Quantifying Data-Driven Campaigning Across Sponsors and Platforms

Abstract

Although modern data-driven campaigning (DDC) is not entirely new, scholars have typically relied on reports and interviewers of practitioners to understand its use. However, the advent of public ad libraries from Meta and Google provides an opportunity to measure the scope and variation in DDC practice in advertising across different types of sponsors and within sponsors across platforms. Using textual and audiovisual processing, we create a database of ads from the 2022 U.S. elections. These data allow us to create an index that quantifies the extent of DDC at the level of the sponsor and platform. This index takes into account both the number of unique creatives placed and the similarity across those creatives. In addition, we explore the impact of sponsor resources, the office being sought, the competitiveness of the race, and the goal of the ad on the measure of DDC sophistication. Ultimately, our research establishes a measurement strategy for DDC that can be applied across ad sponsors, campaigns, parties and event countries. Understanding the extent of DDC is vital for policy discussions surrounding the regulation of microtargeting and data privacy.