In the first of this two-part blog post, Sam Gilbert questions the widely-held view that the targeting of digital advertising based on individuals' online activity is morally wrong. Building on an argument made in his book Good Data, Sam suggests that from a utilitarian perspective, business models based on targeted advertising can be said to benefit the Global South, and could even be seen as an enabler of distributive justice.
Digital advertising is often targeted at individuals by inferring their consumer preferences from data about their online activity – the websites they visit, the social media posts they like and so on. In academia and public policy circles, the ubiquitous surveillance of online behaviour that is required for targeting systems to work is widely seen as morally objectionable.
But is there something to be said in its favour? In this post, I want to develop a counterintuitive claim made in my book Good Data: that from a utilitarian perspective, the surveillance advertising business model serves the interests of the Global South, and can even be seen as an enabler of distributive justice. In a second post, I turn to the policy implications of this analysis, showing that the subscription business model advocated by many critics of surveillance advertising could be expected to have regressive effects were it applied to social media. I focus on Meta, but the argument is applicable to other widely-used services funded by surveillance advertising – including those provided by Google’s parent company Alphabet.
Those who object to surveillance advertising typically do so for three reasons. First, they view the collection of user data for the purposes of targeted advertising as intrusive and covert, and therefore regard it as a violation of privacy. Second, they worry that the use of this data may expose individuals to manipulation based on inferences about their psychological vulnerabilities, further diminishing their freedom. Third, they view the social relations created by the surveillance advertising business model as extractive, seeing user data as a form of digital property or labour surplus, coercively taken from users, whose condition is therefore one of serfdom.
This view of surveillance advertising business models is widely-held, demonstrably influencing actors such as the European Union (EU) Competition Commissioner Margrethe Vestager and United States (US) Senator Elizabeth Warren. It has gained particular political traction in Europe, where the European Parliament and the European Data Protection Supervisor have recommended that surveillance advertising be phased out and eventually banned. Others, including the Danish government’s Tech Expert Group and the Norwegian Data Protection Authority, have gone further, respectively proposing and implementing an immediate ban.
However, in both a geographical and a philosophical sense, it is a rather Eurocentric view – one that implicitly assumes that the interests of users in the Global North are what matters, and neglects the interests of users in the Global South. This is especially problematic given the geographic distribution of the users of services funded by surveillance advertising. If we look at monthly active user numbers for Meta’s Facebook platform, we find that only one of its top 10 country markets is in the Global North. By this metric, Facebook’s largest market is India (315 million (mn) users), and not the US (175mn). The Philippines (80mn) has significantly more Facebook users than the United Kingdom (34mn) and France (30mn) combined, while there are more Facebook users in Egypt (42mn) than in Germany (25mn). From a utilitarian perspective, if we are to make moral judgements about surveillance advertising as a business model, we must make sure its consequences for users in the Global South are appropriately weighted in the calculation.
When Meta is looked at in global terms, it becomes clear that the economic effects of its surveillance advertising business model are surprisingly progressive. This is a function of the huge differences in the value to advertisers of users’ attention and clicks depending on their geographic location. Crudely, advertisers are willing to pay much more to reach a user in Germany than a user in Egypt, because the German user is likely to have much more money to spend on the advertiser’s products than the Egyptian user. At the same time, subject to their devices, bandwidth, and ability to speak at least one of 111 officially-supported languages, all users can access the same services from Meta – whether they are in Germany, Egypt, or elsewhere. As David Runciman puts it, “the provision of free networking services in exchange for access to personal data creates inadvertent advantages for the disadvantaged, whose data are worth less to advertisers but who get the services anyway”.
Measuring the economic value to users of free digital services is notoriously difficult, so how are those “inadvertent advantages” to be quantified? Proponents of a subscription model for social media often suggest that users should pay a subscription fee equivalent to Meta’s average revenue per user to access Facebook and Instagram, in exchange for not being targeted with advertising. In 2021, according to Meta’s annual report and accounts, this figure was $43.60 overall, but varied widely based on geography: it was $202.12 in North America, $70.78 in Europe, $23.09 in Asia-Pacific, and just $12.02 in the Rest of the World region. If we follow these scholars and commentators, and treat this revenue per user metric as a reasonable proxy for the value users receive from the services Meta provides, we can say that each user in North America and Europe subsidized users elsewhere to the tune of $158.52 and $27.18 respectively, while each user in Asia-Pacific and the Rest of the World gained $20.51 and $31.58 respectively (see Table 1).
The overall outcome was that in 2021 some $51.5bn of value was effectively redistributed away from the Global North to the Global South – slightly more than the $47.8bn spent by the US government on overseas development aid that year.
Table 1: Transfer of value between Meta users in different world regions, 2021 – Revenue Basis
Some might object to the characterization of this effect as a subsidy, on the basis that Meta might simply view the different regions of the world as different markets with different profit opportunities. However, when Meta’s $71.2bn of costs and expenses are brought into the analysis, it becomes clear that both the Asia Pacific and Rest of the World regions are unprofitable when these costs are allocated to users equally (see Table 2). Meta does not provide a regional breakdown of its costs, but we could reasonably assume that the pro rata cost of providing services in the Global South is actually higher than in the Global North given Meta’s investments in physical infrastructure such as cell tower networks and submarine cables. In that case, the transfer of value from the Global North to the Global South would be even greater than $51.5b.
Table 2: Transfer of value between Meta users in different world regions, 2021 – Gross Profit Basis
It seems that – at least in the case of Meta – the surveillance advertising business model delivers huge economic benefits to the Global South. If we were to turn to John Rawls, we might even call it a form of distributive justice, since it satisfies the difference principle by most benefiting the least advantaged.
Read the second part of this blog: “Good Surveillance?” The implications for public policy
Acknowledgements
The author thanks Dr Maha Rafi Atal, Professor Lisa Ann Richey, Sofie Elbæk Henriksen and participants in the workshop “Commodifying Compassion in the Digital Age” at Copenhagen Business School for helpful comments and discussion during the development of this work.
The views and opinions expressed in this post are those of the author(s) and not necessarily those of the Bennett Institute for Public Policy.
