Say you, like me, went to bed a little early last night. And when you woke up this morning, you decided to catch the episode of the Daily Show that you missed. So you pointed your browser over to thedailyshow.com, and there, as you expected, is John Oliver. But there's something else there too, at least if you're me: flashing deals for hotels in Annapolis, which just so happens to be where I've been planning a weekend away.
We all are familiar at this point with the targeted ads that follow us around the web, linked to our browsing history. In this case, Google (who served me this ad) only got it half right: I had already booked a place.
And yet, I am planning a trip to Annapolis, and Google "knows" this, and is using this information to try to sell me stuff, a practice commonly criticized as "creepy." But as philosopher Evan Selinger asserted in Slate last year, the word "creepy" isn't particularly illuminating. What, really, is wrong with ad tracking? Why does it bother us? What is the problem?
A new paper by professor Ryan Calo at the University of Washington goes the furthest I have seen in elucidating the potential harms of digital-ad targeting. And his argument basically boils down to this: This isn't about the sanctity of the individual or even, strictly speaking, about privacy. This is about protecting consumers from profit-seeking corporations, who are gaining an insurmountable edge in their efforts to get people to part with their money.
But those are my words. Here are Calo's:
The digitization of commerce dramatically alters the capacity of firms to influence consumers at a personal level. A specific set of emerging technologies and techniques will empower corporations to discover and exploit the limits of each, individual consumer's ability to pursue his or her own self-interest. Firms will increasingly be able to trigger irrationality or vulnerability in consumers -- leading to actual and perceived harms that challenge the limits of consumer protection law, but which regulators can scarcely ignore.
Calo is taking the long view here. Digital marketing techniques haven't quite gotten sophisticated enough to take advantage of a consumer's idiosyncratic irrationalities. Right now, he writes, digital advertising's main strategy is relevance: putting the relevant ad in front of the right person. But Calo foresees a much more personalized approach down the road -- not just the right good, but a customized pitch, delivered late at night, when the company knows you, particularly, have a tendency to make impulse purchases.
This, of course, is not all bad. Calo describes many potential upsides to a marketplace in which companies have much, much more information about their customers. "Firms," he writes, "have incentives to look for ways to exploit consumers, but they also have powerful incentives to look for ways to help and delight them." So perhaps website will appear in color palettes that are highly appealing to you, or a hotel you book at will know about your allergies and prepares your room with a hypoallergenic pillow, or, most obviously, Google will serve you an ad for just the right hotel, at just the right price, before you've already booked elsewhere. And consumers too are getting more information about the products they are considering buying: There's Amazon reviews, apps that scan bar codes for price comparisons, and sites like TripAdvisor and Yelp that will help you avoid rip-offs of all kinds.
But that said, Calo still sees an imbalance in how this will play out, and that's because consumers are not perfectly rational, as the field of behavioral economics has demonstrated over and over. This leaves them vulnerable to persuasion to make decisions that are counter to their own self-interest. Oftentimes, this has negative but ultimately small consequences: "Maybe a consumer pays a little extra for a product, for instance, or purchases an item on impulse," Calo writes.
But the possibilities for exploiting those vulnerabilities are amplified dramatically when bolstered by the kinds of data Google and other firms have access to. Marketing has always been about getting consumers to spend money, but Calo argues that data tracking enables a level of sophistication that is different in kind, not just degree. "Digital market manipulation combines, for the first time, a certain kind of personalization with the intense systemization made possible by mediated consumption," he explains.
This is not a problem when the interests of firms and consumers align, as in the above examples, but, as Calo writes, "it would be highly surprising were every use to which a company placed intimate knowledge of its consumer in fact a win-win." It's where interests diverge, and actual harms are incurred, that the trouble lies.
So where does that trouble lie? What are those actual harms? Calo outlines three distinct types of damages. The first are economic: market failures, not unlike others that the government has decided merit corrective regulatory measures in the past, such as the regulation of cigarette ads. But in the case of digital marketing, Calo says, the inefficiencies aren't going to be such clear cases. Rather, the failures will come in the form of consumers being systematically charged more than they would have been had less information about that particular consumer.
Sometimes, that will mean exploiting people who are not of a particular class, say upcharging men for flowers if a computer recognizes that that he's looking for flowers the day after his anniversary. But other times there could be troubling equity concerns. For example, Calo points to the work of NYU professor Oren Bar-Gill who has shown how companies can use complexity in credit-card contracts, mortgages, and cell-phone contracts to "hinder or distort competition and impose outsized burden on the least sophisticated consumers." Calo says such price-discrimination tactics, applied en masse online, could "lead to regressive distribution effects," also known as preying on the vulnerable.
But perhaps you think those inefficiencies will be balanced out at the level of the aggregate market. Calo says there is still reason to be concerned at the level of the individual. He writes, "Even if we do not believe the economic harm story at the level of the market, the mechanism of harm at the level of the consumer is rather clear: The consumer is shedding information that, without her knowledge or against her wishes, will be used to charge her as much as possible, to sell her a product or service she does not need or needs less of, or to convince her in a way that she would find objectionable were she aware of the practice." There may be nothing particularly embarrassing or personal about my vulnerabilities as a consumer, but I do not especially want to share them with companies so that I can be manipulated for their financial gain. For Calo, that discomfort, the feeling I experience knowing that my vulnerabilities are being tracked in order to be used, is a violation of my privacy, the second area of harm Calo sees.
This is closely related to Calo's third area of concern: autonomy, which in the context of the consumer means, he says, "the absence of vulnerability, i.e., the capacity to act upon the market in our self-interest." When corporations purposely seek out a consumer's vulnerabilities and use them to direct her dollars back to them, that is a violation of that person's autonomy.
(It should be clear that Calo here is only looking at corporate data tracking techniques for the purpose of selling goods and services. This, while troubling for the reasons detailed above, is "not tantamount to massive surveillance by the government. Firms do not have a monopoly on coercion and their motive -- profit -- is discernible, stable, and relatively acceptable when compared with the dangers that attend tyranny.")
The practices Calo describes may be unprecedented, but the harms are familiar, and consumer protection is something courts and legislatures have had to deal with before. They've had to "decide what makes a contract term 'unconscionable,' what kinds of enrichments are 'unjust,' when influence is 'undue,' what constitutes 'fair' dealing, where strategic behavior becomes 'bad faith,' when interest rates becomes 'usury' or higher prices 'gauging,' and on and on," Calo writes. "Such line drawing is endemic to consumer protection and other areas of the law concerned with basic notions of fair play."
That line drawing will undoubtedly be complicated and uneven. How can we know what someone's autonomous consumption desires are? When does consumption slip from autonomously directed to manipulated? (I for one am not even convinced people have ex-ante consumption desires that are identifiable beyond what marketers somewhere have shaped, though Calo's point is specifically about certain big-data-enabled strategies, not kludgy, old-fashioned marketing.) But despite the complexities, regulation is one route, and, if done well, will not suppress the online ad industry in general -- only the predatory practices at its margins.
But Calo also offers another option: "Imagine," he writes, "if major platforms such as Facebook and Google were obligated, as a matter of law or best practice, to offer a paid version of their service." Web companies do need revenue, after all. Fees may be the best way to both protect consumers and fund the tools we all use online. There may be other hazards with such an approach (e.g. will this systematically relegate poorer people to lesser version of these services?), but Calo says that there would be ways to address them.
But I'm getting ahead of myself. There will be time and processes for sorting out how to protect consumers in such a landscape. The value of Calo's paper is not in laying out where we should go from here but in disentangling the mess of problems related to identity and privacy online, and extracting from that mess a set of issues that are recognizable: consumer protection. In doing so, he gives, finally, a shape and a texture to that creepiness we know so well, but have understand so poorly.