Digital Communication and Elections: Online targeting of voters on social media

m.palese

In recent elections and in the EU referendum, concerns have been raised about the online targeting of voters on social media and the use of voter data. At a Constitution Unit seminar, Dr Martin Moore spoke about the shifting nature of online campaigning and examined its impact on the regulatory and legal landscape in the UK. In this blog post, Michela Palese summarises what he said.

Election campaigns have shifted significantly online in recent years. This reflects trends in news and media consumption. Nowadays, around 90% of British adults are online and 84% use social media, with approximately 30 million Britons using Facebook on a regular basis. Furthermore, around three quarters of internet users get their news online and four in ten use social media for news. It is therefore unsurprising that campaigning has shifted to the digital world, given the ease with which voters can be reached in a direct and highly personalised way.

Funding has shifted to digital as well. Campaigners admit that digital is where elections can be won and lost, and this is proven by the successful use they have made of social media. For example, Dominic Cummings stated that the Vote Leave campaign spent around 98% of its budget on digital campaigning. Jim Messina, a former political advisor to Barack Obama and founder of consulting firm The Messina Group, claimed that Facebook was ‘the crucial weapon’ in the Conservative Party’s successful general election campaign in 2015. In the last general election, by contrast, the Labour Party invested significant effort in social media advertising, with its videos being viewed by around ten million internet users.  

At a Constitution Unit seminar, Dr Martin Moore, Director of the Centre for the Study of Media, Communication, and Power at King’s College London, argued that the changing nature of campaigning has highlighted some of the shortcomings in the UK’s regulatory and legal landscape. In particular, Moore noted the concerns which have been raised about the lack of fairness and openness in how campaigns are conducted online. Campaigners have, in fact, much more leeway in what they can do in the digital realm than on print or broadcast platforms. In 2017, both the Information Commissioner’s Office and the Electoral Commission opened investigations into how funds were spent during the EU referendum and into the manner in which voters were targeted on social media.

What is digital campaigning?

While similar to traditional marketing communications, Moore showed how digital campaigning can create new opportunities for campaigners to target voters more precisely and with highly tailored messaging. In this context, Facebook acts as a bridge between the ‘real’ and the online worlds. By gathering users’ contact details, especially email addresses, campaigners can create their own ‘custom audience’ and direct specific advertisements to this target group. Working from this initial set, campaigners can build up their audience by reaching users with similar features and interests.

Moore explained how the process of digital campaigning can be broken down into six steps:

  1. Gather as much data as possible about potential voters. There are different regulations on what you can collect and aggregate. In the US, for example, there are very limited restrictions, which leads to a race to accumulate information. Data can also be gathered in creative ways. For example, the Vote Leave campaign ran a football competition on Facebook, where users could have won £50 million if they correctly predicted the result of all 51 matches in Euro 2016. Entrants had to supply certain personal data, such as email addresses, and this was used by the campaign to profile specific demographics of voters.
  2. Organise the data so as to identify where to target resources and prioritise the campaign.
  3. Analyse the data so as to identify who to target and with which communications in order to win a campaign.
  4. Test the campaign on an initial group. This allows campaigners to understand what motivates voters politically and what makes them ‘tick’ behaviourally.
  5. Deliver the campaign to an initial target audience.
  6. Iterate: Campaigners continually evolve their messaging on the basis of how people respond, how engaged they are and how they share communications. This is done until the most effective targeting mechanism is found.

Challenge #1 – Fairness

The current regulatory framework of political campaigning aims to ensure a level playing field by regulating what parties and campaigners can spend at the local and national levels, and in what period. Ensuring compliance with these regulations is the Electoral Commission’s responsibility. But according to Moore, the increase in digital campaigning has exposed some of the shortcomings in the current framework.

In fact, Facebook has made it possible for campaigners not only to identify who to target, but also where to target. Digital spend can thus be focused on advertising to swing voters in marginal constituencies. In their reporting to the Electoral Commission, parties and campaigners do not need to provide a breakdown of how much money was spent on digital campaigning or where they spent. This lack of clarity means that online spend can be allocated to national spend, even if used in a few key constituencies. This undermines current restrictions on spending at the local level.

Furthermore, Moore explained how the role of external agencies is changing. These organisations can collect, organise and analyse data off the books and outside of an election or referendum campaign. The extent of their involvement is also hard to verify. It is thus becoming more difficult to keep track of support and to scrutinise expenditure.

How and where money is spent has thus become increasingly hard to monitor. Data can be gathered by external agencies and for purposes other than an election or referendum campaign; it can be stored and analysed offshore. Spending on personal data and data expertise can be done outside the campaign period and can be taken care of by agencies; spending on delivery is between the tech platform and the campaigners, and has been opaque up to now.

Challenge #2 – Openness

Another concern raised by digital campaigning is the difficulty of verifying what advertisements are being disseminated and by whom. According to Moore, three aspects are particularly significant.

First, Facebook launched ‘dark posts’ in 2013 – advertisements which are only visible to Facebook, the advertiser running the posts, and the individual recipients. They were originally created for commercial purposes, as companies wished to test different advertising campaigns on customers without the advertisement appearing on the company’s Facebook page. They were then adapted for political use. Given such direct and hidden targeting, dark posts can include false claims, misinformation, inflammatory allegations, or conspiracy theories without anyone being able to monitor the information provided. Given heightened debate about the spread of ‘fake news’ on its platform, in late 2017 Facebook announced that it would introduce a ‘View Ads’ tab so as to make political posts more transparent and effectively end dark posts. This feature will allow users to view all the ads an advertiser is running on the platform.

Second, Facebook introduced the ‘instant articles’ feature in 2015. Originally, this allowed selected news organisations to publish stories directly on the platform, as a way of keeping people on the site rather than referring them to an outside website. Soon after, this feature became available to everyone, meaning that any user could publish news stories to the Facebook News Feed. This made it easier for creators of fake news sites to attract visitors through Facebook and increase both their traffic and advertising revenue. Hyper-partisan publishers also began to take advantage of this feature to distort political communication by publishing false or misleading content.

Third, laws that exist offline to help with sourcing and verification of content currently do not apply online. For example, while imprint laws require all printed materials to expose who published them on behalf of whom, they currently only apply as guidelines to online communication. Political advertising on social media does not include any information of where it originated and who paid for it. Attribution is thus a significant problem.

Concluding Remarks

Moore concluded by remarking that there are lots of unknowns in the area of digital campaigning and there are no easy solutions to the threats to fairness and openness. Data now acts as an alternative currency in campaigning and the role of external agencies is increasing. But neither are accounted for in electoral law and have few precedents.

Social media is, by its nature, open to almost anyone and therefore equally open to misuse. But despite this high level of accessibility, it is for the most part also opaque to independent scrutiny. This opacity explains why to date it has been difficult to identify filter bubbles, micro-targeting, and dark posts, and thus determine how best to respond.

Moore noted that digital campaigners have considerable freedom within the current regulatory framework. If there are ways in which campaigners can use online platforms to win an election or referendum, they will do so even if this means bypassing existing laws and regulation, such as spending limits.

About the author

Michela Palese is a Research Assistant (McDougall Fellow) at the Constitution Unit.

One thought on “Digital Communication and Elections: Online targeting of voters on social media

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s