Nick Srnicek's platform capitalism framework explains how companies extract value by controlling infrastructure. IS researchers can study what this does to worker agency.
Nick Srnicek's 2017 book "Platform Capitalism" gave me a vocabulary for something I had been noticing but could not name. The argument is that a new type of firm has emerged whose business model is not to produce goods or services in the traditional sense, but to control the infrastructure through which other people's economic activity flows. The platform does not make the ride. It controls the system through which rides get matched, priced, dispatched, and evaluated. It does not make the deliveries. It owns the algorithmic layer that connects drivers to orders and restaurants to customers. The value the platform extracts comes from its position as the indispensable intermediary, the entity you cannot route around if you want to participate in the market.
What struck me about Srnicek's framework is that it made visible something that market descriptions of these companies tend to obscure. The standard story is that Uber is a technology company that happens to coordinate transportation, that DoorDash is a logistics platform, that Amazon Mechanical Turk is a marketplace for human intelligence tasks. What Srnicek's analysis shows is that these companies derive their power not from technology per se but from the data they accumulate as a byproduct of platform activity. Every ride, every delivery, every task completed on Mechanical Turk generates data that the platform controls. That data is used to improve matching algorithms, to price labor, to evaluate workers, and to shape the conditions under which workers operate. The workers generate the data that is then used to manage them.
This is where the IS research angle becomes interesting to me. The phenomenon Srnicek describes, platforms extracting value by controlling the infrastructure through which workers and customers interact, is fundamentally an information systems phenomenon. It depends on specific technical architectures, specific data collection and processing designs, specific algorithmic decision systems. IS researchers study exactly how systems are designed, how they function in organizational contexts, and what the consequences of those design choices are for the people who use them.
The "algorithmic management" concept is particularly relevant here. On traditional platforms, workers are managed not by human supervisors who observe their behavior and give feedback through relationships, but by ratings, performance metrics, dispatch algorithms, and dynamic pricing systems. A driver on a ride-hailing platform is managed by an algorithm that determines which requests they receive, how much they are paid, whether they receive a warning, and whether they are deactivated. The decisions that a human supervisor would make in a traditional employment relationship are made by a system, at scale, with minimal human review. I find this striking from an IS perspective because it means that the design of the algorithmic management system is, in a real sense, the design of the employment relationship.
The IS questions that follow from this are not primarily about adoption or acceptance in the traditional TAM sense. They are about worker agency: how much discretion do workers retain over their own labor when the management system is an algorithm? They are about transparency: do workers understand why the algorithm made a particular decision, whether to lower their rating or to reduce their dispatch frequency? They are about collective action: how do workers organize when there is no single employer to organize against and when work is performed at scattered times and locations? These questions push IS research toward political economy in ways that most IS papers still avoid.
The data dimension is the one I think IS researchers are most uniquely positioned to study. Workers on platforms generate what I would call behavioral data about how work actually happens: routing decisions, timing, the sequence of micro-decisions that constitute skilled labor. This data is economically valuable, and the platform controls it entirely. The worker who has spent three years learning the most efficient routing in a city has generated data that encodes that expertise, and that data belongs to the platform. From an IS governance perspective, this raises questions about who owns data generated through labor, whether workers should have rights to data about their own work, and what oversight mechanisms would look like for algorithmic management systems.
The research I would want to see from IS scholars is fieldwork and experimental work that goes inside platform labor relationships. Not aggregate statistics about how many gig workers there are or what they earn on average, but fine-grained study of how specific algorithmic management systems shape the day-to-day experience of platform work. What information do workers have access to? What signals does the algorithm give them? How do they adapt to algorithmic management, develop workarounds, or resist? Möhlmann et al. (2025), working on algorithmic stakeholder governance, are starting to ask some of these questions at the platform governance level. But the worker-side story still feels undertold in IS research.
The platform capitalism framework also raises a question that IS has historically not engaged with much: the political economy of infrastructure control. When a small number of platform companies control the infrastructure through which a significant share of labor operates, questions about data sovereignty, pricing power, working conditions, and regulatory oversight become infrastructure questions, not just employment questions. That is squarely within IS territory, and I think the field has not caught up to the stakes yet.
About the author
Share
More notes
Related notes