Surveillance Advertising places advertisements based on consumers’ characteristics – their demographics, their interests, and their past behavior over time and across platforms and devices.
Contextual Advertising places advertisements based on characteristics of the content a consumer is currently browsing – the subject matter of only that webpage or app. Contextual advertising in its pure form does not take the consumer characteristics into account or depend on that person’s past behavior.
It is important, however, to look closely at what a company is actually doing under the label “contextual advertising.” Not all contextual advertising systems are the same, and some so-called contextual advertising systems look more similar to surveillance advertising, taking into account consumer data from online behavior and outside sources.
Both surveillance and contextual advertising involve the same automated processes such as real-time bidding auctions. The difference is that placing surveillance ads requires broadcasting a consumer’s profile to potential advertisers, while placing contextual ads only requires broadcasting information about the contents of the webpage. Both types of advertising use machine-learning algorithms to make inferences. For surveillance ads, these inferences are about consumers, and for contextual ads, they’re about content.
Surveillance advertising is generally more expensive than contextual advertising because it requires companies to collect data and maintain profiles on every consumer to understand what ads might be relevant. That is not necessary for contextual ads. Much of the cost of surveillance advertising goes to the ad tech industry – the companies that conduct the tracking, create profiles of consumers, match consumers with ads based on their profiles, and place those ads where they’ll see them. Publishers – the businesses that operate the websites and apps where the ads appear and make money when consumers click on them – do not significantly benefit from surveillance advertisements. A 2019 study found that publishers only see a four percent increase in revenue from hosting targeted ads over non-targeted ones, or $0.00008 per ad. Both the New York Times and Dutch public broadcasting company NOP have actually seen ad revenues increase after they stopped accepting surveillance advertising.
A 2019 study found that demographics and interest categories used to target ads are often inaccurate across leading data brokers, resulting in low gains for targeted ads over random ad placement. Another study found contextual ads to be more cost-effective than surveillance advertising.
Privacy and Other Concerns
Surveillance advertising can perpetuate discrimination in housing, credit, employment, and other economic opportunities. It also hides personalized pricing from consumers, leaving them unaware that a company has charged them a different amount than others. Surveillance advertising is also sometimes used for promoting unhealthy products, encouraging gambling, and perpetrating fraud. The data that are fed into algorithms to profile consumers may be inaccurate, but even when they are correct, the fact is that surveillance advertising is unfair. It uses invisible and invasive techniques to manipulate consumers and robs them of real choice in the marketplace. Furthermore, the enormous stores of personal data collected for surveillance advertising put consumers at risk for exposure, identity theft, and more malicious tracking. It can also lead to erosion of their 4th Amendment rights, as government agencies can purchase data that otherwise requires a probable cause warrant to obtain.
Contextual advertising does not raise these concerns because it does not require tracking and storing consumer data, as long as it truly is contextual advertising and not a surveillance advertising hybrid.