by Jeff Sovern
I've been listening to the audio version of Stephen Baker's fascinating book The Numerati, and while I'm not done with it, I thought I would blog about some of what I've encountered in it thus far, as well as some thoughts it's prompted. Baker's thesis is that data miners are attempting to construct mathematical models of individuals which will enable them to predict how those individuals will respond to particular stimuli. Though Baker ranges across politics (how to get people to vote the way you want) and the workplace, I'm going to focus on consumer matters, as befits our Blog.
Suppose a grocer wants to persuade consumers to buy a particular item, say Advil. One option is to offer coupons, but those will reach many shoppers (or even non-shoppers) who won't respond to them. Thus, the grocer will waste money advertising to people who aren't interested, while also burdening shoppers who read the ad but are not interested in Advil. Inefficient. But now suppose the grocer offers frequent shopper cards that enable it to identify consumers who have used such coupons in the past. If the grocer can target such consumers with coupons, it can reach those likely to be influenced by the coupons without wasting money on (and wasting the time of) those who would not be interested in the coupons. So far, it sounds like a winning proposition all around.
Baker reports that existing technology permits that. Some stores use "smart carts," shopping carts that can flash to shoppers specials, etc., that are likely to be of interest to the particular shopper. The smart cart reads the chip embedded in the frequent shopper card to figure out what is likely to matter to the owner of the card. OK--but what if the item is something that the consumer might not want to have flashed on a screen, like an ad for savings on condoms or laxatives? Probably grocers would figure out not to use the technology on such items. Even so, some consumers might be troubled by the invasion of privacy this technology implies, but they could always not use the frequent shopper card, if they know what use of that card permits retailers to do. But that's something that probably few consumers are aware of. For what it's worth, I wrote about whether collecting and using information about consumers without informing them of that fact violates the FTC Act and UDAP statutes in an article, Protecting Privacy with Deceptive Trade Practices Legislation, 69 Fordham Law Review 1305 (2001).
The book also has implications for consumer discrimination law. Baker describes how the technology can be used to drive away undesirable shoppers, the so-called "barnacles" who buy only a few items, and those on sale. Barnacles may cost grocers money by buying loss-leaders (items priced at below cost to attract consumers in the hope that the consumers, once in the store, will also buy more profitable items) and not buying anything else. Smart carts could flash barnacles specials on caviar and other expensive items rather than the items that more conventional shoppers might see, in an attempt to cause the barnacles to go elsewhere. Put that way, it's kind of funny, but what if the grocer uses the smart card to show different specials to consumers of different races or genders--something that the technology would not only be able to do, but because of its ability to target consumers who are interested in different products, seems like a real possibility, if it turns out that people of different races or genders respond differently to the same offers. Could we end up with price discrimination based on such classifications, with some groups getting access to specials that others never learn about?
Baker writes about how when one cell phone company charged an extra fee for a particular service, it received numerous complaints from consumers almost all of whom shared certain characteristics. Suppose, he goes on, the company had offered that service for free to consumers with those characteristics, but charged consumers who were less likely to complain? Again, we have price discrimination. Is that troubling? Suppose those of a particular race or gender generally end up paying more for that service. I think most people would find that troubling. I know I do. How would we even know?
Anyway, it's a thought-provoking book and I'm looking forward to the rest of it.