Identifying User Needs for Advertising Controls on Facebook

Abstract

We conducted an online survey and remote usability study to explore user needs related to advertising controls on Facebook and determine how well existing controls align with these needs. Our survey results highlight a range of user objectives related to controlling Facebook ads, including being able to select what ad topics are shown or what personal information is used in ad targeting. Some objectives are achievable with Facebook's existing controls, but participants seemed to be unaware of them, suggesting issues of discoverability. In our remote usability study, participants noted areas in which the usability of Facebook's advertising controls could be improved, including the location, layout, and explanation of controls. Additionally, we found that users could be categorized into four groups based on their privacy concerns related to Facebook's data collection practices, objectives for controlling their ad experience, and willingness to engage with advertising controls. Our findings provide a set of user requirements for advertising controls, applicable to Facebook as well as other platforms, that would better align such controls with users' needs and expectations.

Computers as Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially

“…In considering these ambiguities, we will use the term 'dark patterns' in a broad sense, regardless of intention, to refer to design patterns that involve inducing certain emotions (e.g., guilt, fear, shame) or misleading mental models (e.g., a misconception of the system or their actions) to manipulate user behaviour towards certain ends-involving mechanisms or goals that the user is unaware of or did not consent to. 6 As noted earlier, the vast majority of literature on dark patterns has focused on how information is phrased or presented in graphic user interfaces (GUIs), most prominently cookie consent banners [5,6,19], ads on online platforms [30,66], and e-commerce websites [44,47,58]. Recently, a handful of papers have started considering dark patterns in the ways that other modalities interact with people, i.e., embodied interactions with robots [33,57], auditory interactions with voice user interfaces (VUIs) [52], and proxemic interactions with movable platforms [28].…”

Facebook Data Shield: Increasing Awareness and Control over Data used by Newsfeed-Generating Algorithms

“…uses data to profile users and serve targeted ads based on such profiles [14]. The data they collect usually consist of both "voluntarily" offered information when signing up, a.o.…”

“…It has been proven that they can lead to discriminating users [4], spreading fake news [3,21] and creating filter bubbles [12]. Previous studies have shown that people are not always aware of these practices and when they discover them, they often feel uncomfortable with the data that is being collected [6,7,14]. However, a segment of users also considers these targeted ads useful [14,22].…”

“…Previous studies have shown that people are not always aware of these practices and when they discover them, they often feel uncomfortable with the data that is being collected [6,7,14]. However, a segment of users also considers these targeted ads useful [14,22].…”

“…Users can exert some control by changing their privacy settings or how their newsfeed is generated. However, these controls are hidden inside nested menus, which makes them not only hard to find but also less known to users [13][14][15]. For example, Facebook currently allows users to choose how their newsfeed is generated, but this is a hidden setting that needs to be activated each time the platform is accessed.…”