Digital platform economies: Value from data?

Introduction ¹

Digital platform economies are heterogeneous terrain. Platform economies are constituted by disparate digital and non-digital elements. They include datasets, algorithms, application programming interfaces (APIs), programming languages, networked computational systems, business models, distributed storage facilities, as well as diverse public and private stakeholders and participants. And these arrangements entail various computational forms (algorithms, datasets) and modalities for producing value (multi-sided markets, monetary and non-monetary metrics, governance rules). Moreover, there are myriad types of platforms: financial technology platforms, media platforms, social media platforms, e-commerce platforms, content management platforms, network infrastructure platforms, cloud computing platforms, machine learning platforms, artificial intelligence platforms, digital marketing platforms, blockchain platforms, decentralized autonomous organization platforms, and so forth. And these operational and automated functionalities are often combined, making “platformization” a dispersed and sometimes contradictory process.

Given this constitutive heterogeneity, the contributions to this Themed Issue focus on a common set of questions to illuminate the manifold ways that “value” is realized on digital platforms: How does data become a value form? How are the artifacts produced by distinct digital platforms (e.g. datasets, content, licenses, tokens) translated into value forms (e.g. asset, commodity, property, capital, rent, revenue)? What aspects of digital infrastructures are essential for these translations (e.g. APIs)? How are forms of data-value made actionable (monetized, priced, exchanged, marketed, securitized)? And what are the consequences for conventional approaches to power or governance, most typically taken as an a priori that is enacted through material infrastructures? Answers to these questions provide insights into the design and actualization of data-valuation practices. Ultimately, they help us to better understand continuities and discontinuities across different modes of digital value production.

Digitalized value production is most often understood in terms of capitalist accumulation, or “platform capitalism,” which depends on data capture and enclaves that generate value through either monetization as a commodity or extraction as a monopoly rent (Birch, 2023; Langley and Leyshon, 2017; Sadowski, 2019; Srnicek, 2016; Zuboff, 2019). This process is characterized, more precisely, as “dual value production,” whereby monetary value is generated through service provision while, simultaneously, speculative value is derived from datasets and analytics (van Doorn and Badger, 2020). This description pertains most aptly to the gig economy and some aspects of Big Tech platforms. Questions arise, however, when one considers the general validity of this interpretation, both in sectoral and geographical terms.

For instance, while digital twins—or virtual real-time representations of physical objects, processes, people, and environments—intensify digital enclosures, this cannot be reduced to a function of their commodity status or speculative value. As Andrejevic, Horn, and Richardson show in this Themed Issue in their examination of digital twins at different scales—from the human body to the warehouse to the planet—by exploiting disaggregated, non-interoperable systems, digital twins constitute data streams in novel ways through aggregation, integration, and coordination, thus intensifying data enclaves. However, as they argue, digital twins invert the traditional hierarchy, consisting of top-level managerial processes presiding over the realms of bodies, labor, and enterprise operations, because the coordination of decision-making itself is the primary site of value creation. The commodity value and speculative value of a digital twin are supplemented by, if not dependent upon, its operational value.

In some cases (Huber, Rennie, Nabben), value is generated not through the operations of data capture and enclaving, but rather through interoperability and enhanced mediation. For instance, Huber's account of the development of data infrastructures in the US healthcare industry illustrates the role of APIs in value generation. Here, APIs are not developed for monopolistic enclaving, as is typically the case for Big Tech firms, which give third-party access to operating systems through APIs in order to maintain closed systems that maximize value produced through multi-sided markets. In contrast, Huber shows that new health data regulations in the US establish interoperable data infrastructures to thwart closed data systems. However, their aim is not to disrupt monopoly power; instead, they seek to replicate platform business models to generate both private and public goods. The point is that APIs, or interoperability, cannot be reduced to a technical feature of platforms because they are constitutive of particular modes of value creation—such as distributed, cross-sectoral forms of coordination, and the capacity to exchange and mediate across data infrastructures rather than siloing these systems toward rent extraction.

Indeed, because “value” is constituted and made actionable in diverse forms (capital, property, asset, commodity, rent, revenue stream), the contributions to this Themed Issue take value itself as a primary category of investigation. This is in keeping with a pragmatic approach to value production—informed by both the anthropology of value (Appadurai 1986; Guyer 1995; Munn 1986; Strathern 1988, 1992) and the social studies of finance (Callon, Millo and Muniesa 2007; Latour and Lépiney 2009; MacKenzie 2006). And it entails suspending the assumption that data and the operations of digital platforms have intrinsic value—whether understood through specific moral or political axioms or more diverse underlying presuppositions—and, further, that they necessarily take the form of either “rents” or “commodities.” To that point, several commentaries (Rennie, Nabben, Cooiman) indicate how particular modes of value production depend upon translation and commensurability between value forms, as opposed to the production and reproduction of a singular form (e.g. commodity).

These points of translation (see Strathern 1992) can be observed through various analytical devices: for instance, as valorization chains (Cooiman), as “relational value” (Nabben), and as “interobjective value” (Rennie). Cooiman's study of a venture capital fund illustrates how disparate financial forms (data-asset, revenue, fund-asset) are made commensurable along what she calls the valorization chain. Making those points of translation and commensuration visible means shifting from a singular focus on the platform firm to an analysis of the entire investment chain, which allows one to account for the generation of both financial and operational value for the fund. Rennie likewise shifts the analytic from the firm to distributed resources and distributed agency. Her presentation of contribution systems, or decentralized platforms based on mechanisms for the validation and potential reward of contributions, demonstrates that value is not expressed through price (exchange theory of value) or commodity production (labor theory of value). Instead, interdependent interactions continuously validate and reward contributions made by both human and machine agents, thus generating “interobjective value.” Similarly, Nabben shows how a decentralized data ecosystem integrates economic (market mechanisms), legal (a data trust regulatory framework), and social enforcement mechanisms to create a model of “located accountabilities” (cf. Suchman 2002) between data contributors, intermediaries, and consumers. Here, in contrast to Big Tech and most gig economy platforms, the role of data contributors is accounted for, and a decentralized “data mesh” generates relational value for the nexus of stakeholders.

These commentaries provide insights into platform economies by specifying the operations of financial practice: for example, by substantiating how data figures in processes of both capitalization and monetization (see also Birch and Muniesa 2020; Birch et al., 2021). These processes depend fundamentally on the qualification of data as an economic or financial form, as we see quite distinctly in the case of foundational models, or machine learning and deep learning models trained on vast datasets—the new frontier for value creation in the digital economy (Goldenfein). Ultimately, that process depends on conceptual integrity, or whether these datasets are qualified as expression (“content”) versus latent statistical relations within that expression (“data”). As we know, vast datasets are monetized through consumer-facing AI products, such as ChatGPT. However, the increasing unlicensed use of these datasets raises the question of whether their value derives from their definition as “content”—and hence rights defined by copyright law—versus “data,” in which case they can be commodified or monetized in alternative ways. As Goldenfein shows, this question of data-versus-content is crucial because it establishes different claims to the very right to generate value from data.

Existing and emerging regulatory frameworks don’t merely set rules for the existing and emerging digital economy: regulations constitute digital value and hence the contours and substance of the digital economy (Goldenfein, Huber, Wu, Rennie, Nabben, Cooiman). Regulatory mechanisms are based on category distinctions (e.g. data versus content) that determine modes of value creation. They establish the contours of data enclaves, but they are also the basis for distributed data-assets; the valorization of new categories of mediation; and the creation of public, private, and distributed goods (Cooiman, Rennie, Nabben, Huber). Indeed, they raise fundamental questions about the public-private distinction, or the ways that this distinction is constituted or displaced. This is exemplified by the case of “the Golden Shares,” or equity stakes taken by the Chinese state in local private media companies that grant the state special voting rights. Ironically, this ostensible state subsidy led to the dramatic devaluation of these private media platforms on global capital markets (Wu). As Wu argues, “maximizing financial value in one context doesn’t [necessarily] translate into another context.” Here, financial value depended not on technologies of capture and surveillance, but on the constraints and affordances of public regulatory and governmental mechanisms, and the ensuing practical management of private capital market valuations—a highly instructive point, given the prospects of government-granted monopolies in both China and the United States today.

Documenting data-valuation practices and processes is not a matter of affirming the inexorable efficacy of digital platforms and the totalizing effects of a digital economy. By suspending assumptions about the teleology—or presumed ends and achievements—of value generation, studies of data-valuation consider both successful and failed attempts to produce actionable forms of value (per Wu; Goldenfein; Rennie; Nabben; Andrejevic, Horn and Richardson herein; and see Roitman, 2024). These studies build from prior work on digital platforms, which defines them as structures for value-generating interactions, or multi-sided markets (Constantinides et al. 2018; Kenney and Zysman 2016; Pon et al. 2014; Rochet and Tirole 2003; Sanchez-Cartas and León 2021; Smedlund et al. 2018; Van Alstyne et al. 2016). However, this focus on structures of value generation is reductive in its narrow concern with commodity market exchange and underlying assumptions about market efficiency. A corrective to this restrictive focus builds from past scholarship on the materiality of information networks (Burrell 2016; Dourish 2016, 2017; Helmond 2015; Jaton 2020; Lowrie 2017; Mackenzie 2017; Neyland 2016, 2019; Seaver 2017) and yet accounts for the potential heterogeneity of value forms, thereby documenting diverse modalities of data-valuation. This research demonstrates how digital platforms entail both extremely effective and sometimes surprisingly ineffective economic and financial operations (Bernards, 2019; Birch and Muniesa 2020; Birch et al. 2021; Breckenridge 2019; Eben 2018; Hansen and Thylstrup 2024; Meese 2023; Mellet and Beauvisage 2020; Narayan 2022; Roitman 2025; Westermeier 2020).

The Themed Issue offers a multidisciplinary perspective on the question of data valuation by bringing different forms of expertise to bear upon a set of shared questions regarding the production of digital value and emerging modes of valuation. The various cases establish the grounds for the examination of continuities and discontinuities across different modes of value production, thereby setting forth and clarifying programmatic research. While much current research seeks to establish a general analytic and consequential critique of platforms, few studies examine the concepts that are either assumed or mobilized in academic inquiry—let alone the conceptual shifts that render foundational terms, such as “economy” and “capitalism,” both legible and operational (see Tribe 1981). Looking forward, the aim of this programmatic research is to evaluate the extent to which the foundational concepts of social theory (e.g. capital, property, commodity, rent, labor) are useful and/or relevant to the study of platform valuation processes and platform economies, more generally.