Smart meter rollout activities worldwide generate high-resolution energy consumption data that has so far been underutilized to support private households in measuring their individual energy consumption and improving efficiency. This paper presents the development and evaluation of a user-centered Energy Information System (EIS) for private customers, following the Design Science Research (DSR) paradigm. Based on 119 user-identified requirements, 5 Design Principles, and 17 Design Features, a web probing study (n=35) validated individual UI elements in a first formative step. Afterwards three incremental system variants were built and evaluated in a controlled between-subjects experiment (n = 46). Task-based metrics (Task Success Rate, Efficiency), the Technology Acceptance Model (TAM), and the System Usability Scale (SUS) were applied. Looking at the results, the objective performance metrics (TSR and Efficiency) show significant differences between the instantiation variants. Interestingly, TAM and SUS do not replicate these subjectively, their scores are clustered close together across the instantiations. All variants achieve acceptable usability scores, and Actual Use shows a large effect favoring the feature-richer variants (d=0.87). This two-stage approach is methodologically relevant and enables the precise identification of potential feature-specific understanding and acceptance barriers, before they are integrated into a full system prototype. The study highlights the need for multi-method evaluation and provides recommendations for stepwise implementation, market launch, or revision of existing Energy Information Systems.