like Google according to Vaidhyanathan is “colleting information about our habit and predilections so that it can more efficiently target advertisements at us” (9). All the products Google offers, contribute to achieve this goal. Whether it is Google AdWords, Google+ for business, Google Analytics, Google Trends, G-suits or Google Multiscreen Strategies, start-ups are happy to use them and Google collects all the data created. This aspect of the advertising market will be examined in more detail in paragraph 4.2.1.
Beside the importance of these platforms for the relation with costumers, they are also essential for business-to-business (B2B) relations. Networking and marketing of the business is indispensable for a digital entrepreneur. The discussed articles from Turner and Kennedy in paragraph 2.2 already showed the importance of project-based labour and Deuze shares this idea. He argues that these projects occur within a company, “as well as between companies, and between companies and
individual cultural entrepreneurs (27). For these projects to succeed platforms can be helpful for networking and the creation of co-working environments.
In addition, in their article Net Working: Work Paterns and Workforce Policies for the New Media Industry Batt et al. state that one of the challenges for companies in the new media industry is the spirit of collaboration that pervades the industry (60). Batt et al. argue that if companies can benefit from this spirit they can also benefit from the flexibility of the industry and more importantly “they can build or secure the new instruments and institutions it will need to survive and flourish”. With this statement Batt et al. imply, that if a start-up decides not to use the support of these platforms they can also not benefit from the flexibility of the ecosystem. The analysis of the conducted interviews will tell how entrepreneurs experience the benefits of the ecosystem.
2.3.4. API’s
Finally the last domain relates to the open source software culture and business dependency on (free) API’s. This is also a central topic in the bundle from Herzwurm and Margaria. Various researchers analyse the so-called software-ecosystem (Lucassen, Rooij, and Jansen; Handoyo, Jansen and Brinkkemper; Schütz, Kude and Popp). An important aspect of this ecosystem is the use of API’s.
API is short for Application Programming Interface. The website techterms.com provides the following definition: “An API is a set of commands, functions, protocols, and objects that programmers can use to create software or interact with an external system” (API). In other words, API’s can be seen as packages of pre-coded software. In there article From the Service-Oriented Architecture to the Web API Economy Tan et al. state that web services and in particular these Web API’s are the backbone of the Internet (64). They refer to an API as a ‘mashup’ of software, a service that combines multiple other services.
Hence, web application developers do not develop every part of their service themselves. A good example is the storage services AWS offers. These hosting services explained in paragraph 2.3. are also API’s: packages of software that can be integrated in a start-up as a building block. Another example is the increasing presence of machine-learning services. Tan et al. found that many business owners outsource these kinds of “non-essential but still very important features” to a third party (65). The popular machine-learning API’s are only noticeable on the back-end of a digital platform. However, there are also a lot of API’s that can be found on the user interface.
A good example is the often-used API Google Maps. A lot of websites have integrated this service from Google. Think about restaurant websites and their location information. The source code for integrating the API can be copied and pasted to a certain extent for free. Providing these software codes for almost nothing is a business model that has been applied to a lot of services from Google and Amazon, and also to social media platforms like Facebook and Instagram.
Take for example Facebook Connect. This free API provides websites the option to use Facebook accounts to register on another platform. These kinds of API’s are also called plugins (Gerlitz & Helmond 1328). Social plugins seem to work as a decentralization mechanism. Facebook algorithms are now spread to the whole Internet to offer services. However, Gerlitz and Helmond argue in their case study about Facebook: “while the Open Graph presents an attempt to decentralise opportunities to connect external web content to Facebook, it at the same time recentralises these connections and the processing of user data” (1355). Data runs through these API’s back to the service provider. That is why Google offers the Google Maps source code for practically nothing. These companies stimulate the use of these API’s by providing them for free, with the prospect of collecting more interesting and valuable data.
Offering online services for hosting, distribution and marketing/advertising provide the companies mentioned in the previous paragraphs a lot of valuable data. Thus, besides being managed by these companies another problem of being dependent is the issue of privacy and data integrity. This notion will be central in the remaining chapters of this research.
3. Problematizing digital infrastructure
Two different types of research will be conducted for this thesis. In this chapter the cases and the methods will be explained. Both cases require another type of method. The different types of methods will produce different sorts of outcomes. In the next paragraph an outline of the cases will be given.
3.1. Cases
The two cases for this research are (1) the website startup.google.com and (2) companies in the start-up scene of Amsterdam. These two cases are chosen because they both represent a different side of the start-up ecosystem. The website startup.google.com represents the companies that produce the services start-ups are dependent on and the companies in the start-up scene of Amsterdam represent those who use these services.
The purpose of the first case study – an analysis of the website startup.google.com – is to support the findings in the theoretical framework above. An in-depth analysis of one of the main actors of the digital infrastructure market can better help understand the advantages and disadvantages of using these digital infrastructures, since the specific moments privacy issues and problems with data integrity emerge can be demonstrated. In addition, the analysis can also distinguish to what extent established digital companies are able to manage start-ups.
The reason why particular this case – instead of other digital companies – is interesting for this research is bipartite. First, the Google products and services are one of the most used in the start-up ecosystem. It certainly is an institution every entrepreneur has to deal with. A telling example is the amount of professional users of the product G Suite. More than three million businesses pay for this
specific product alone (Lardinois). In addition, Startup.google.com is a sub-website of the Google for Entrepreneurs platform. This platform partners with start-up communities to help grow their businesses. The Google for Entrepreneurs platform also builds campuses where starters can learn, connect and create companies. On this platform – together with startup.google.com – an entrepreneur can find all the products he or she needs to create a successful company: tools to ‘grow your skills’, ‘bring your idea to life’, ‘understand your users’ and ‘reach your users’.
The second reason is the notion that Google can offer a complete digital infrastructure. The in the theoretical framework main distinguished digital practises (hosting, distribution, marketing and advertising and API’s) can be analysed in a practical way, by looking at the products and services of Google. For the domain of hosting the Google Cloud Storage and complementary Google Cloud Web Hosting services will be analysed. Both services display the in the theoretical framework criticized complexity of digital infrastructures with potential privacy issues and problems with data integrity as a consequence. The usability of the service will also be central to the analysis, because it is this characteristic of cloud computing that makes it a key resource for many digital entrepreneurs.
For distribution Google Play (store) will be analysed. This digital infrastructure is chosen, because it is the application store with the most professional users (app developers), as mentioned before in paragraph 2.3.2. Central to the analysis will be the in the theoretical framework criticised curated character of the infrastructure. Again, the analysis can help understand to what extent start-ups are dependent on digital infrastructure.
For the domain of marketing and advertising two complementary products of Google will be analysed. Both Google AdWords and Google AdSense are digital infrastructures start-ups depend on. As explained before in paragraph 2.3.3, online advertising is one of the main digital practices to reach an audience and new potential customers and the advertising market is also an infrastructure with a potential revenue stream for start-ups who earn money with selling ad-space to other companies. Beside the importance for the potential revenue and customers, both services are also essential for Google, because selling user generated data to advertisers is one of the main revenue streams of the platform. Having a closer look to the products that enable this revenue model can help understand what the implications are of using these digital infrastructures.
To analyse the last domain distinguished in the previous chapter, the Google Maps API will be further examined in the last part of this case study. The Google Maps API is a good example of the success of the API economy described in the theoretical framework. An in-depth analysis can help to better understand the advantages and disadvantages of using these software packages. Start-ups thrive on emergent technologies like Google Maps, but the question again is to what extent this results in a dependency problem.
Although in this case study only one digital company is analysed instead of the several established digital companies mentioned in the theoretical framework, the findings of this analysis will provide a more specific insight on how digital infrastructures work in general. These findings can then
be compared to the experience of the interviewed entrepreneurs in the second case study of this research.
To answer the question for the second part of this research – Are entrepreneurs with a start-up aware of the problematic tendency of being dependent on established digital media companies? – four companies within the start-up scene of Amsterdam were interviewed. The interviews have been conducted in Dutch (see paragraph 7.2. for transcriptions) and the quotes used for this research have been translated back into English. The start-ups are partly selected because of practical reasons. It is not easy to find people who have time to talk about their new ventures. Owning a start-up is a 24/7 job that takes a lot of time. To make it easier to find start-ups willing to cooperate, two different tech hubs in Amsterdam were contacted to make use of their network: ACE Incubator program and TQ. Both tech hubs help start-ups growing their businesses. The help consist of training, networking, coaching and providing a location to run the business from. ACE is located at Amsterdam Science Park and provides meeting rooms, workshop areas and flex-working spaces. TQ is located in the heart of Amsterdam and also provides their residents the right facilities to start their business.
Beside these practical points, the start-ups were also chosen because of more substantive reasons. The chosen start-ups all work with digital technologies and the products they deliver are software related. Start-ups with software related products were chosen for this research, because they cannot survive without digital infrastructures of other digital companies as explained in the theoretical framework above. The start-ups all meet the following characteristics: (1) the products are innovative; (2) they have potential to grow exponentially; (3) the business models are based on digital infrastructures provided by other companies. Within this sector of software-orientated start-ups, the following four companies were chosen to measure up to a broad representation.
Peerby is a business-to-consumer-company (b2c) and can be seen as a classical sharing platform. Peerby brings the supply and demand of daily stuff the customer needs together. The idea is that people do not value possessing stuff, but that borrowing stuff from someone else works fine as well. The product they deliver is a website and an app for android and IOS. Dyme is also a b2c company. It is a company that manages all the subscriptions their clients have. They are not launched yet and are currently testing their product. Dyme offers an app that automates the entire switching process for utilities, telecom and insurance subscriptions.
The next start-up CTque is a business-to-business company (b2b). CTque delivers software for hospitals. This software is used for analysing medical data of patients. The product CTque offers can be seen as a search engine that can be integrated in hospitals. This makes information of medical records of patients instantly available. The last start-up 904Lab is also a b2b company. 904Lab also delivers search engine. They focus on online webshops. They use artificial intelligent and machine learning technology for their service. The product 904Labs delivers is a software as a service solution.
Hence, the start-ups interviewed for this research are all software-orientated and within this sector they represent both b2b and b2c companies.
3.2. Methods
The method for researching the website startup.google.com will be a platform analysis. This consists of a text analysis and an interface analysis. The text and interface will both focus on the front-end of the platform. Therefore, in this research only the user (in this case the start-up) experience of the websites will be analysed. This research follows the platform architecture and design complemented by a focus on users (Plantin et al. 6). In this way the analysis can explore the friction between “on one side, users’ goals of expression and, on the other side, platforms’ profit-seeking aims” (van Dijck quoted in Plantin et al.).
In addition, this focus on the user experience is also they way the platform will be analysed. For the analysis the same steps of using the website will be made as an entrepreneur interested in the products and services would have done. Therefore, the focus will be slightly more on the text and products the platform offers, than the interface and functionality of the website itself. However, the interlinking of the different (sub-) websites can be interesting for the outcome of this analysis. By looking at these aspects the analysis is “examining the assumptions built into interfaces as the normative or path of least resistance – though, like all norms, this is not deterministically guaranteed in the actual encounter with a site visitor (Stanfill 1060). The way products and services are presented also depend on the place of the website where they can be found. This certainly applies to the fine print of these products. The question is what a company like Google gets in return and if it presents this information in a transparent normative way.
For the interviews the method of reception research is used. In the interviews the focus will be on similarities and patterns between the different answers. Together these patterns form a repertoire that can serve as a general finding out of the interviews. Schroder describes this method in his article reception research toolbox: the qualitative interview.
Reception research methodology is predicated upon the qualitative research interview, which is used as a discursive generator for obtaining an insight into the interpretative repertoires at the disposal of the informants as they make sense of a specific media product (143).
In the quote above Schroder talks about media products. For this research the topic of the interviews is not a specific media product, but the infrastructure of these media products. This slightly changes the perspective, but the same method of analysis can still be used. Before the interviews were taken, a topic list was conducted. The topics correspondent with the different aspects of digital infrastructures
