Smart meter data can support more relevant household energy comparisons by grouping users according to their consumption patterns. A tested energy-cyber-physical system combined load-shape analysis with classification to update these groups over time and achieved high accuracy in weekly and monthly feedback cycles.