This paper examines how service providers may resolve the trade-off between their personalization efforts and users? individual privacy concerns. Finding that neither an optimized one-size-fits-all strategy, nor a market-driven specialization of providers or choices between different usage scenarios can solve the problem, we analyze how negotiation techniques can lead to efficient contracts and how they can be integrated into current technologies. The analysis includes the identification of relevant and negotiable privacy dimensions for different usage domains. Negotiations in multi-channel retailing are examined as a detailed example. Based on a formalization of the user?s privacy revelation problem, we model the negotiation process as a Bayesian game where the service provider faces different types of users. Finally an extension to P3P is proposed that allows a simple expression and implementation of negotiation processes. Support for this extension has been integrated in the Mozilla browser.
