Mastering knowledge of user profile is one of the technical cornerstones for service providers who handle a large amount of service consumption data and are well positioned to dynamically infer user interests. This paper presents a technology allowing to gather usage data from different multimedia services, create and track users profiles in real-time and monetize them by targeting content or other personalized services. In particular, we address the critical issue of a large service provider having to manage services in both controlled domains (where the metadata structure is under its control) and domains out of its control. We present a solution based on keyword inference and mapping to enable profiling and personalization on such heterogeneous domains.