Introduction

Over the past 20 years, the collection and monetization of internet users’ personal and behavioral data for the purpose of delivering targeted advertising has emerged as a ubiquitous business model for the internet economy. According to projections made in 2019 by eMarketer—a market research company that focuses on digital marketing, media, and commerce—U.S. advertisers would spend more than $129 billion on digital advertising during the year, whereas they would spend only $109 billion on more traditional advertising methods such as television, radio, and newspapers.¹

The digital advertising industry has grown from an online version of the traditional advertising system, featuring advertising agencies and publishers, to a data-driven environment in which internet platforms such as Google and Facebook take center stage. Today, the two companies account for approximately 60 percent of the U.S. digital advertising market, with the next closest competitor, Amazon, only accounting for 8.8 percent of the industry market share.² On a global level, Google accounts for approximately $103.73 billion in net digital ad revenues, followed by Facebook with $67.37 billion, Alibaba with $29.20 billion, and Amazon with $14.03 billion.³

One of the core components of the success of the online advertising industry has been the introduction of new targeting tools that have enabled advertisers to segment and select their audiences along very specific lines. These targeting tools categorize consumers using a range of data points, which can include demographic characteristics, behavioral information, and personally identifiable information (PII).

Many companies assert that advertising can enhance the lives of their consumers, as it can connect users with products or services that they may find useful and may not otherwise be aware of. Although a user may find a certain advertised product or service relevant, it is important to recognize that the main goal of an advertisement is to drive purchases and revenue, or to elicit changes in behavior or beliefs. In addition, by delivering useful content to users, platforms aim to maximize the amount of time users spend on their services, which in turn increases the number of ads users will view and the amount of revenue the platforms will earn.⁴

Over the past decade, the digital advertising industry has increasingly adopted automated tools to streamline the targeting and delivery of advertisements. The goal of this targeting is to reach customers who are most likely to be interested in particular products and services.⁵ However, in many cases this targeting has instead exacerbated discriminatory and harmful outcomes. It is important to recognize that ad targeting and delivery practices inherently involve “discriminating” among users as they require advertisers to delineate which categories of users they would like to target; therefore their reach “discriminates” in favor of reaching certain users and against reaching others.

Not all of these delineations are based on sensitive or protected categories, nor do all of them result in harmful outcomes. However, some distinctions can generate particularly damaging outcomes,⁶ especially where they involve discrimination against protected classes or other sensitive categories. This can harm protected classes and exacerbate hidden societal biases. For example, ad targeting and delivery systems could be calibrated and optimized so that users of a certain gender or race do not receive ads for certain employment opportunities. As a result, these users are deprived of the opportunity to apply for these jobs. Given that the digital advertising ecosystem features an array of ads for services such as employment, financial services, and housing, these online discriminatory outcomes often have very real offline impacts.

In addition, the targeted advertising ecosystem can be (and has been) used to spread propaganda and disinformation, enable media manipulation, and even incite genocide.⁷ These destructive outcomes have heightened concerns regarding fairness, accountability, and transparency around algorithmic decision-making. In addition, some researchers have discussed the harms posed by the advertising business model in terms of human rights risks.⁸ Because of these concerns, some researchers, scholars, and members of civil society believe that the targeted advertising business model as a whole must be overhauled in order to enact meaningful change. This is an ongoing debate that will develop as internet platforms further expand their ad targeting and delivery practices.⁹ While eliminating this business model would solve many problems described in this report, the advertising business model is likely to last for the immediate future. This report explores actions that could help to mitigate some of these harmful effects.

This report is the third in a series of four reports that explore how automated tools are used by major technology companies to shape the content we see and engage with online. It focuses on the use of automated tools to target and deliver ads to internet users and relies on case studies of three internet platforms—Google, Facebook, and LinkedIn—to highlight the different ways algorithmic tools can be deployed by technology companies to enable ad targeting and delivery. These case studies will also highlight the challenges associated with these practices. This report also offers recommendations on how internet platforms, civil society, and researchers can promote greater fairness, accountability, and transparency around these algorithmic decision-making practices. This report also provides recommendations for policymakers in this regard. However, because the First Amendment limits the extent to which the U.S. government can direct how internet platforms decide what content to permit on their sites, this report provides only limited recommendations for action by policymakers.

Editorial disclosure: This report discusses policies by Google, Facebook, and LinkedIn. Google, Facebook, and Microsoft (owner of LinkedIn) are funders of work at New America, and LinkedIn's co-founder is on New America's board of directors, but none of these companies or individuals contributed funds directly to the research or writing of this report. New America is guided by the principles of full transparency, independence, and accessibility in all its activities and partnerships. New America does not engage in research or educational activities directed or influenced in any way by financial supporters. View our full list of donors at www.newamerica.org/our-funding.