In this week’s CPGJ platform capitalism reading group, we turn towards education for the first time with a paper by José van Dijck and Thomas Poell looking at the influence of social media platforms on education, particularly within schools. Much of the literature has addressed social media as tools, with varying interpretations offered about how these might harm or hinder teaching and learning. The ubiquity of social media is often cited as a reason to try and integrate their use into the curriculum, with some arguing they could play a crucial role in helping with particular tasks such as information retrieval. Others frame social media as a disruptive force within the classroom, undermining existing routines and creating problems for teachers. Optimists and pessimists are united in their “social media-as-tools approach: social media are considered as technical tools that may either enhance or disrupt learning experiences”. In contrast, van Dijck and Poell insist on framing these as platforms, which are “driven by a complex interplay between technical architectures, business models, and mass user activity” and “introduce new mechanisms in social life”.

This helps broaden the focus of our analysis, away from “student behaviour and teaching practices” towards “the organization of schools and universities and, one might argue, (public) education as such”. Their analysis rests upon two distinct mechanisms: datafication and commodification. In doing so, they draw on work which has explored social mediain terms of a transformation of the landscape within which young people become civic actors, creating a range of possibilities for how education might change. The development of this perspective by van Dijck and Poell involves seeing social media as “more than mere technical facilitators: they are simultaneously technological, economic, and socio-cultural frameworks for managing online social traffic”. The main focus of their paper is upon how ratification and commodification reshape the organisation of education at primary and secondary levels.

• Datafication is “the tendency to quantify all aspects of social interaction and turn them into code”. This incorporates two aspects: quantification and digitisation. The affordances of digital technology facilitate quantification to an extent that would not otherwise be possible. This can have descriptive and predicative dimensions to it: tracking developments in real time but also producing predictions which feed back into practice. In a sector like education, “emerging digital policy instruments transfer the assessment of didactic and pedagogical values from teachers and classrooms to (commercial) online platforms deploying real-time and predictive analytics techniques”. But datafication will have a similar tendency in others sectors because it circumvents the situational judgement of professionals by creating an analytic apparatus which operates in the background. There might be a degree of variability in how much leeway the professional continues to enjoy (consider for instance the way data can be used to enhance the performance of elites) but the broader trend is towards the diminution of agential prerogative.In the educational context, mechanisms of datafication includes data trackers and dashboards, facilitating personalisation of a sort similar to that found in content-streaming platforms like Netflix. As they write of AltSchool, it “favors technology over teachers; online personalized learning takes over classroom instruction; and the primacy of predictive analytics downgrades teachers’ professional judgment”. Digitalising a process, rendering it data and quantitative, imposes epistemic constraints on the ensuing knowledge, creating a bias towards the immediate and the atomistic. The specificity of educational is eviscerated by a generic architecture of likes and upvotes.
• Commodification involves the “monetization of online social traffic through business models and governance structures” and is closely connected to datification. A limited number of business models all revolve around how data can be used to generate profit, incentivising continual expansion of datafication and economies of scale giving rise to fewer and larger data actors. It is hoped that was is datafied can be commodified.Data-driven commodification facilitates the unbundling of education. As the authors write, “[t]he conventional business model reflects the ideology of higher education as a curriculum-based, comprehensive experience that offers an education at a price that includes not only lectures or course content but certification, advising, tutoring, and testing”. The market for educational data, coupled with the near-zero marginal costs of digital communications, means that the curriculum can (technically) be delivered purely as content and there is a (financial) motivation for doing so. The potential implications of this educational data have barely been recognised, with the authors plausibly suggesting they might in future replace CVs in the eyes of employers.

Their analysis refuses to separate off education platforms from the wider ecosystem in which they emerge, dominated as it is by the major actors of Facebook, Google, Apple, Amazon and Microsoft. While education platforms might not threat existing institutions in the manner of Uber and taxi firms, Dijck and Poell identify three features which might lead to systemic change:

1. Principles of social media architecture have primacy over pedagogical principles on educational platforms. When young people are “growing up immersed in the compelling social interaction these platforms offer in terms of connecting, liking, rating and following each other” and free education services (e.g. Google Scholar, Google Docs, Gmail) offered by major players like Google already play a prominent role in young people’s educational lives. This ubiquity is liable to be reinforced by continued growth in use amongst young people and funding shortfalls leaving organisation’s looking to free services which enables costs to be cut. The result is that “corporate platforms such as Google, Facebook, LinkedIn, and Microsoft are able to position themselves strategically, at the gateways of educational infrastructures such as libraries, administrative and communication systems”.
2. The capacity of education platforms to solve what are perceived as endemic problems of education is widely assumed yet little demonstrated. This reflects the broader influence of ‘solutionism’ (seeing technical fixes to social problems) and the narrative of sustained failures within the education system. These platforms are presented as emerging from off-stage to solve the problem, riding to the rescue of something their own emergence is intimately bound up in the creation of as part of the political economy of late capitalism.
3. The growth of online educational globally might lead to a cultural shift in the understanding of education as a public good. They suggest we can identify “how education is increasingly defined as a technological challenge developed by tech companies and decreasingly as a service carried out by dedicated teachers and funded by taxes”. The scaleable and free logic of digital education seems enticing against a backdrop of austerity politics and a drive towards the retrenchment of the welfare state.

The second paper analysis the platform as evaluative infrastructure. They are evaluative in the sense of deploying a wide array of ranking mechanisms to establish orders of worth. They are infrastructure because they provide the background conditions which makes interaction possible. An infrastructure consists of “technical artefacts, institutional arrangements, cultural habits and social conventions” (“people, language, numbers, categories, cultures, practices, artefacts but also pipes and hard-wired circuits”) to produce material forms which facilitate exchange over time and space. Power within them operates through protocols (rules and standards governing behaviour within networks) rather than familiar hierarchical forms of influence. Evaluative infrastructure “consists of an ecology of devices that disclose values of actions, events and objects in heterarchically organized systems (such as platforms) through the maintenance of protocol”. Their mechanisms co-ordinate and condition interaction which takes place between distributed parties, with the platform being the means through the platform owner facilitates the interaction and seeks to profit from it. Evaluative infrastructures facilitate platform owners to operate distinctive types of platform organisation. The evaluative infrastructure is what makes platform capitalism possible.

An immense amount of activity takes place on them: “as of 2014 eBay had 165 million active users,3 Uber was hosting over 1 million rides per day, and Airbnb was facilitating 155 million guest stays annually, surpassing the Hilton Worldwide by 22 percent”. The evaluative infrastructure establishes shared orders of worth which makes this interaction meaningful, stabilising expectations and generating trust between parties who do not stand in a prior relation to each other or have much context in common. In doing so, they “relate and recombine people, ideas, and things” through “the invisible infrastructures that coordinate and control platform activities”. Their operation rests on a “an ecology of accounting devices in the form of rankings, lists, classifications, stars and other symbols (‘likes, ‘links’, tags, and other traces left through clicks) which relate buyers, sellers, and objects”. The value creation this gives rise to takes place horizontally across the platform, defying any traditional vertical attempts to organise it by the platform-owner, necessitating a new accounting regime on the part of the platform owners and new concepts for social scientists to analyse their operation. Part of the challenge stems from the capacity of these infrastructures to bring new worlds into being rather than capturing the traces of what is already there.

Community plays a significant role in this, with the eBay founder once saying that “eBay’s success as a company de- pends upon the success of the community”. What I take them to be saying, in slightly different theoretical lingo to the one I’d used, concerns the capacity of platforms to generate relationality within groups. It produces thick relations through the mechanisms designed to counter the fact thin relations are the starting point. In doing so, the interests of the platform are effectively baked into the relational web, as much as it remains possible for its evaluative orientation to run counter to the problem in exceptional cases. Users can resist a platform but they do so in spite of their status as users. Recognising this will be crucial to understanding the lived experience of platform participation, generating thick descriptions of actions within and through infrastructures which “constantly link events, actions, behaviours, decisions (clicks), assessments and other traces left unintentionally and unconsciously (such as speed of typing, time of access, or browser used to access site) all of which are used to build a web of context around objects and subjects”. The power of platform owners operates under these conditions “through its infrastructural design, maintaining standards, imposing what counts and how to count, excluding users, and introducing rules” so as to structure the field of possibilities, rather than guiding actors within it.

Questions for discussion:

1. What is at stake in whether we define social media as platforms or tools?
2. What does it mean to say “All platforms are equally defined by a set of mechanisms”?
3. Where are the agents behind evaluative infrastructures?