The wide spread of smart metering roll out enables a better understanding of the consumer behavior and tailoring demand response DR programs to achieve cost-efficient energy savings. In the residential sector smart metering allows detailed readings of the power consumption in the form of large volumes time series that encodes relevant information for distribution network operators DNOs to manage in optimal ways low-voltage networks. Further, DNOs may leverage the smart meter data to identify customer group for energy efficiency programs and demand side response DSR (e.g., dynamic pricing schemes). In this paper, we outline the application of smart meter data mining to identify consumers who are more responsible for the peak system using responsibility factor and consumption variability. Identification of consumers having higher responsibility to the peak system may yield to better enhance energy reduction recommendations and enable more tailored dynamic pricing plans depending on the consumer's influence on the utility peak. Responsibility factor and consumption variance have been investigated as input features of the clustering algorithms. Two clustering techniques, hierarchical clustering and self-organising map SOM, have been used to study the resulting customer groups and to have an effective graphical visualization of the customer's cluster distribution on the input feature space.