We Say We Want Privacy Online, But Our Actions Say Otherwise


In polls and surveys, consumers indicate profound and increasing concern for their privacy. Yet from the posting of suggestive photographs on social networking sites to the impulsive broadcasting of illicit activities on Twitter, consumers’ behavior often suggests a remarkable lack of discretion. This “privacy paradox” has also been documented empirically — in various lab settings, people who indicate serious privacy concern nevertheless reveal intimate details of their lives for trivial rewards. And it is also evident in commerce — when asked, most consumers say they reject behavioral targeting gleaned from online behavior by companies, yet research suggests that the tactic can be highly effective.

As consumers spend an increasing amount of their time in cyberspace, companies are developing techniques to maximize the marketing capabilities afforded by cookies and other online surveillance tools. Yet companies first need to understand how consumers think about their privacy online, which isn’t entirely

. In an upcoming chapter of the Cambridge Handbook of Consumer Psychology, I draw on insights from behavioral psychology and the social sciences to explain the prevalence of the privacy paradox; I have summarized the reasons for consumers’ inconsistent views about privacy below:

Privacy is a faceless issue. People’s behaviors are much more strongly affected by specific and concrete issues. Privacy, on the other hand, is an intangible, hard-to-quantify concept and is therefore often not on the forefront of people’s minds. That may explain why covert tracking of online behavior fails to rouse concerns. It also explains how, when it comes to targeted advertisements, overly personalized advertisements can backfire because they bring privacy concerns to the fore. For example, the retail chain Target endured a public relations nightmare when it marketed diapers to a teen who the company (correctly) inferred to be pregnant due to her shopping patterns.

Sharing feels good. A wealth of research has documented the benefits of confiding in others; recent neuroscientific research suggests that self-disclosure is intrinsically rewarding. The internet offers endless opportunities to seek this reward. Users of Facebook, for example, are perpetually posed the question “What’s on your mind?”

Websites use defaults. People generally don’t opt-out of preset options. Privacy policies generally go unread, and for good reason. One study estimated it would take Americans 54 billion hours annually to read the privacy policy of each new website they visit.

Privacy promises can backfire. Some websites have learned that heavy-handed confidentiality assurances can cause people to “clam up” — they can serve as a cue that triggers privacy concerns, resulting in decreased disclosure in the face of increased protection.

Insight Center

People adapt to deteriorating privacy controls. Consumers’ willingness to part with personal data is impacted by changes in, rather than absolute levels of, the protectiveness of privacy policies. For example, in 2006, Facebook launched the “News Feed” feature, a running list of activities of users’ friends. By making public previously searchable (but obscure) information, News Feed generated backlash. But the outrage waned, and now News Feed is a central feature of Facebook.

People follow the herd. Economists and psychologists have long demonstrated how people’s behavior conforms to that of others. Such social norms affect online disclosure. In one study, my colleagues and I found that subjects were more likely to disclose that they engaged in sensitive behaviors when they were given information that others had engaged in those behaviors before them.

People disclose more to computers than each other. There’s a reason Google tells you that it’s only bots and algorithms that read your emails, not human beings. I recently conducted an experiment in which half the participants were told that engineers were looking through their emails to generate targeted advertisements and others were told computer programs were scanning their emails for the same purpose. As predicted, the human scanning was deemed to be more intrusive.

People don’t realize the implications of what they are revealing. Small but frequent requests for personal information give rise to isolation errors—the failure to appreciate the broader impact of one’s choices. For example, social security numbers can be predicted by combining a person’s data and place of birth (both commonly divulged on Facebook) with an algorithm; medical records can be tracked simply using a person’s date of birth and zip code. Shopping habits can reveal major life changes such as pregnancies and bereavement.

People don’t project far into the future. Projection bias refers to the misguided belief that future tastes will resemble current ones. This bias suggests that people fail to appreciate that how they value their privacy is likely to change — hence the prevalence of college students who post sensitive personal information only to regret it later.

People underestimate the threat of privacy violations. Research shows that humans are inherently optimistic about their likelihood of engaging in positive behaviors such as donating to charity, exercising, or losing weight. Over optimism extends into cyberspace. In one large survey, 56% of respondents were overly optimistic about their likelihood of avoiding identity theft.

The internet itself is a paradox of sorts. Seemingly abundant with ephemera, its content is actually permanent and searchable, and disclosures are forever catalogued in cyberspace (and, in the case of tweets, also in the Library of Congress!). Given our apathetic, often conflicting attitudes toward privacy, it is unlikely that consumer demand for privacy protection will force companies to instate it. Online companies have strong interests in making information open and easily accessible. Ultimately, government intervention may be necessary to regulate and protect consumer privacy in the face of the privacy paradox.