A fit-for-purpose platform research agenda for a broken world

The term ‘platform’ has multiple meanings and roots and there has been ample debate around it (see for instance Gawer, 2014; Langley and Leyshon, 2017; Montalban et al., 2019). There has also been an attempt to universalize the idea of platforms, which sees contemporary capitalism as going through a platformization process (Poell et al., 2019). The next step is to work on two fronts. Research should focus on conceptually integrating platforms into broader capitalist dynamics without attempting to see it all as a platform. At the same time, more scholarship on the heterogeneities among platforms is needed. By opening up the concept of platforms while connecting it to other emerging features of our time we could build a much more thorough and critical analysis of the present.

A starting point for developing such a research agenda could be to recognize that if there is one thing that all the critical literature on platforms agrees on is that, either seen as a company or a distinct type of architecture, for-profit platforms have neither democratized nor horizontalized society. Platforms have cemented existing power relations and generated new ones. Precisely speaking of power relations and platforms, the ongoing brain drain by Big Tech companies highlights the need for more independent and critical research, especially considering that some academics who are investigating platforms from the humanities and social sciences – in particular in relation to artificial intelligence (AI) – work for or are directly funded by leading tech companies. This signals that their outcomes will likely be complicit with industry interests (Abdalla and Abdalla, 2021; Ochigame, 2019).

We also need more critical research because, across existing conceptualizations, there is a risk of completely detaching platform studies from capitalism's general dynamics or replacing the latter with the former. This is particularly relevant for a journal focused on platforms. Instead of seeing this as a threat, I think of it as a trigger for new investigations populating this journal. For instance, we need more research delving into what is common between platforms and other architectures of control beyond ownership, such as global value chains and franchising. How do platforms intersect with them? This type of question inscribes platforms within global capitalism.

Such a more comprehensive inquiry into the place of platforms in global capitalism could be tackled from multiple angles. One of them refers to the transformations of corporate planning amid digital technologies, noting that leading companies from the most diverse sectors are combining the development of platforms with their pre-existing architectures. McDonald's gathers data from its customers and restaurants using digital technologies in-store and its own applications that it uses to plan its over 40,000 restaurants, of which it owns only 5%. The rest are franchises. ¹

More generally and in relation to corporate planning, platforms are both a means of information and knowledge appropriation and a means of spatiotemporal projection to exercise control beyond ownership (Bensussan et al., 2023). Data are generated, curated and gathered according to the parameters and standards set by platform operators. They rule not only what can and cannot be done on their platforms, but also define the type of data that they want users to produce. There is a need to advance our knowledge of ruling and platforms, how this interplay shapes and is shaped by economic power relations, and what are their social effects.

Another angle of analysis refers to core–periphery relations. Research on platforms has mostly focused on the United States and China, which have become the world's core digital economies (Rikap and Lundvall, 2020; UNCTAD, 2019, 2021). There is a double opportunity here. On the one hand, research could explore different core–periphery scenarios in relation to platforms. For instance, the place of platform companies from Brazil and Argentina expanding towards most of Latin America could be seen through the lenses of a regional core and its periphery. For instance, a recent comment along these lines builds on Ruy Mauro Marini's concept of sub-imperialism to study the role of Brazilian platforms in Latin America (Seto, 2024). On the other hand, and of crucial importance for updating research on the contemporary causes of underdevelopment, there is a lacuna in terms of simultaneously studying the particularities of platforms in the peripheries while looking at the relationship between regional platforms and global leaders. So far, while some authors have considered the former (see for instance Andreoni and Roberts, 2022), others have focused on the latter, such as those advancing the idea of digital or data colonialism (Couldry and Mejias, 2019; Kwet, 2019). The interplay of these dynamics remains underexplored.

In a recent study on Mercado Libre, Latin America's e-commerce and fintech behemoth, we made a first attempt to consider both dynamics together (Fernández Franco et al., in press). We found that Mercado Libre structurally depends on digital technologies – algorithms and processing power – provided as black box cloud services by Amazon and Google. It intertwines those purchased services with in-house frontier digital technologies to control and capture value from those partaking in its platform business. This is not simply digital colonialism, because the regional platform accumulates at the expense of other local actors – companies and users – while favouring global players. Dependency Theory authors in Latin America have precisely acknowledged the role of complicit local actors in the extraction of value from the peripheries. And this has been observed both by the less and more radical streams of thought within this approach in the region, as shown by Bresser-Pereira (2006) in his account of the different streams within the Dependency Theory. These results are just one step forward into an uncharted – and yet crucial – territory.

The relationship between Mercado Libre and US Big Tech companies is not exclusive to regional platforms. From Netflix to Meta, global platforms from core countries also depend on digital technologies offered on Amazon, Microsoft and Google Clouds. These three giants, in that order, concentrate over 66% of the cloud computing market globally. ² On the public cloud, these giants sell hardware (infrastructure), software, platforms and datasets as services through the Internet and running in their data centres.

The cloud is itself a platform where third-party companies offer their software as a service. Just like on Amazon's marketplace, a share of every rented service is charged by the cloud provider. Third-party companies follow their rules. The co-production of technology is distributed between these Big Tech companies and countless third-party players, including other large tech companies such as Salesforce, IBM, Meta and Oracle. They all offer their computing services on Amazon, Microsoft and Google Clouds, yet profits are disproportionately captured by the latter.

The cloud is all the more important since the AI fever. AI models are trained and processed on Big Tech clouds. Complementary applications are developed and work by running the models hosted on those clouds. Organizations also use AI models on the cloud. ChatGPT runs exclusively on Microsoft's cloud Azure. Whereas Meta's large language model, Llama, is available on the three clouds. Many research avenues spring from these entanglements. Power relations are being exercised not only within but also among tech and platform companies. Moreover, the interconnectedness of the system points to network relations that transgress the tech sector proper, as companies from around the world migrate to the cloud to develop their ‘AI solutions’.

A related puzzling topic is platforms’ heterogeneities. While there is some research on why some platforms succeed or fail (see for instance Montalban et al., 2019), this remains an understudied topic. Heterogeneities go beyond classifications of winners versus losers. It is precisely the fact that some successful platforms, such as Mercado Libre or Netflix, structurally depend on the cloud hegemons, which calls for more complex studies of the heterogeneity and interplay among platforms.

Finally, as some platforms become fundamental for every organization and individual around the world, as is prominently the case for the cloud but also other platforms such as search engines, there is space to elaborate further on how platforms should be governed. While the literature is mostly inclined towards promoting competition, implicitly accepting (and promoting) for-profit models, there are serious doubts as to whether federated solutions that encourage a less asymmetric digital capitalism can overcome network effects and the quasi-infinite content available on platforms such as Google YouTube, Facebook, or TikTok (O'neil et al., 2021).

In fact, many platforms have a tendency towards natural monopolies that is only partly explained by strong network effects that lead to situations in which the winner takes all or most (Ducci, 2020; Rikap and Durand, 2022). Search engines such as those of Amazon's marketplace and Google's multiple platforms are AI models that get better the more they are used. Every search contributes to improving the algorithms and thus the provided service. This creates a preference for centralizing provision in one supplier, which is missed by competition claims.

There is an open door to examine and suggest alternatives to dismantle for-profit natural monopolies. One could aim for truly publicly funded platforms governed as a global or local commons for services seen as socially necessary. For instance, internet search engines could be offered as global commons. Amid the ecological breakdown, online marketplaces, on the contrary, could instead be local commons. On such a publicly funded platform, advertising must be forbidden. Advertising not only induces consumption and broadly shapes behaviours but also downgrades the search engine service. There is no reason to believe that those paying to be listed at the top will be the results that best match the search. All in all, there is space for exploring alternative governance models while further analysing platforms, their heterogeneities and their interplay amid capitalism. Platforms & Society could be the place where such a research agenda blossoms.