Hard to Discover, Harder to Use: The Widespread Failure of Ad Transparency Settings
Have you ever seen a strangely specific ad follow you around the internet and wondered why? You dive into your account settings, looking for a way to control what data is being used, only to find a confusing maze of menus, vague descriptions, and options that don’t seem to do what you want. Or maybe, you don’t even notice that your data is being tracked at all.

This lack of ad transparency in social media is at the heart of new research from Kevin Bryson, a fifth-year PhD student in SUPERgroup, supervised by Associate Professor Blase Ur. In their recent paper, Characterizing the Usability and Usefulness of U.S. Ad Transparency Systems, Bryson and his team found that the “ad transparency” tools provided by major online platforms are often hidden, inconsistent, and ultimately not beneficial to average users.
“For me, the really interesting thing was to focus on ‘what can my dad do?’” Bryson said. “Especially as it’s related to advertisement tracking on these platforms. That was the main focus: let’s look at the things that somebody could find in the app or on the website, and from that, what they can learn or control about their personalized advertising experience.”
Bryson broke the research down into two main phases. First, the team embarked on a qualitative analysis to understand the current most commonly used platforms. To do this, they began with the Tranco list of the top 150 most popular web domains, filtering it down to 23 platforms to perform their main analyses with.
“In existing research, there’s a lot of focus on really large sites, like Facebook, Twitter, and Google,” Bryson stated. “This is for good reason, but there are many other platforms that also contribute to this economy of tracking data. So it was important for us to look at a really broad range of different platforms, because those impact people too.”
Their analyses involved manually taking hundreds of screenshots of every settings page on both mobile (iOS and Android) and web. They then systematically coded these images to catalogue what types of data were mentioned (such as location data), whether data was from the platform itself or third parties, and what affordances were offered to users. This was complicated by interfaces that would often change without warning, forcing the team to constantly rescan and update their data every few weeks. 
“To answer the question of what my parents would be able to do with these interfaces, we conducted a user study,” Bryson said. “We recruited a bunch of people who use any one of eight different apps, a representative sample of the original 23. We chose them based on level of control and information, and asked whether users find those resources useful and usable. That will help us understand what an average person is able to do and access.”
In the user study, participants were guided through four specific tasks: navigating to the platform’s ad transparency system (ATS) landing page, exploring controls for third-party data usage, reviewing the interest topics the platform had inferred about them, and finding the explanation for why they were shown a particular ad. They found that across the 200 participants, many were not able to successfully find the landing page and that people had hundreds of questions that could not be answered with the existing tools, citing a lack of transparency and agency over their own data.
Bryson believes that this problem stems from a near-total lack of regulation in the US. While EU laws such as the General Data Protection Regulation and Digital Services Act mandate features like the “right to erasure” or ability to download data, specific control over user data access and transparency is still not under any legislation.
“The personal advertising space is largely a lawless land,” reflected Bryson. “There’s a real lack of transparency, mainly because it’s not mandated, and there’s no monetary incentive for them to do so.”
Looking ahead, Bryson suggests investigating new systems that give users more agency by letting them specify what ads they want to see, which could also create a monetary incentive for companies. Another path is policy, such as requiring a simple, one-button way to turn off personalized ads, similar to the Federal Trade Commission’s recent ruling on one-click subscription cancellations.
To learn more about SUPERgroup’s work, visit their lab website here.