Ad Delivery Algorithms: The Hidden Arbiters of Political Messaging (1912.04255v3)

Abstract: Political campaigns are increasingly turning to digital advertising to reach voters. These platforms empower advertisers to target messages to platform users with great precision, including through inferences about those users' political affiliations. However, prior work has shown that platforms' ad delivery algorithms can selectively deliver ads within these target audiences in ways that can lead to demographic skews along race and gender lines, often without an advertiser's knowledge. In this study, we investigate the impact of Facebook's ad delivery algorithms on political ads. We run a series of political ads on Facebook and measure how Facebook delivers those ads to different groups, depending on an ad's content (e.g., the political viewpoint featured) and targeting criteria. We find that Facebook's ad delivery algorithms effectively differentiate the price of reaching a user based on their inferred political alignment with the advertised content, inhibiting political campaigns' ability to reach voters with diverse political views. This effect is most acute when advertisers use small budgets, as Facebook's delivery algorithm tends to preferentially deliver to the users who are, according to Facebook's estimation, most relevant. Our findings point to advertising platforms' potential role in political polarization and creating informational filter bubbles. Furthermore, some large ad platforms have recently changed their policies to restrict the targeting tools they offer to political campaigns; our findings show that such reforms will be insufficient if the goal is to ensure that political ads are shown to users of diverse political views. Our findings add urgency to calls for more meaningful public transparency into the political advertising ecosystem.

• The paper demonstrates that Facebook’s political ad delivery algorithms skew distribution based on ad content and inferred political leanings.
• It reveals that ads reaching non-aligned audiences incur higher costs, complicating efforts to engage diverse voter groups.
• The study shows that user engagement further amplifies delivery biases, highlighting the need for transparency and regulatory oversight.

Analyzing Political Ad Delivery on Facebook

The paper "Ad Delivery Algorithms" by researchers Muhammad Ali, Piotr Sapiezynski, Aleksandra Korolova, Alan Mislove, and Aaron Rieke provides a rigorous analysis of Facebook's ad delivery algorithms and their impact on the political advertising landscape. The paper explores the mechanics of ad delivery, exposing how these mechanisms can inadvertently contribute to increased political polarization and informational echo chambers.

Summary

The research scrutinizes Facebook's role in political ad delivery, during which campaigns aim to reach diverse voter demographics. Instead, the platform's algorithms may asymmetrically skew ad distribution along political lines, affecting both the visibility and price of ads based on perceived political affinities of users. The paper effectively demonstrates that the content of an ad significantly influences which audience segments see it, irrespective of the advertisers' intended targeting.

To explore this, the authors conducted methodical experiments, running various political ads with identical targeting parameters on Facebook. They examined how these ads were distributed among audiences with differing political involvements, including registered voters and political donors, contrasting delivery patterns across different demographic subsets. Their experiments revealed that Facebook's ad delivery algorithm not only skews toward affinity audiences but also imposes higher costs on ads reaching non-aligned political groups.

Key Findings

1. Ad Content Influences Delivery: Facebook's ad delivery algorithms substantially impact ad distribution based purely on the political content of the ads. Ads for Democratic and Republican candidates were distributed differently even when targeted at the same audience, reflecting a systematic bias based on inferred political alignments.
2. Increased Costs for Non-Aligned Audiences: Ads targeted at audiences with differing political inclinations from the ad content incurred higher costs. For example, a Democratic ad targeting a conservative audience cost more per 1,000 impressions than a Republican ad targeting the same conservative set.
3. User Engagement Exacerbates Skews: While initial skews were algorithmically determined, user engagement metrics such as likes and shares further amplified the differences in ad distribution, highlighting a feedback loop where popular ads or those resonating with the inferred views received preferential treatment.
4. Implications on Political Campaigns: These biases signify that campaigns attempting to engage diverse political viewpoints may find it both harder and more expensive to achieve their outreach goals, complicating democratic discourse on digital platforms.

Implications and Future Research

The research highlights pressing concerns about Facebook's influence over political communication, emphasizing that ad delivery can inadvertently isolate voters from diverse political expressions. The findings imply an urgent need for transparency and possible regulation of digital ad platforms to ensure equitable political discourse.

Further, this paper builds a foundation for future explorations into algorithmic biases on social platforms, advocating for the development of external auditing mechanisms to scrutinize the complexities of ad dissemination algorithms. Future advancements might include exploring algorithm design changes to mitigate political skewness or the creation of neutral channels for political and civic communication.

Given its methodological rigor, this paper adds significant clarity to our understanding of digital advertising ecosystems and their broader societal ramifications. It raises critical awareness about the responsibilities of technology platforms in fostering a balanced political landscape in the ever-evolving field of digital communication.