Book Review: The Platform Society. Public Values in a Connective World (2018)

Van Dijck, José, Poell, Thomas and Martijn de Waal (2018). The Platform Society. Public Values in a Connective World. Oxford: Oxford University Press, 226 pages.

By Matthias Finger

This book is already five years old and probably older if one considers the time of writing, an eternity when it comes to rapidly evolving digitalization. Yet, its analysis of how platforms work and what they do to society is still among the best available today. Written by Dutch academics specialized in media studies, the book by Van Dijck, Poell and de Waal brings yet another unique perspective to digitalization. Grounded mainly in their analysis on how digitalization has already transformed the media sector, the authors reflect more generally on the effects of digitalization on society, social life, and social interactions. Hence, I will focus my review on their excellent analytical chapters and offer a critical review of the remedies they propose in conclusion.

To begin, the broad focus makes this book particularly interesting and relevant: “… platforms are at the core of an important development, but we think of them neither as an exclusive economic phenomenon nor as a technological construct with social corollaries. Rather, we prefer a comprehensive view of a connective world where platforms have penetrated the heart of societies affecting institutions, economic transactions and social and cultural practices … platforms produce the social structures we live in.” (page 2).

The authors’ main argument is that digital platforms, in general, and the main infrastructural platforms, in particular (i.e., the “Big Five” that are Amazon, Apple, Facebook, Google and Microsoft), have become so powerful and so pervasive that they have by now permeated every corner of society and social life to the point that they determine, at least in part, the way society operates. This raises the question about whether this evolution is ultimately in the public interest. It is indeed this broad focus, on the societal transformation as a result of digitalization, which attracted my interest for this book. This is a rather rare approach among scholars who write on digitalization and digital platforms.

Infrastructural and Sectoral Platforms

In their first two chapters, the authors argue for the need of such a comprehensive approach. They do this by introducing the very useful distinction between sectoral platforms, on the one hand, and foundational infrastructural platforms as organizers of a “platform ecosystem”, on the other. Infrastructural platforms can, and do, also operate sectoral platforms. Still, this distinction is extremely valuable, as it has profound regulatory implications. According to the authors, “an online ‘platform’ is a programmable digital architecture designed to organize interactions between users – not just end users but also corporate entities and public bodies. It is geared towards the systematic collection, algorithmic processing, circulation and monetization of user data”. … “A ‘platform ecosystem’ is an assemblage of networked platforms, governed by a particular set of mechanisms that shape everyday practices. The Western ecosystem is mostly operated by a handful of tech companies (Alphabet-Google, Apple, Facebook, Amazon, and Microsoft) whose infrastructural services are central to the ecosystem’s overall design and distribution of data flows” (page 4).

Sectoral platforms are typically sector- or domain- specific and use a range of business models but basically “create value out of data, content, user contacts, and attention by selling advertisements, subscriptions, and user data or by charging fees; moreover, they can sell data to other companies or governments …” (page 10). In doing so, “they steer user interaction but simultaneously shape social norms” (page 11). In order to be able to operate, these sectoral platforms build on the infrastructural platforms operated by the Big Five: “these (infrastructural platforms) form the heart of the ecosystem upon which many other platforms and apps can be built. They also serve as gatekeepers through which data flows are managed, processed, stored, and channeled. Infrastructural services include search engines and browsers, data servers and cloud computing, email and instant messaging, social networking, advertising networks, app stores, pay systems, identification services, data analytics, video hosting, geospatial and navigation services and a growing number of other services” (page 13). On the one hand, “the Big Five profit most from the burgeoning development of sectoral platforms and millions of websites and apps integrated with their basic services, enabling the collection of user data throughout the Web and app ecosystem” (page 15). On the other hand, the sectoral platforms that are “not connected to the ecosystem’s core can hardly profit from its inherent features: global connectivity, ubiquitous accessibility, and network effects” (page 15).

This consideration also applies to governments, public institutions, and non-governmental organizations that “can, of course, operate their (own) platforms; but it is increasingly difficult to do so as autonomous actors” (page 15). In other words, the sectoral platforms, even those operated by governments, become merely “complementors” of the Big Five’s respective infrastructural ecosystems. They do so because they are connecting all these complementary platforms and“are turning into infrastructures that are inherently essential” (page 16) to the point that “they have become paramount to the functioning of economies, as well as democracies” (page 20).

Futhermore, at any moment in time, infrastructural platforms can blur the boundaries by venturing into sectoral platform activities: “The potential for vertical integration between infrastructural and sectoral platforms is endless … Some platforms’ near monopoly status in the infrastructural core coupled onto the sectoral platforms’ dominant position make these companies become ‘fluid’: they introduce a new type of organization, defying the classical definitions that are tied to sectors. … More importantly, accumulation of power typically happens between sectors as data streams can be manipulated across sectors via infrastructural platforms that are sector-agnostic” (page 19).

Platform Mechanisms

In their second chapter, the authors delve into the inner workings of the platforms – something they call “platform mechanisms” – by describing the processes of “datafication”, “commodification”, and “selection”. Although these processes are already well known, the authors show how they reenforce each other and how they shape social values. “Datafication” is about “rendering into data many aspects of the world that have never been quantified before” … “every form of user interaction can be captured as data” … “datafication endows platforms with the potential to develop techniques for predictive and real-time analytics …” (page 33).

Capturing data is not a neutral endeavor: “data are always already prefigured through a platform’s gathering mechanism. Platforms do not merely ‘measure’ certain sentiments, thoughts, and performances, but also trigger and mold them …” (page 34). In other words, “data are not raw but precooked” (page 35); raw data is an oxymoron” (page 34). In short, through datafication, values are not only codified, they are also “reified”, a concept that the authors unfortunately do not use, yet that captures their main idea.[1] In other words, thinking and behaving is standardized by the platforms, according to the values that are embedded in their codes.

“Commodification” furthermore transforms these “online and offline objects, activities, emotions, and ideas into tradeable commodities. These commodities are valued through at least four different types of currency: attention, data, users, and money” (page 37). Due to the network effects they can generate, platforms not only shape behavior, they also “aggregate, facilitate, and control the connections and transactions between different groups of users” (page 38). Of course, the infrastructural platforms that control the ecosystem can do this more effectively and on a larger scale than the sectoral platforms, hence they can extract more power and money than the sectoral platforms. This is also because the infrastructural platforms operate globally, i.e., above and beyond national jurisdictions. In other words, commodification, as performed by platforms, is not only global but also almost without limits and, as such, is beyond the reach of regulators.

Note that platforms not only datafy and commodify, they also “select” and “curate” and, by doing so, further standardize and shape social values, social interaction, and social behavior: “Datafication and commodification are closely related with the ways in which platforms steer user interaction through the selection or curation of most relevant topics, terms, actors, objects, offers, services, etc.” … They “replace expert-based selection with user-driven and algorithm-driven selection” (page 40; see also next section). Keep in mind that selection is not driven by the users themselves, but again precooked, i.e., value-laden and standardized by interface features and algorithms that “are anything but transparent to users” (page 41). Moreover, these algorithms “are constantly modified in response to evolving business models and user practices” (page 41); yet, these model and practices are   almost impossible to follow hence to regulate.

Together, these three mechanisms make up for the complex and multi-dimensional dynamics of platforms; specifically, the dynamics between the individual treatment of every user (“personalization”) and the standardization of collective behavior, as well as the dynamics between the past-oriented aggregation of historical behavior and the shaping the future behavior. The authors also show that there is not much science at play here, instead there is much experimentation and trial and error. There are also many implicit value-laden judgments, often in reaction to societal fashions and pressures. All of this is governed by the overarching objective to augment network effects and, by doing so, to increase power and revenue.

Of course, such platform dynamics and mechanics take place outside of and mostly beyond the control of institutional actors and institutions more generally. Even worse, institutional actors, including governments, most public institutions, and non-governmental organizations have become part and parcel of the ecosystems as controlled and operated by the infrastructural platforms, for which they are now also merely “complementors”. Even their (public) values, their (public) services and their activities (in the public interest) are now datafied, commodified, and curated.

Undermining the Public Interest

Although the first two chapters, are more conceptual in nature, the subsequent four chapters are about the sectoral illustrations, i.e., about how sectoral platforms affect social values, social interactions and social behavior in the news,  and in the urban transport, healthcare and education sectors. Of these four chapters, the chapter on the media sector is clearly the best one, as this is the domain of expertise of the authors. The chapters on health care, transport, and education are rather anecdotal and grounded in secondhand knowledge. Not only is the media sector the subject the authors know best, the media and news is also the domain where digitalization most directly and most profoundly affects and even shapes society. For the authors, the key question is about how digitalization, as described above, “reshapes public values in the news (and media) sector – values that have historically guided the journalistic profession and are deemed of vital importance to journalism’s role in democratic politics” (page 50).

The authors operate with two concepts that illustrate the effects of sectoral platforms on the media sector, namely unbundling and personalization. Both concepts have not yet really been theorized in the context of digital platforms. Unbundling means that media content is disaggregated: “each individual story becomes a separate product standing naked in the marketplace which lives or dies on its own economic merits” (page 52). Instead of following an editorial policy that organizes the content and aims at conveying a message and a pedagogical intent, platforms disaggregate the content into isolated articles, stories, or pieces and feed them separately to the reader. At best, content is reaggregated but, most of the time, it just becomes a shopping list from which the reader can then choose. This is, of course, a form of “elimination of the middleman”, i.e., the editor and curator of the newspaper, a process already well-known from the platformization of the retail and the taxi industries.

Furthermore, and due to ever more sophisticated algorithms, journalists (the producers of the content) can now follow the take-up of their articles in real-time by the readers and consequently feed them even more of what they want, thus reenforcing the echo-chamber phenomenon. To some extent the editorial intent, the pedagogical line, etc. get lost. Information and media are now a matter between the writer and the reader, with the platforms’ algorithms – datafication, commercialization and selection (curation) – acting as the engine that fuels this process at an ever-accelerated pace.

Simultaneously, and to a certain extent paradoxically, newsfeeds are also becoming ever more personalized, thus “putting additional pressure on journalists to produce content that triggers user engagement” (page 54). Although user engagement is something personal, it is also highly standardized and only of commercial interest if it takes place at a massive level. Moreover, “the integration of platform data in news operations effectively creates path dependencies as the data infrastructures of the Big Five platforms shape the scope of editorial decision-making” (page 54). “Instead of relying on editorial decisions based on journalistic judgments, the ‘automated’ news process would be determined by quantified user demand” (page 55).

Instead of making a contribution to the shaping of opinion and to a democratic debate, “the Big Five platform corporations, which operate internationally, tend to set global standards regarding content that can be shared by professional news organizations. Given that most infrastructural platform corporations are US-based, this effectively entails a globalization of American cultural standards concerning what is and is not permitted” (page 63). In other words, the role of media in shaping public opinion and in creating a healthy public debate, is structurally undermined by digitalization (the digital architecture). It is not the content that is the real problem, rather it is that the content is disaggregated and personalized. This occurs both in the infrastructural and the sectoral platforms.

In their chapter on education, the authors identify exactly the same mechanisms at play: The educational content and the curriculum are disaggregated and personalized to the individual learners who can choose content as they like. “Bildung” (formation), as a societal project, is replaced by individualized learning activities, e.g., the consumption of YouTube videos. A collective objective is being replaced by a purely individualistic one that plays to ambient individualism, while furthering it. The next step in this process, currently in the making, is to re-bundle such individual learning activities in the form of degrees; degrees given by the platform (such as Coursera) and no longer by public authorities.

In both media and education, one can see how the collective objectives of fostering socially responsible behavior with the public interest are being replaced by a neoliberal individualistic approach whereby every individual is now a consumer of commodified media and educational content. Undermining social life and collective values is one result, but the other is the fact that all this individualized media and educational content is again totally standardized and heavily value-laden, thus only giving the illusion of a personalized treatment. Consequently, everyone is individualized (as the result of the structural nature of digital platforms) yet, at the same time, totally shaped by the standardized values embedded in the algorithms. A true “One-Dimensional Man” to refer to a philosopher, Herbert Marcuse (1898 – 1979), the authors do not seem to know. Whereas the authors do not go as far as making this statement, they nevertheless perfectly identify and illustrate the mechanisms at play in the digital media and educational platforms that lead to it.

Weak on Remedies

So, what are the remedies to this evolution, as outlined in the last chapter of the book? To be clear, the book is weak on remedies, which is very disappointing after such a powerful analysis. This weakness is, in my view, mainly due to the fact that the authors have no intellectually solid understanding of the public interest, a term that they use randomly. First, their vocabulary is fuzzy, whereby they confuse “public interest”, “collective interest”, “common interest”, “public services”, “public values”, and others. Although they are good at showing the way values are profoundly embedded inside the platforms’ architectures, including in the ecosystems’ architecture of the Big Five, they can only lament the fact that the value-laden nature of platform architectures and algorithms is not openly recognized, or more precisely, is obfuscated behind technical discussions. They also lament that these embedded values favor individualistic and narcissistic behavior that ultimately undermines social life, as well as the ability of society to collectively determine its future. One could have expected more.

Regarding remedies, the reader is struck by the authors’ naiveté: Although they have identified profoundly structural issues that are at the heart of digitalization (namely, datafication, commercialization and selection), their answer is basically cosmetic and contradictory, when they say that  public values (which ones? whose values?) have to be embedded into the very infrastructural architecture of the Big Five, as well as into the algorithms used by the sectoral platforms. In other words, they propose to remedy a structural problem with an ethical answer; a naïve approach that sadly has a long tradition. At the minimum, one would have expected a more political remedy, given that the public interest is a political and not an ethical concept.

However, this would have required a much more elaborate analysis of governments and their abilities to act in light of pervasive global platforms that, as they authors show themselves, have come to underly government activities and services, be it at the local the national and the supranational levels. In summary, all the authors propose is a stakeholder analysis, which is very typical in business school literature: They call for governments (which ones?), companies (it is not clear which ones, but the reader understands that it is probably the Big Five) and civil society (represented by whom?) to come together and jointly decide about the public values that should “guide” the platforms’ underlying architectural design. The fact that the emphasis is on “guide”, rather than on “regulate”, makes this a disappointing remedy and conclusion … but then this might be the only realistic remedy, considering that governments are already platformed, if not directly in bed with the Big Five.

[1] „Reification“: the act of treating something abstract, such as an idea, relation, system, quality, etc. as if it were a concrete object (www.dictionary.com).

This edition of the Digital Governance Book Review was authored by: Matthias Finger, C4DT

Image credit: Cover of The Platform Society. Public Values in a Connective World by Van Dijck, J., Poell, T. and M. de Waal, published by OUP.

David Viollier