by Iva Nenadic
Political microtargeting practices of campaigns came into the spotlight in the aftermath of the 2016 US presidential elections and the UK’s Brexit referendum. The public discussions around these practices have largely been guided by the thought that microtargeting might have a devastating effect on democracy. However, the existing evidence seems not to support this thought. Moreover, there is a lack of relevant research on this matter outside the US context. The recently published special issue of Internet Policy Review, highlights the need to further understanding and analytical capacities around political microtargeting and to assess this practice taking into account different political, economic, social, and cultural conditions.
Following the definition of Bodó et al. (2017), political microtargeting is a technique of political communication based on the use of data and analytics to tailor messages to a subgroup or individuals via different channels (mail, phone, canvassing, direct mail, and social media advertising, etc.) in order to build a relationship with prospective voters and supporters. Ultimately, of course, they aim to persuade citizens to give them their vote. Political targeting is nothing new. Geographic and demographic targeting, based on the analysis of data sources, such as precinct-level results from past elections, have long been applied by campaign managers to focus and maximise their efforts (Kruschinski & Haller 2017). What is new today is that, unlike only several years ago, modern technology and its widespread application now provide the means to collect countless sets of individual data points, to apply sophisticated data mining tools and to use this information strategically in campaigns (Schipper & Woo 2017). Furthermore, the scope is no longer on data labeled as “political” – but increasingly about consumers behaviour and very personal information that people reveal more or less consciously through their online activity. Modern microtargeting classifies voters not only on the basis of demographics and issues, but also based on the personality traits that can be predicted with relatively high accuracy by analysing public data that people share online, including textual and behavioural features (see: Adali & Golbeck 2012 & 2014, Yang & Li 2013).
As already highlighted, big data deriving from citizens’ online behaviour, including from their social media use, is the main fuel of contemporary political microtargeting. Micro-targeted political persuasion can still be done offline – for example in the form of door-to-door canvassing – however, the focus of this paper, and of most discussions on this topic, is on online political microtargeting usually conducted via social media. It involves tailoring messages to individual voters, framing them around issues that match the interests of targeted voters (Zuiderveen Borgesius et al. 2018). The most relevant platform for this is Facebook. Facebook is by far the most popular social media network, being used by almost two billion people worldwide (Digital in 2017). It’s business model is based on amassing data on its users to personalise their experience and to target ads, including political ads. Political actors, therefore, can rely on Facebook’s low-cost service to convey targeted wide-reaching advertising or can opt to invest in their own capacities to develop and maintain comprehensive and multi-channel oriented data-driven targeting, with the latter being incomparably more expensive.
It is still not clear how efficient microtargeting is or may be as it is a novelty, not even being tested in some countries. What is clear is that data-driven political targeting is undoubtedly on its rise. As it was emphasised during the Workshop on Social media and data driven targeting in election campaigns, held in Perugia 16-17 February 2018, scholars, regulators, policy makers and civil society organisations need to join forces in monitoring this new practice and in ensuring that it does not undermine the legitimacy of the electoral process. Especially highlighted was the request for a comprehensive and a comparative understanding of this phenomenon, as different political and electoral systems largely affect applicability and efficiency of data-driven political microtargeting. For example, unlike the presidential system in the US, most political systems in the EU member states are parliamentary, and there is also a significant difference in relation to whether the electoral system is proportional, winner-takes-all, or a variation thereof. Political and electoral systems, as well as political culture, shape also the relationship between voters and candidates, which, within the framework of existing regulation, may also affect campaign budgets (donations). Then, but not least important, are the differences in the legal frameworks of the EU and the US, in particular with regard to data protection and privacy laws, which are more strict and comprehensive in Europe.
These differences affect political microtargeting. Kruchinski and Haller (2017) explored implementation of data-driven political microtargeting in Germany during the 2016 Rhineland-Palatinate state parliament elections, and clearly showed specific constraints related to electoral and party system, legal framework, political and party culture, budgetary restraints, and individual capacities of campaign managers. For example, as highlighted by Kruchinski and Haller (2017), Germany is covered by very strong data protection rules that limit personal data gathering only to party members or persons who have regular contact with the party, requiring the explicit consent from the person in question. The rules further forbid the matching of individual data with other datasets, to re-purpose it or to store for long-term. Then, there is an issue of financing as political parties in the EU member states have significantly lower budgets than those in the US (Kruchinski and Haller 2017, Zuiderveen Borgesius et al. 2018). All this affects the application of political microtargeting in Europe, as this practice is highly dependent on detailed and reliable datasets, and it requires a lot of resources and knowledge investments to obtain, assess and effectively use the data.
So far, both the promises and threats of online microtargeting to democracy have been observed (Zuiderveen Borgesius et al. 2018): On one hand, this individualised political communication has the potential to “diversify political campaigns” by discussing more issues, and to “increase political participation and therefore strengthen democracy”: by mobilising citizens in times of elections – in particular ones who opted out of traditional media exposure, increasing their political knowledge, and helping them make more informed voting choices. On the other hand, Zuiderveen Borgesius et al. (2018) list threats, such as the invasion of privacy, data breaches, decrease of political knowledge and pluralism, manipulation and exclusion of certain groups. There has been a concern that political microtargeting may enforce filter bubbles and political polarisation by exposing people only to information that corresponds to their expressed preferences.
In his essay from the 1963, Stanley Kelley Jr. explained the role of mass media in the electoral process: media set the agenda by prioritising issues and selectively giving time and space for candidates; they frame their reporting within a certain field of meaning and considering the characteristics of different types of media; they themselves act as opinion makers by endorsing certain candidates; and they sell time and space for political advertising within the existing legal framework. When Kelley Jr. was writing this essay he probably didn’t have in mind social media becoming amongst the primary sources of news and other information for an increasing number of people, especially youth (see Newman et al. 2017). However, his explanation could easily be extended to digital platforms that are not fully recognised as media but share many of the news media functions: filtering and prioritising certain content over others; framing the content within a certain field of meaning; and selling the time and space for political advertising. So far, these platforms do not endorse candidates, at least not as explicitly as traditional media do, but they provide other elements that affect modern political campaigns: including data and data-driven microtargeting services. News media have long been recognised as having a crucial role during elections. Free and fair elections rely heavily on free and fair media that, on one side, serve as wide-reaching platforms for political actors to deliver their messages, and, on the other, should help voters to make informed choices. So far, across the EU, electoral reporting and political advertising during electoral campaigns are mainly regulated for audiovisual media. Facebook, as one of the main sources and channels for political microtargeting, is not recognised as a media company, hence is not subject to legal, regulatory or policy requirements that are established for traditional media.
Adali, S. & Golbeck, J. (2012). Predicting Personality with Social Behavior. Proceedings of the 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012), p.302-309, August 26-29, 2012: https://pdfs.semanticscholar.org/84ad/47e877a2d90f2dca4476ebe4d92689d04381.pdf
Bodó, B. & Helberger, N. & de Vreese, C. (2017). Political micro-targeting: a Manchurian candidate or just a dark horse?. Internet Policy Review, 6(4). DOI: 10.14763/2017.4.776
Bozdag, E., van den Hoven, J. (2015). Breaking the filter bubble: democracy and design. Ethics and Information Technology, 17(4): https://link.springer.com/article/10.1007%2Fs10676-015-9380-y
Caplan, R. & Boyd, D., (2016). Who Controls the Public Sphere in an Era of Algorithms?. New York: Data & Society: https://datasociety.net/pubs/ap/MediationAutomationPower_2016.pdf
Kreiss, D. (2012). Yes We Can (Profile You), A Brief Primer on Campaigns and Political Data, Stanford Law Review: https://www.stanfordlawreview.org/online/privacy-paradox-yes-we-can-profile-you/
Kruschinski, S. & Haller, A. (2017). Restrictions on data-driven political micro-targeting in Germany. Internet Policy Review, 6(4). DOI: 10.14763/2017.4.780
Newman, N. et al. (2017). Reuters Institute Digital News Report 2017. Available at: https://reutersinstitute.politics.ox.ac.uk/sites/default/files/Digital%20News%20Report%202017%20web_0.pdf
Piccio, D. R. (2016). The state of political finance regulations in Western Europe. Stockholm: International IDEA: https://www.idea.int/sites/default/files/publications/the-state-of-political-finance-regulations-in-western-europe.pdf
Schipper, B. & Woo, H. Y. (2017). Political Awareness, Microtargeting of Voters, and Negative Electoral Campaigning. Available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2039122 or https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2039122
Yang, H. & Li, Y. (2013). Identifying User Needs from Social Media. IBM Research Report: https://domino.research.ibm.com/library/cyberdig.nsf/papers/28778C32C50DDB9185257BFA005AF49B/$File/rj10513.pdf
Zuiderveen Borgesius, F.J. et al. (2018). Online Political Microtargeting: Promises and Threats for Democracy. Utrecht Law Review, 14(1): 82–96.