Author : Eric Zeng
Publisher :
ISBN 13 :
Total Pages : 0 pages
Book Rating : 4.:/5 (136 download)
Book Synopsis Characterizing and Measuring "bad Ads" on the Web by : Eric Zeng
Download or read book Characterizing and Measuring "bad Ads" on the Web written by Eric Zeng and published by . This book was released on 2022 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Online advertising is a core part of the modern web; ads sustain websites that provide free content and services to consumers, and inform people about products that they may be interested in. However, the ubiquity of the online advertising ecosystem makes it a potent vector for abuse; malicious actors can use the infrastructure of ad networks to serve scams, malware, and other misleading or detrimental content to millions of users across millions of websites. Though the ad networks that provide this infrastructure make efforts to prevent inappropriate and harmful content from appearing on their platforms through content moderation, many kinds of deceptive and unpleasant ads regularly appear on people's screens. Due to the opacity of the online advertising ecosystem, it is challenging for external observers to assess the harms and scale of problematic online ads. This dissertation presents a systematic investigation of the nature and prevalence of problematic content in online advertising, or "bad ads", on the modern web, through four studies. First, this work investigates users' perceptions of online advertising, characterizing the reasons why people dislike (and like) ads, and identifying types of ad content which engender negative reactions. Second, this work quantitatively measures the phenomenon of "clickbait" advertising on news and media websites. Using data crawled from over 7000 news and media websites, this work finds that native advertising networks are strong drivers of problematic content such as content farms and advertorials, and are extremely common across a variety of news websites. Third, this work examines problematic content in online political advertising during the 2020 U.S. elections. In a longitudinal measurement study, this work finds evidence of multiple categories of deceptive political content in online ads, including misleading polls and petitions, political clickbait, and misleading political-themed product ads, and found that these ads were targeted at partisan news sources. Lastly, this work empirically measures the targeting of online ads more broadly, through a unique field study using data collected from 286 real users. This dataset provides measurements of the prevalence of different categories of ad content, how such categories are targeted across websites and demographic groups, the monetary value placed on users by advertisers. Together, these works provide a foundation for future regulation, policy, and research aiming to curb problematic content in online advertising, and improve the overall experience for users on the web.