The Platform Decides Who Wins

In 2018, Apple removed the app Parental Controls from its App Store. The stated reason was privacy. The practical effect was removing a product that competed directly with Apple's own Screen Time feature, which launched the same year. The developers had agreed to terms giving Apple the authority to make exactly that decision.

That story does not make sense if you think of Apple as a company that sells products. It makes sense only if you understand Apple as a platform, and platforms as a particular kind of governance structure.

Parker, Van Alstyne, and Jiang (2017) drew the distinction as sharply as anyone has. A pipeline business creates value in a line: raw materials come in, transformation happens, a product goes out. A platform creates value by facilitating interactions between producers and consumers who would not have found each other. The platform owner does not produce the core value. The platform owner sets the rules under which others produce value and then takes a share. That shift from production to coordination sounds minor. It is not. It changes who holds power, how value is distributed, and what governance even means.

Platform ecosystem theory treats the platform as an architecture with a core and a periphery. The owner controls the core: the interfaces, the APIs, the terms of service, the right to participate. Complementors build in the periphery, adding products and services that make the platform more attractive, which draws more users, which draws more complementors. The cycle is self-reinforcing. It is also asymmetrical. The platform captures value from both sides without being directly exposed to the risks either side carries.

Tiwana (2014) argued that platform ecosystem governance has three components: who holds decision rights, what control mechanisms the owner uses, and who owns the interfaces and architecture. Those three questions determine how value flows inside the ecosystem. And the answers are made almost entirely by the platform owner, based on what maximizes the platform's own position.

Amazon is the cleanest case of what the literature calls platform envelopment. Amazon hosts third-party sellers on its marketplace. Amazon also competes with those sellers directly, often with private-label products that appear beside the seller's listings in the same search results. Amazon can see which products sell well, at what price, and in what volume, before any competitor outside the platform can. The sellers agreed to terms that permit this arrangement. They built their businesses inside a governance structure they did not design and cannot renegotiate. When Amazon enters a product category itself, the seller discovers their landlord is also their competitor.

Uber makes the governance question visible for labor. Uber drivers are classified as independent contractors, not employees. That classification is a governance decision, not an objective description. The platform sets the prices. The platform sets the surge multipliers. The platform decides when drivers receive rides and when their accounts are deactivated. Drivers cannot negotiate any of these terms. They can accept the arrangement or leave the platform. The independence in "independent contractor" describes schedule flexibility. Every economically significant variable is set by the platform's governance rules.

Möhlmann et al. (2025) made the stakeholder dimension explicit: on content platforms, algorithms do not just recommend content or automate moderation. They govern how stakeholders interact and how conflicts among them are resolved. Visibility, ranking, monetization, and safety are not neutral outcomes of a technical system. They are policy decisions encoded in algorithms. Platform governance is not only platform owner control over technology. It is also algorithmic stakeholder governance, where different groups have goals that conflict and the algorithm mediates those conflicts according to rules the platform wrote.

Mayer et al. (2025) showed how GenAI changes this governance problem. When AI becomes a boundary resource on a platform, it creates validation problems because outputs need checking, standardization problems because generative outputs resist stable platform rules, and complementor-skill problems because complementors need new capabilities to work with unpredictable AI outputs. The governance challenge is not just about who gets to participate. It is about whether the boundary resources that coordinate participation can even remain stable when the outputs they produce are open-ended.

The thing that keeps surfacing as I read this literature is that governance rules in platform ecosystems are not neutral. They are designed to protect the platform's core interests. When Apple removes a competing parental control app on privacy grounds while launching its own, that is governance working as designed. When Amazon uses data generated by third-party sellers to make private-label decisions, that is governance working as designed. When a content platform's algorithm downranks posts that do not generate engagement because engagement drives ad revenue, that is governance working as designed. None of these are governance failures. They are governance succeeding at what it is structured to do: protect the platform first, the high-value side of the market second, and the low-value side third.

IS research on platform governance is asking the right questions: who decides, who benefits, and who gets excluded. Platform companies are mostly asking one question: how to extract more value from both sides of the market simultaneously. That is not a cynical reading. It is a structural feature of the business model. The platform's function is to increase the value of interactions while increasing its share. Governance rules are instrumental to that function.

For how platform logic intersects with IT value, the argument in the productivity paradox applies directly. Platform businesses invest in infrastructure that creates enormous value for users while capturing a fraction of it through data and advertising, and the question of where value actually lands is more complicated for platforms than for traditional IT. And for why platform logic clashes with regulatory logic, institutional logics and competing rationalities captures the tension: the market logic of the platform and the public interest logic of the regulator disagree about what counts as value and for whom.

The governance question is not just who wins. It is what winning requires you to do to everyone else.