Nowadays, keeping track of buyers has become easy and inexpensive, and the established theories that had firms or sellers treat buyers as anonymous are enhanced by ones where the sellers not only keep track of the buyers with whom they are dealing, but also design their pricing schemes (coupons, loyalty programs) based on the information that they have obtained thus far. In an influential paper, Odlyzko (2003) argues that most compelling reason for collecting and storing information about consumers is price discrimination. At the same time, the ability to collect and store information cheaply has raised privacy concerns, and justifiably so. There is considerable debate among academics and policy makers about the value that individuals attribute to privacy and the government’s role in protecting it, as well as works that try to estimate the willingness to remain anonymous (Acquisti, John, and Loewenstein (2010). There is also a literature that builds on the classical price discrimination literature by a monopolist, and studies optimal pricing in environments where the seller can condition price offers on past behavior (eg. Acquisti and Varian (2005), Fudenberg and Villas-Boas (2006) or Conitzer, Taylor and Wagman (2011). Conitzer, Taylor and Wagman (2011) examine optimal pricing, but allow consumers to take actions in an effort to protect their privacy. Experimental evidence reported in Tsai, Egelman, Cranor and Acquisti (2011) shows that people are willing to pay a premium for privacy.
This paper considers a seller who interacts multiple periods with the same buyers, but instead of deciding what price to post in each period, the seller can choose any selling procedure. In such an environment the amount of information that a seller collects in each period is crucial for the subsequent mechanism offers which, in turn, affects the buyers’ behavior early on.
The results of this paper answer the following questions: If the seller can hire an intermediary (or choose a platform) that runs the mechanism and discloses information, what intermediary will she choose? An opaque one, a fully transparent one, or something in the middle? Does the seller benefit from having access to such intermediaries? Do optimal intermediaries disclose too much information about buyers? Would buyers want to take actions to protect their privacy?
We show that minimally transparent platforms (intermediaries) are optimal. The literature (McAdams and Schwarz (2007)) stresses the desirability of hiring intermediaries: If the seller can hire someone with greater commitment power that is beneficial and she would be willing to pay up to the expected revenue loss due to her lack of commitment. Our analysis highlights, however, that if the seller can employ sophisticated procedures to make inferences at subsequent periods, this can nullify the potential benefit from hiring an non-transparent intermediary. How? Opaqueness can create correlation in the buyers’ private information, which can be exploited by the seller by using Cremer-McLean lotteries, resulting to the same outcome as if the seller was using maximally transparent intermediaries. More surprisingly, we show that minimally transparent platforms help the seller generate strictly higher revenue if the seller promises to treat buyers more favorably, by employing mechanisms that satisfy voluntary participation constraints ex-post. How does it work? Ex-post voluntary participation constraints work as a commitment devise for the seller: The seller commits not to use sophisticated procedures to learn information about the buyers. This additional constraint on the seller’s problem strictly increases the revenue generated. Ex-ante this constraint appears to be in favor of the buyers, but ends up helping the seller. This shows that laws aiming at protecting consumer privacy, may actually end up helping sellers, because they provide a commitment device not to observe information that sellers would be tempted to exploit in the future.