Last week, Facebook landed in hot water – yet again. This time, it was because of the revelation that two years ago, some data scientists at Facebook ran an experiment to gauge whether users’ emotional state (as measured by status updates) changed depending on the content that appeared in their News Feeds. The news kicked off the very sort of criticism we’ve come to expect, with the usual suspects lining up to decry the social network’s psychological experimentation. It probably didn’t help matters that this news followed Facebook’s announcement last week that it would start using users’ web browsing data for ad targeting. Much grumbling about privacy concerns ensued.
Why does cycle keep repeating itself? It’s mostly because of Facebook’s scale and reach – the sheer ubiquity of the service means that virtually every person on the web, whether a Facebook user or not, has some exposure to it along the way. The company is relevant in some way to almost everyone, and their business model puts them on the front lines of the modern debate over online privacy.
Discussions of privacy are everywhere you look now – and they are usually very stale, the debating points having changed little over time. Those advocating maximalist individual privacy are unlikely to cede any points to those arguing for greater openness, and vice-versa. But while this debate rages with no clear resolution in sight, consumers continue to make choices that change the terms of the debate. In fact, the whole question of how far individual privacy reaches online is becoming increasingly irrelevant.
The world moves on
In a lot of ways, the fundamental issues surrounding the debate over online privacy were settled years ago. The bargain is this: consumers receive free, high-quality products and services in exchange for seeing targeted advertising. It is one of the oldest business models around made new again by the addition of better methods for segmenting and targeting consumers and the opportunity to monetize huge volumes of data about users. Consumers around the world have flocked to the services supported by this model: everything from Google to Facebook and Twitter and Pandora. As user bases have steadily grown, the majority of consumers have voted with their feet, choosing free ad-supported experiences that monetize user data over paid membership-based, tracking-free ones wherever the choice is offered.
I’ve had an ongoing dialogue with Aurelie Pols from Mind Your Privacy on this topic for quite a while. She argues, as have others, that users of these online services should be given a choice of paid usage options that offer an ad-free experience, but also opt the user out of sharing personal data.
The problem with that model was well summarized recently by Ben Thompson. In his recent post “Privacy is Dead,” he shows that “Advertising is lucrative, and free is often imperative:”
“There are other services that can’t even realistically choose between advertising and member-supported. Facebook is a great example: the utility of Facebook is directly correlated with how many people you know who are also using Facebook, and the only way to maximize that number is to make the service free, supported by advertising. Google is in a similar boat: the efficacy of search is in many ways tied to how many people are using search.”
In other words, where network effects are key to the value the service provides, “free” is not just the best model – it is the only feasible one. That leaves aside all of the practical problems: what do you do for customers without a credit card? A membership-based social network would essentially disqualify most of the developing world from entry.
Where network effects are not crucial, and customers pay for a better-defined good or service, a membership model can make more sense. Spotify has seen tremendous success by offering a differentiated paid product against a free version; Google Apps is another example, not to mention the galaxy of SaaS products for businesses that exist today. The journalism industry is now trying desperately to convince consumers that they can provide a paid product that’s demonstrably superior to free news. (For now, I’ve found my New York Times digital edition subscription to be worth it.)
Complaints, but low abandonment
And yet, we know that the vast majority of users do not abandon these services. This seems to be the best demonstration that users accept the tradeoff inherent in the huge value that such free services offer them. If the prospect of living without the conveniences of Google and social media makes you uncomfortable, then that actually suggests you agree. (Here’s a link to my presentation at South by Southwest this year on this very topic.)
But isn’t there a lot of public concern about online privacy?
It seems that every week, a new survey is released confirming the conventional wisdom that concern for privacy is front-and-center among consumers. Doesn’t this belie the argument that consumers have mostly accepted the proposition of personal data for free services?
I actually think it doesn’t. We all know that surveys can be highly unreliable measures of public sentiment, and strongly sensitive to subtle changes in wording. I have personally read very few surveys touching on privacy that used neutral language, likely because most organizations doing the surveying are issue-based. Moreover, a big problem with consumer surveys on this topic is that most consumers simply have no knowledge of the subject – either about what benefits they receive for “paying” with their personal data, nor how that data is actually used for marketing purposes. Indeed, many surveys will show overwhelming consumer opposition to using any personal data for “marketing purposes” – even though that practice has been around for almost a century.
The results of polling like this reminds me of the research showing that most Americans believe 28% of our government’s budget is spent on foreign aid (it’s actually well less than 1%), and a third believes Iraq actually did possess weapons of mass destruction when the U.S. invaded (no evidence was ever found). The results suggest more about public confusion and lack of knowledge than to any realistic policy choices.
What accounts for the high levels of reported concern, then? There is assuredly some normal anxiety created by shifting cultural norms driven by technological change, but I think that much of the public concern has been ginned up by irresponsible media coverage. “Journalism” like the recent 60 Minutes report on data brokers is a perfect example of straight-faced fear mongering among a mystified public, and it’s far from unique. Particularly in conjunction with utterly unrelated stories like the NSA surveillance leaks, the media has created a moral panic around “privacy” – and left the definition of what that means to the viewer.
What does “privacy” mean, anyway?
The problem with “privacy” as a topic for concern is that, like most personal values, it has no specific definition. My concept of privacy is different from yours, and it’s very difficult to build any kind of public policy based on such a shifting foundation. The European Court of Justice recently tried with its “right to be forgotten” legislation, which by most accounts has led to a complete mess. When each individual’s perception of “privacy” is a law unto itself, the result is a dramatically inferior experience for all.
In our shared digital future, personal data is going to be shared, likely indefinitely – and we’re substantially there already. While that will lead to changes in how people choose to share and disclose information about themselves, those individual decisions will be hard, if not impossible, to meaningfully regulate. As a matter of policy, we need to move beyond “privacy” and towards a focus on the applications of consumer data by the companies that own it.
A better approach would be talking about the proper and improper uses of data which involve actual consumer harms. It’s very difficult to quantify any actual consumer harm related to ad targeting, for example – but using credit scores to influence employment decisions, on the other hand, has a much clearer personal interest at stake. Using consumer data to influence price discrimination is probably inevitable (indeed, it’s been happening for decades), but there may be limits that it would make sense to set.
As more sophisticated uses of analytics with customer data have increasingly become competitive advantages, it’s unlikely that companies’ data uses themselves can be removed from the black boxes where they sit today. But the same situation exists for financial data, and to provide necessary protections, both business and government agreed to auditing standards that are required by regulation but enforced by independent bodies. Auditors provide public assurance of quality and above-board uses of company funds, which makes reliable investing by the public possible. It’s not always perfect, but the systems in place generally work exceptionally well. Such a system could be highly applicable to uses of consumer data in the future.
In any case, it’s time for the “privacy” debate to move on. Users have spoken, and they have chosen a techno-social landscape dominated by free online products and services paid for by an advertising model that requires personal and behavioral data to operate. That model is not going to change – now we’re just waiting for the public to catch up.