The Internet, Policy & Politics Conferences

Oxford Internet Institute, University of Oxford

Anne Helmond, Fernando N. van der Vlist: Big Data Advertising Infrastructures: A Comparative Study of Social Media Ad Platforms

Anne Helmond, University of Amsterdam

Fernando N. van der Vlist

In 2015, social media companies generated over $25 billion in advertising revenue (eMarketer, 2015). For Facebook, advertising revenue accounted for 95% of its total revenues, for Twitter 90%, and for LinkedIn, 20% of their revenue came from marketing solutions. The largest player by far is Facebook with over 2.5 million active advertisers, generating $5,637 million in revenue in the fourth quarter of 2015, 80% of which now comes from mobile advertising. Social media platforms do not only facilitate connections between people but also constitute marketplaces, mediating between the platform itself, its end-users, marketing partners, and advertisers. Our aim is to investigate social media as advertising platforms and move beyond studies that merely focus on ‘connectedness’and which do not take the ‘connective’aspects of platformed sociality into account (Van Dijck, 2013). Building on previous work by Rieder and Sire (2014) who have examined Google as an advertising platform instead of as a search engine only, we similarly propose considering social media platforms from a techno-economic perspective to foreground platform economics and conceptualise these platforms as multi-sided markets (cf. Rochet and Tirole, 2003; Evans et al., 2006). In doing so, we are contributing a political economic perspective to the field of platform studies (Bogost and Montfort, 2009) of social media, which mainly focuses on how platforms enable and constrain platform users and activities.

The platform, as the dominant infrastructural and economic model of the social web, extends beyond its boundaries by decentralising and recentralising platform content and functionality and imposes its logic on external actors through the process of ‘platformisation’(Author, 2015). Platformisation in this context means that ad networks are not only enabling advertisers to create and serve ads within a specific platform, but also on external websites, platforms, apps, and across devices. Studying this expansion of social media platforms through advertising therefore requires a cross-platform perspective that also takes the larger platform ecosystem into account. As such, we contribute to existing work investigating the politics of platforms (Gillespie, 2010), addressing the ways in which platforms negotiate relations between different stakeholders such as institutions, corporate actors, end-users, and third parties. Furthermore, we situate the transformation of social media companies into advertising platforms within the larger history of the development of the online advertising industry (cf. Turow, 2011). Since the mid-1990s, this industry has changed significantly, not only in terms of the composition of the major players, but especially in terms of changing techniques for audience aggregation (i.e. how advertising audiences come into being) and differentiation (i.e. how internet companies differentiate themselves from their competitors). The design of software platforms, including social media platforms, and their business models have consequences for how the advertising industry evolves (cf. Evans, 2006: 39).

Our investigation focuses primarily on Facebook, Twitter, and LinkedIn, the top three social media companies in terms of relative share of advertising revenues in the fourth quarter of 2015 (eMarketer, 2015). By following the medium (Rogers, 2013: 24) and the traces of actors (Latour, 2005: 12), including financial traces, we are able to map the relations between different stakeholders and their interests and study how platforms negotiate these conflicting interests. We trace these associations by making use of the platforms’own advertising interfaces and available online platform documentation. This includes developer documentation, business, marketing, and advertising pages as well as partner directories. Our methodology connects three levels of analysis: first, we identify relations between each platform and its commercial and business partners, and then identify and map cross-platform relations in order to gain insights into the larger social media advertising ecosystem. As such we contribute a novel empirical methodology for doing comparative studies of platforms as well as cross-platform analysis. Second, we characterise individual platforms by investigating their relationships to advertisers and end-users as structured through Application Programming Interfaces (APIs), Software Development Kits (SDKs), and myriad business apps built for advertisers and marketers. Third, we look into the medium-specific techniques and interfaces of advertising tools created by the platforms themselves that enable such things as the creation and management of ads, ad delivery, analytics, and reporting.

Our findings so far concern all three levels of analysis: first, social media platforms facilitate distinct ways of producing audiences and delivering ads. The systematic capture of platform-native engagement metrics based on user activities (e.g. Likes, shares, and tweets) facilitates the real-time production of dynamic or ‘fluid’audiences (e.g. users liking a certain Facebook Page). These are audiences that do not need to be stabilised in composite profiles because the data on which they are based are the very same data that are used for targeting advertisements. Moreover, platforms offer their own techniques such as the Facebook Pixel used for tracking conversion rates, app installs, and other ‘App Events’. This technique has enabled a shift from ‘proxy metrics’(e.g. clicks and ‘impressions’) to business results (e.g. in-store sales). Similarly, Twitter’s platform distinguishes between clicks on hyperlinks, media objects, and app install links, enabling different accounts of ‘Tweet Engagement’. While metrics are native to each platform, some of the partners aggregate them in order to offer software services across devices and platforms. Second, Facebook has an extensive and highly diverse range of techniques and applications compared to Twitter and LinkedIn. It offers countless APIs geared towards advertisers, such as the Marketing API, Atlas API, Business Manager API, Ads Management API, Ads Insights API, and Native Ad API, to only name a few. Distinct forms of advertising each come with their own specific APIs, facilitating and structuring different advertising products and practices. In addition to characterising platform advertising infrastructures, we are able to specify and map partner relations and their marketing specialties, advertising goals, pricing models, target industries, countries, and languages. Third, social media advertising infrastructures do not only consist of APIs, SDKs, and apps, but also include ad networks which are accessed through several of these interfaces. Ad networks play a central role in bringing together and coordinating the many different different stakeholders and their interests. Finally, social media platforms share a great number of marketing partners with each other, which raises questions relating to media concentration in big data advertising. This means we should consider investigating these partners and their role in the social media advertising industry. For example, we should enquire into the capacities of major social media platforms such as Facebook to link their own data to databases from external partners such as Acxiom, Datalogix, and Epsilon, which enable Facebook to integrate advertising data with external consumer data sets.


Bogost, Ian, and Nick Montfort. 2009. ‘Platform Studies: Frequently Questioned Answers.’In Proceedings of the Digital Arts and Culture Conference. University of California, Irvine. eMarketer. 2015. ‘Social Network Ad Revenues Accelerate Worldwide.’eMarketer. September 23.

Evans, David Sparks, Andrei Hagiu, and Richard Schmalensee. 2006. Invisible Engines: How Software Platforms Drive Innovation and Transform Industries. Cambridge, MA: MIT Press.

Gillespie, Tarleton. 2010. ‘The Politics of ‘platforms.’’New Media & Society 12 (3): 347’64.

Latour, Bruno. 2005. Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford: Oxford University Press. Rieder, Bernhard, and Guillaume Sire. 2014. ‘Conflicts of Interest and Incentives to Bias: A Microeconomic Critique of Google’s Tangled Position on the Web.’ New Media & Society 16 (2): 195’211.

Rochet, Jean-Charles, and Jean Tirole. 2003. ‘Platform Competition in Two-Sided Markets.’Journal of the European Economic Association 1 (4): 990’1029.

Rogers, Richard. 2013. Digital Methods. Cambridge, MA: The MIT Press. Turow, Joseph. 2013. The Daily You: How the New Advertising Industry Is Defining Your Identity and Your Worth. New Haven: Yale University Press.

Van Dijck, Josщ. 2013. The Culture of Connectivity: A Critical History of Social Media. New York: Oxford University Press.

Anne Helmond, Fernando N. van der Vlist