English summary (Nederlandse samenvatting). Download the full dissertation as PDF (43Mb).

Helmond, Anne (2015). “The Web as Platform: Data Flows in Social Media.” PhD dissertation, University of Amsterdam, 25 September 2015. Honorable mention, AoIR 2016 Best Dissertation Award.

The web as platform: Data flows in the social web

In October 2004, at the first Web 2.0 conference, Tim O’Reilly rhetorically repositioned the web after the dotcom crash as “the web as platform.” With this claim he suggested to understand the web not only as a medium for publishing information but also as a computational development platform for building applications. Not only the web as a whole, but also websites themselves are developed as platforms by offering Application Programming Interfaces, APIs. It is through APIs that websites and most notably social network sites can provide structured access to their data and functionality and be turned into platforms. In this dissertation I trace the transition of social network sites into social media platforms to examine how social media has altered the web.

The key aim is to develop the concept of “platformization” in order to understand this process from an infrastructural perspective. The platformization of the web refers to the rise of the platform as the dominant infrastructural and economic model of the social web and the consequences of the expansion of social media platforms into other spaces online. Platformization, I argue, rests on the dual logic of social media platforms’ expansion into the rest of the web and, simultaneously, their drive to make external web data platform ready. As an infrastructural model, social media platforms provide a technological framework for others to build on which, I argue, is geared towards connecting to and thriving on other websites and their data. Making external web data amenable for their own databases is, so I suggest, central to the economic model of social media platforms. These two processes of decentralizing platform features and recentralizing platform ready data characterize what I call the double logic of platformization. This double logic is operationalized through platform-native objects such as APIs, social buttons and shortened URLs, which connect the infrastructural model of the platform to its economic model. I argue that these platform-native objects serve as prime devices for social media platforms to expand into the web and to create data channels for collecting and formatting external web data to fit the underlying logic of the platform. That is, I show how social media platforms are building data-intensive infrastructures to reweave the web for social media.

This argument is organized around five case studies in which I chronologically trace the platformization of the web and its consequences in terms of 1) the transformation of social network sites into social media platforms 2) the restructuring of the blogosphere and the introduction of new linking practices, 3) the changing nature of the hyperlink from a navigational tool into an analytical tool for data capture, 4) the transformation of the currency of the web from link to like and 5) the boundaries of a website and the end of it as a bounded object. Adopting an approach that combines software studies, platform studies and digital methods, I analyze the underlying platform infrastructure and platform-native objects of the social web to ask what social media has done to the web. As part of this undertaking, I put forward new methods that I frame as digital methods for platform studies which utilize medium-specific features to explore dynamics of platformization.

In the first chapter on ‘The platformization of the web’ I provide a detailed material-technical perspective on the development and emergence of what we understand as social media platforms today. I trace how social network sites have become social media platforms by outlining three pre-conditions for platformization: the separation of content and presentation with XML, the modularization of content and features with widgets and interfacing with databases through APIs. Taken together, these aspects turn websites into programmable platforms allowing them to extend beyond their boundaries and establish two-way data flows for data exchanges with third parties. I conceptualize these data channels as ‘data pours’ that not only transfer data from database to database but also format external web data according to the logic of the platform.

In the second chapter on ‘The coming of the platforms’, I examine the changing structure of the blogosphere in relation to the rise of social media. This is achieved by reconstructing the historical Dutch blogosphere per year using a collection of archived blogs retrieved from the Internet Archive Wayback Machine between 1999 and 2009. Within this archived collection, I trace how social media platforms introduced new linking practices through widgets and show how these widgets transform the hyperlink structure of the blogosphere. To identify the role of social media in this historical Dutch blogosphere, I develop a method to further examine the types of links between blogs and social media platforms. It becomes apparent that with the rise of social media, bloggers are no longer predominantly linking to establish an interlinked network of blogs, a blogosphere, but increasingly connect to social media and, as a result, weave their blogs into the social media ecosystem.

In the third chapter on ‘The algorithmization of the hyperlink’, I retain my focus on the changing role of links and examine their advancing commodification by social media platforms. Here I show how platforms use their infrastructure to render the web-native object of the hyperlink into a platform-specific shortened URL. In doing so, social media platforms change the function of the link from a navigational into an analytical device amenable for data capture. Social buttons play a central role in this, as they create new forms of automated, data-rich shortened URLs that are formatted to fit the purpose of the platform; that is, to feed the underlying algorithms and analytics suites. In the web as platform, the link becomes a database call and a device to make external web data ‘platform ready.’ I develop a method to examine the actors involved in this reconfiguration of the hyperlink by following shortened URLs.

The fourth chapter on ‘The Like economy’ shows how social media platforms employ social buttons as part of their technical infrastructure to turn social activities into valuable data, conceptualized as a so-called ‘Like economy’. I contextualize the rise of social buttons as metrics for user engagement and link them to different web economies: the hit, link and Like economy. I explore how the platformization of the web shifts the currency of the web from web-native links to platform-native likes which are tied not to the web at large, but to the mechanics and logics of specific social media platforms. Facebook’s Like economy is enabled by the interconnected dynamics of the decentralization of data production—by offering social buttons to like content across the web—and the recentralization of data collection through these buttons. I devise a method to map the presence of social buttons on a collection of websites to show how they create new forms of connectivity between websites beyond hyperlinks, introducing an alternative fabric of the web.

In the final chapter on ‘Website ecologies’, I explore the changing boundaries of the website in the web as platform. Websites are increasingly shaped by and assembled from content and functionality such as embedded content, social plugins and advertisements, thereby complicating the notion of the website as a bounded object. The third-party objects present on websites draw attention to the larger techno-commercial configurations of the web that these sites are embedded in. I therefore suggest to reconceptualize the study of websites as website ecology which analyzes how various relations between the different actors on the web have become inscribed in a website’s source code. In this chapter, I propose a method that uses the source code of an archived website to study a website’s ecosystem over time as a way to examine the spread of platformization. In addition, I employ the affordances of social media platform APIs to retrieve missing platform content in archived websites.

The five case studies demonstrate that the consequences of social media platforms’ tight integration with the web—platformization—typify a significant change in how the web’s infrastructure is put to use. To study the platformization of the web, I therefore argue, one should engage with data exchange mechanisms, new means to connect websites, the transformation and commodification of the hyperlink, the introduction of new web currencies for web content such likes, shares and retweets, and the redrawn boundaries of the website. That is, one should recognize the platform-specific objects that have been introduced by social media platforms that take on various social and technical functions, one of them being to reweave the fabric of the web.

In this dissertation, I develop a platform critique that revolves around the notion of platformization that is positioned as a contribution to the emerging fields of software studies and platform studies, and draws on digital methods to study the effects of social media on the web’s infrastructure. In doing so, I answer current calls for taking platforms as computational infrastructures seriously, and respond to the need for new methodological development to advance the fields of software studies and platform studies. Ultimately, I propose a new branch of platform studies that I call platform infrastructure studies, which analyzes the ecosystem of software platforms with platform-specific digital methods.

In the conclusion, moreover, I ask whether this sort of critique still applies with social media access shifting from the web to mobile apps. In presenting a future research agenda, I address a number of developments that show how social media platforms are integrating themselves into apps (and vice versa), pointing towards a platformization of the app space. Most prominently, apps are supposedly killing the web as well as the mobile web with far more time being increasingly spent using apps than on the mobile web. More specifically as a case in point, with the launch of the Facebook Messenger Platform for building apps that integrate with Facebook’s Messenger app, we can observe the reconceptualization of the app as a platform.

These developments show the urgency of continually developing new methods to study the platformization of the web as well as the platformization of the app ecosystem, together with the consequences. Whilst the rise of apps may introduce a different dynamic of platformization, there are likely shared concerns. In all, the concept of platformization and the methods developed for what I have called platform infrastructure studies provide ways to examine not only a platform’s ecosystem on the web but also changes to the web’s infrastructure. They also provide means to examine the app space, including the platformization of apps as well as an app’s ecosystem.