Companies and the UK government pulled advertisements from YouTube after it had been revealed that some of these organisation’s ads were presented together with videos featuring anti-Semitism, white supremacist ideology, hate speech inciting violence, and right-wing extremism. Organisations that have removed ad investments from YouTube include Audi, BBC, Channel 4, EDF Energy, Havas ad agency (representing e.g. Domino’s, HSBC, Hyundai, O2, Royal Mail), L’Oréal, Lloyds, Marks & Spencer, McDonald’s, RBS, Sainsbury’s, Tesco, The Guardian, Toyota, Transport for London, UK Cabinet Office, UK Financial Conduct Authority, and Volkswagen.
The mainstream media reported about this phenomenon, which resulted in public complaints that Google did not properly control YouTube content. Such arguments misunderstand the political economy of algorithmic advertising. The problem is not simply a moral one, but has to do with how big data capitalism works. One is surprised that large organisations have only by now become aware of some of the pitfalls and problems of big data capitalism and have invested billions into targeted online advertising. Big data promises big audiences, but in the end results in big scandals.
According to the Forbes 2000 list, Alphabet/Google is the world’s 27th largest transnational corporation. The company’s profit amounted in 2016 to US$ 19.5 billion. Advertising accounted for 87.9% of Alphabet’s revenues. YouTube has a double use-value: It provides user-generated videos to viewers and a targeted advertising platform to ad clients. The first use-value is given away as a “free lunch” to users in order to amass large numbers of users, viewers, engagements and klicks. Big data gathered about the users allows the targeting of ads. YouTube sells ad space that is targeted to user groups. Google sells the combination of online ad space and users’ attention as big data commodity.
YouTube is only in respect to the use-values it provides a media content company. In respect to what it sells and in terms of exchange-value it is the world’s largest advertising company. It lives by turning user activity into labour that creates big data commodities. YouTube’s data is big in two respect: One hour of new videos is uploaded to YouTube per second — that is 31,536,000 hours or 1,314,000 full days of video content per year. One person would need 3600 years to watch all the content uploaded in one year to YouTube. There are more than one billion YouTube-users. YouTube involves big content data and big user meta-data. Big data capitalism is a capitalism of large numbers. YouTube is part of the capitalist openness industry: It favours the open distribution of content without access restriction and the de-commodification of access in order to commodify user data in the form of targeted advertising.
In 2016, Alphabet/Google had “72,053 full-time employees: 27,169 in research and development, 20,902 in sales and marketing, 14,287 in operations, and 9,695 in general and administrative functions”. Big data — and relatively small employee numbers. Chemical company BASF and car manufacturer Hyundai’s revenues are in the range of Google’s. But the number of employees of both companies is around 115,000, 60% higher than at Google. One of the online economy’s ideologies is that data and the Internet are new economic wonders. At least in respect to jobs this is certainly not the case.
Not humans, but algorithms decide on which YouTube and Google-content specific advertisements are featured. Algorithmic advertising and sales only require relatively little human labour. Data trade is a highly automated capitalist business, in which algorithms organise the commodity sales process.
One can target Google-video ads by location, demographics (gender, age, parental status, household income) and interests. The ad algorithm identifies individuals with the chosen characteristics and presents ads to them while they watch YouTube videos. A YouTube ad client can select particular videos or leave the choice open. Popular YouTubers can become YouTube partners and earn money by allowing ads on their video content. The ad revenue is shared between YouTube and its partners.
Digital capitalism entails big data commodity fetishism: Users’ sociality and activity disappear behind screens and algorithms. The algorithmic machine organises the data commodity, the viewing process, and the mapping of ads and users. Google’s profits live and grow through algorithms, instrumental logic, online automation, and machinic behaviour.
The commodity and the algorithm are blind for morality and meaning. They empty online communication from its meaning and fill the created void by automatically presenting advertisements. The algorithm does not care if the content it matches with ads contains dog poo, beef stew or a fascist cuckoo. The logic of algorithms and the logic of the commodity know no ethics and no morals. They only know the instrumental logic of numbers. The immorality apparent in the pairing of fascist videos with mainstream organisation’s ads is the very consequence of the automation of human activity in big data capitalism.
It is certainly “unacceptable that taxpayer-funded advertising has appeared next to inappropriate internet content”. But why is taxpayers’ money not used for funding public services, but rather for advertisements? Why is it transferred to Google? A company that has come under attack for alleged tax avoidance. First and foremost, big data capitalism, targeted advertising, and the automation of human online activity are inappropriate. What appears as immoral ad practice, where online ads of respected organisations meet fascism on YouTube, is just the outflow of big data capitalism’s inherent immorality of automation, accumulation and dehumanisation.
What can be done? Clear guidelines and policies that ban fascist content as well as employment of more staff for checking policy violations would certainly be a good step. But it is not enough. The problem is of a more fundamental nature: Big data capitalism negates public interest content and public interest platforms. Part of the problem is that Google holds a de-facto monopoly in online search and online video. The logic of minimising human intervention by algorithmic automatic is part of big data capitalism’s business models. Algorithmic advertising is part of the problem. The only solution is that we re-invent Internet platforms. We need public interest platforms that are not about advertising and selling commodities, but about critical and educational content that engages humans in debate, co-creation and fascination. If a world without advertising works on the BBC, then why should it not also be possible online and on social media? The online world needs socio-technological and institutional innovations that transcend big data capitalism. We need small data, slow media, non-profit social media, and non-commercial platforms.
How do big data capitalism and its possible alternatives work? Ten experts will discuss this question in the conference “Digital Objects, Digital Subjects: An Interdisciplinary Symposium on Activism, Research & Critique in the Age of Big Data Capitalism”. The event will take place on May 20 and 21 at the University of Westminster. It will feature talks by the philosopher Antonio Negri, the political scientists Jodi Dean and David Chandler, the design scientist and historian Orit Halpern, the legal studies scholar Antoinette Rouvroy, the art and design theorist Etienne Turpin, and the media and communication studies scholars Paolo Gerbaudo, Kylie Jarrett, Jack Linchuan Qiu and Christian Fuchs. They will in an interdisciplinary symposium engage in the question whether a human and digital world beyond big data capitalism is possible.