'Fake news', 'online disinformation' or 'computational propaganda' – there are a variety terms to describe coordinated online campaigns to sway elections. Public interest in these practices spiked after the 2016 US presidential elections, when it was revealed that the consultancy Cambridge Analytica obtained data of more than 70.6 million US citizens (and another 17 million users worldwide). The consultancy claimed to have developed sophisticated profiles to both mobilise Trump supporters and suppress turnout for Clinton using a technique borrowed from behavioural advertisement called micro-targeting. Pundits were quick to attribute Trump's victory to this disinformation campaign - despite some experts expressing doubts over its true effectiveness.
What is micro-targeting?
Online disinformation campaigns rely on a variety of different practices. These include 'astroturfing' and 'trolling', where large numbers of online profiles share hyperpartisan content. Yet, the practice that has featured most prominently in recent public debates is that of micro-targeting.
Micro-targeting is not a new phenomenon. Private companies have always valued detailed information on their customers. The 'connectedness' of modern life gives these actors access to ever more personal data. Tools like Facebook Pixel or Google AdSense track user online activity and develop comprehensive databases. These are then sold to advertisers that seek to target specific audiences. If advertisers have a seed list of 'customers', platforms can use this to look for 'similar' audiences. These seed lists can be party membership registries, public electoral rolls or even hacked mailing lists. Private companies use these behavioural insights to find new customers for their products. Political actors use online (and offline) constituent data to tweak their campaign messages.
In short, micro-targeting is the practice of inferring individual preferences from a variety of data sources. These insights are then used to personalise campaign messages and ads.
Micro-targeting itself is not necessarily at odds with fair political communication in democracies. It may allow campaigners to inform individuals that are not reached through offline media. Obama’s 2012 campaign was pioneering social media as a tool to this end. Yet, micro-targeting can also be a tool to spread highly-personalised disinformation. When the public is misinformed, it undermines the legitimacy of election results.
How effective is micro-targeting?
There is still little empirical evidence to show how the exposure to online political advertisement affects personal political convictions. After all, it’s easier to make a user buy a product than it is to convince them to support a certain political agenda. Additionally, marketing firms have a clear incentive to make their services appear as effective as possible. Voters aren't as gullible as some political marketeers would like them to be. Most citizens in developed democracies have access to a variety of informational resources, both on and offline. The reality for campaigners is that it is difficult to change people's minds, particularly when it comes to deeply-held political persuasions. This does not mean that political micro-targeting is useless. Rather, it can be crucial tool in the late phase of a campaign: that of 'getting out the vote'. The goal here is to mobilise their own supporters and curb turnout for political opponents. The goal of micro-targeting here is to convince those with weak partisan identities that their vote matters - or is wasted anyway.
This is why the effectiveness of micro-targeting is dependent on the political context. As this article illustrates, one feature that is of particular relevance is the national electoral system.
Sparked your interest? Want to stay informed amidst the daily onslaught of online disinformation? Subscribe to our weekly newsletter!