Facebook and advertisers had a symbiotic relationship. The Cambridge Analytica saga exposed it in a nasty way after the 2016 US election following the . Western countries have regulated it to a large extent since then. However, the advertisers in the third world countries continue to reap the rewards of the business model of Facebook. This is how it works. The online targeted advertising campaign constitutes the primary revenue source of Facebook. The account registration and user engagement process enable Facebook to accumulate demographic, interest, and behavioral profiles of its users. The number of users is in the billions. Advertisers use such data to micro-target the users giving the advertiser the ability to choose the audience with very specific attributes for their ads.
Types of microtargeted ads
There are 3 types of microtargeted ads on Facebook. They are PII, look-alike audience, and attribute-based targeting. They have been used for different purposes such as lifestyle disease surveillance, public health awareness, to study gender bias as well as migration across countries, to infer the political leaning of news outlets, etc. Also, before launching a micro-targeted Facebook advertisement, the advertiser can get the approximate number of people likely to match the number of the target audience. Let’s touch upon each of them briefly to get a sense of what it entails.
The personally identifiable information, otherwise known as PII, is a form of targeting in which Facebook provides advertisers with personal information of the users such as name, email address, and phone number so that the ads can be directed to them. Cambridge Analytica harvested such data to the tune of 80 million users and psycho-profiled each user to target them with chewable divisive messages to influence the US election in 2016.
The look-alike audience target is the form in which advertisers create a list of users and then Facebook attempts to target a similar group created from their vast reservoir of data-trove. These are personal data entrusted in the hands of Facebook by their users at the time of creating the Facebook account as well as throughout their usage. And that includes posts, comments, their friends’ profiles, and their data as well as their browsing history.
The attribute-based targeting is a form in which the advertiser creates a target formula based on a range of elements including behavioral and interest attributes. Facebook then allows advertisers to combine multiple attributes and exclude or include users with each of those attributes as part of a formula. It also presents other attributes that target people with related aspects. A vast majority of high impact ads use this option.
Forms of abuse
A rich body of work has focused on understanding echo chambers, filter bubbles, ideological discourse, and polarization on Facebook through targeted advertising. Also, researches have been conducted as an attempt to understand the abuse on Facebook by misinformation campaigns along political lines.
Malicious advertisers leverage Facebook to target selective groups on different sides of the issues. They then deliberately script messages to stoke each side’s grievances and worsen social discord using socially divisive advertising. Studies reveal that the click rates of these malicious and divisive ads are higher than those of typical ads, meaning that these ads are extremely effective. Interestingly, Facebook not only injects misinformation in social systems through such malicious ad campaigns but also chooses vulnerable people as their target to disseminate fake news. Worse still, they give advertisers the ability to exclude people who may report such ads.
The studies conducted on Facebook ads have highlighted many forms of abuse of microtargeted advertising including exposing the private information of users thereby raising privacy concerns. Also, Facebook ads can be abused as a form of attack to create social discord. Although microtargeting can be very effective, it has a disruptive ability to create divisive ad campaigns on Facebook about controversial societal issues to incite social conflict. Therefore, it has come under scrutiny.
Advertisers take advantage of the Facebook suggestions tool to manipulate the targeted populations. While this method may be helpful in several ways, it needs redesigning. This will prevent a malicious advertiser from reaching vulnerable people. Hence, these days, after much scrutiny from the lawmakers, Facebook sometimes inspects ads manually before launching. That way ads do not divide or discriminate against some people. Additionally, Facebook flag those ads that experience high click-through rates for manual inspection. The way Facebook chooses which ads to manually inspect is based on their targeting model.
Ribeiro, Filipe & Saha, Koustuv & Babaei, Mahmoudreza & Henrique, Lucas & Messias, Johnnatan & Goga, Oana & Benevenuto, Fabrício & Gummadi, Krishna & Redmiles, Elissa. (2018). On Microtargeting Socially Divisive Ads: A Case Study of Russia-Linked Ad Campaigns on Facebook.