Platform Capitalism by Nick Srnicek

Srnicek's Platform Capitalism is an important contribution for understanding the economic structures of the current moment, given the dominance of the platform capitalism model (ideologically and economically), the centralization of power by Big Tech during the COVID-19 pandemic (adoption of digital tech), and the Big Tech Anti-Trust hearings in the U.S.

There are roughly three parts to Srnicek's investigation: 1) historicization of platform capitalism; the current manifestation of platform capitalism is the result of previous crises and as a part of the logic of capitalism; 2) a typology of the various platforms; and 3) the future of platform capitalism.

Platform capitalism, while appearing as a novel emergence, is connected to the past few decades of capitalism. Srnicek focuses on three crises in capitalism in particular: the 1970s downturn, the boom and bust of 1990s, and 2008 financial crisis.

Srnicek's story begins in the post-World War II period, marked by social democratic policies and a large manufacturing sector that provided stable employment for a large number of workers. Manufacturing at that time was marked by a "just-in-case" approach (in contrast to "just-in-time" manufacturing) that justified retaining workers and inventory in reserve. From the 1970s, competition from Japanese and German manufacturers led to an influx of manufacturing goods, a downward pull on prices of goods, and reduced profitability. Firms responded by slimming down their operations (moving from “just-in-case” to “just-in-time” manufacturing) and through a frontal assault on labour power and labour unions, leading to the outsourcing of jobs overseas and contracting labour on increasingly flexible and low wage contracts.

The boom-and-bust period of the 1990s was marked by speculative investment in the Internet, which was still unmatched at the time of the writing of the book (2017). The amount of capital invested led to the construction of key digital infrastructure, like “millions of miles of fibre-optic and submarine cables… major advances in software and network design… and large investments in databases and servers” (p. 22). When the bubble burst, and the bust came, government response came in the form of “asset-price Keynesianism” (p. 24). Instead of deficit spending, the government cut interest rates, leading to investments of capital in financial assets and housing. Loose monetary policy is one of the legacies of the boom-and-bust period.
 
Finally, during the financial crisis of 2008, the government took on large deficits to bail out Wall Street, turning “high levels of private debt … into high levels of public debt” (p. 26). Continuing the trend from the boom-and-bust period, key interest rates fell to near-zero levels. Due to the rise in deficits, federal governments started to impose economic policies of austerity, cutting budgets and raising taxes. In place of fiscal stimulus, governments began to create new types of monetary policy, principal among them quantitative easing, in which money is created by the central bank to purchase various forms of assets, lowering the interest rates of longer term assets and leading to investors seeking higher yields in riskier investments, like in the many untested tech companies in the market today. Meanwhile, the new dominant tech companies in the market are able to move intellectual property to low tax jurisdictions to evade taxes, draining government revenues even further, and leading to a vicious cycle of “tax evasion, austerity, and extraordinary monetary policies” (p. 33). While this is happening, there is trend towards precaritization of work and long-term unemployment.
 
It is in the rubble of these historical and economic trends that platform capitalism is situated. Thanks in part to the boom-and-bust period of the late 1990s, there is a large amount of infrastructure to record, collect, and analyze behavioural data. Data can be used for the following functions: 1) algorithm development for competitive advantage; 2) “coordination and outsourcing of workers”; 3) “optimization and flexibility of productive processes”; 4) “transformation of low-margin goods into high-margin services; and 5) “data analysis,” which is “itself generative of data, in a virtuous cycle” (p. 42). Further technological advancements are expanding the type of data that can be collected and it is making data collection increasingly cheaper.
 
Platforms are a new business model optimized to make use of the new raw material, data. Srnicek defines them as “digital infrastructures that enable two or more groups to interact” (p. 43) as intermediaries that enable user-generated content. Through the platform, businesses are able to record user interactions and user activities. Given these characteristics, digital platforms benefit from “network effects”, where “the more numerous the users who use a platform, the more valuable that platform becomes for everyone else” (p. 45)—social media sites like Facebook are obvious examples. The importance of network effects means that platforms naturally strive toward monopolies, given cheap marginal costs that allow them to scale rapidly. In contrast to the lean business models used by the manufacturing sector in the post-1970s era, platforms use “cross-subsidization” to attract and retain users, where some functions are given for free and are subsidized by other functions that generate the revenue.
 
Finally, design considerations are important for platforms. Design attracts and retains users (i.e., UX) and, more importantly, design of platforms sets the ground rules for interactions and platform governance. Despite apolitical pretensions, platforms are both political in nature (as they represent a system of governance) and affect politics—this is hardly a controversial statement given the events of the past few years (e.g., Cambridge Analytica, the Rohingya genocide, etc.).
 
Srnicek identifies five different types of platforms:
 
1) Advertising platforms: The first iterations of platform capitalism and the primary revenue generating model of Google and Facebook.
2) Cloud platforms: Cloud platforms offer services like storage, computing power, applications, etc. and are becoming basic infrastructure for the digital economy; also allows privileged access to data.
3) Industrial platforms: Platforms are entering traditional manufacturing; industrial IoT produces data and allows for optimization of production processes.
4) Product platforms: Product platforms rent out products; in the case of digital products, they take advantage of zero marginal costs for digital products (e.g., Spotify), while for physical products like Rolls Royce aircraft engines, data is used for maintenance and repair.
5) Lean platforms: Lean platforms like Uber and Airbnb are often assetless and embody the results of previous economic crises; they rely on precarious labour and are driven by the cheap availability of capital—they are rarely profitable.
 
Srnicek ends with a discussion on the future of platform capitalism by focusing on intra-capitalist competition. As per the previous discussion, platforms tend toward monopolies to exploit network effects; these monopolies can be harder to overcome for new businesses due to access to data and established networks. Platform competitiveness is based on data collection and analysis, and the various great platforms tend to converge and encroach into each other’s markets to occupy key positions (e.g., Uber’s attempt to buy a mapping provider). As Srnicek writes, “companies like Facebook, Google, Microsoft, Amazon, Alibaba, Uber, and General Electric (GE) are also direct competitors” (p. 108). While there is a trend towards convergence, large platform companies are looking to silo their platforms and create enclosed ecosystems to tie their users to the platform, potentially leading “from an open web to increasingly closed apps” and “driving the internet to fragment” (p. 113).
 
Srnicek calls for public platforms instead of mere state regulation of corporate platforms. These platforms would be outside of the purview of the state and would not function as a part of the state surveillance apparatus, instead, it would be controlled by the public. These could be offered as public utilities, and the data collected could used democratically, instead of being extracted by private platforms.

---