The increasing amounts of data requested by mobile client devices has given rise to broad research endeavors to determine how network providers can cope with this challenge. Based on real world data used to derive upper limits of web page complexity, we provide an evaluation of web browsing and localized caching approaches. In this paper, we employ two different user-browsing models for (i) individual mobile clients, (ii) mobile clients sharing one centralized small cell cache, and (iii) mobile clients operating in an energy-optimized co-located fashion. We find that for a given content popularity distribution, average group savings due to caching depend highly on the user model. Furthermore, we find that for the purpose of overall savings determinations, an aggregated virtual cache falls within less than ten percent of a more elaborate energy-conscious approach to caching.