The platform – an infrastructure that connects two or more groups and enables them to interact – is crucial to these companies’ power. None of them focuses on making things in the way that traditional companies once did. Instead, Facebookconnects users, advertisers, and developers; Uber, riders and drivers; Amazon, buyers and sellers.
Reaching a critical mass of users is what makes these businesses successful: the more users, the more useful to users – and the more entrenched – they become. Ello’s rapid downfall occurred because it never reached the critical mass of users required to prompt an exodus from Facebook – whose dominance means that even if you’re frustrated by its advertising and tracking of your data, it’s still likely to be your first choice because that’s where everyone is, and that’s the point of a social network. Likewise with Uber: it makes sense for riders and drivers to use the app that connects them with the biggest number of people, regardless of the sexism of Travis Kalanick, the former chief executive, or the ugly ways in which it controls drivers, or the failures of the company to report serious sexual assaults by its drivers.
Network effects generate momentum that not only helps these platforms survive controversy, but makes it incredibly difficult for insurgents to replace them.
As a result, we have witnessed the rise of increasingly formidable platform monopolies. Google, Facebook and Amazon are the most important in the west. (China has its own tech ecosystem.) Google controls search, Facebook rules social media, and Amazon leads in e-commerce. And they are now exerting their power over non-platform companies – a tension likely to be exacerbated in the coming decades. Look at the state of journalism: Google and Facebook rake in record ad revenues through sophisticated algorithms; newspapers and magazines see advertisers flee, mass layoffs, the shuttering of expensive investigative journalism, and the collapse of major print titles like the Independent. A similar phenomenon is happening in retail, with Amazon’s dominance undermining old department stores.
More and more companies are coming to realize this. We often think of platforms as a tech-sector phenomenon, but the truth is that they are becoming ubiquitous across the economy. Uber is the most prominent example, turning the staid business of taxis into a trendy platform business. Siemens and GE, two powerhouses of the 20th century, are fighting it out to develop a cloud-based system for manufacturing. Monsanto and John Deere, two established agricultural companies, are trying to figure out how to incorporate platforms into farming and food production.
And this poses problems. At the heart of platform capitalism is a drive to extract more data in order to survive. One way is to get people to stay on your platform longer. Facebook is a master at using all sorts of behavioural techniques to foster addictions to its service: how many of us scroll absentmindedly through Facebook, barely aware of it?
Another way is to expand the apparatus of extraction. This helps to explain why Google, ostensibly a search engine company, is moving into the consumer internet of things (Home/Nest), self-driving cars (Waymo), virtual reality (Daydream/Cardboard), and all sorts of other personal services. Each of these is another rich source of data for the company, and another point of leverage over their competitors.
Others have simply bought up smaller companies: Facebook has swallowed Instagram ($1bn), WhatsApp ($19bn), and Oculus ($2bn), while investing in drone-based internet, e-commerce and payment services. It has even developed a tool that warns when a start-up is becoming popular and a possible threat. Google itself is among the most prolific acquirers of new companies, at some stages purchasing a new venture every week. The picture that emerges is of increasingly sprawling empires designed to vacuum up as much data as possible.
But here we get to the real endgame: artificial intelligence (or, less glamorously, machine learning). Some enjoy speculating about wild futures involving a Terminator-style Skynet, but the more realistic challenges of AI are far closer. In the past few years, every major platform company has turned its focus to investing in this field. As the head of corporate development at Google recently said, “We’re definitely AI first.”
All the dynamics of platforms are amplified once AI enters the equation: the insatiable appetite for data, and the winner-takes-all momentum of network effects. And there is a virtuous cycle here: more data means better machine learning, which means better services and more users, which means more data. Currently Google is using AI to improve its targeted advertising, and Amazon is using AI to improve its highly profitable cloud computing business. As one AI company takes a significant lead over competitors, these dynamics are likely to propel it to an increasingly powerful position.
What’s the answer? We’ve only begun to grasp the problem, but in the past, natural monopolies like utilities and railways that enjoy huge economies of scale and serve the common good have been prime candidates for public ownership. The solution to our newfangled monopoly problem lies in this sort of age-old fix, updated for our digital age. It would mean taking back control over the internet and our digital infrastructure, instead of allowing them to be run in the pursuit of profit and power. Tinkering with minor regulations while AI firms amass power won’t do. If we don’t take over today’s platform monopolies, we risk letting them own and control the basic infrastructure of 21st-century society.
Nick Srnicek, a lecturer in digital economy at King’s College London, is the author of Platform Capitalism