The use of top-down models, for load forecasting purposes, has been the dominating method over the last decades. However, there is now a discussion regarding the performance of the top-down models, e.g. in situations with unusual weather conditions due to the lack of historical data. This paper considers an alternative bottom-up approach with a stronger relation to the laws of physics. Electricity companies in Sweden are installing automatic meter reading systems for their customers, and using the consumption data gives new possibilities when adapting the modeling parameters in a bottom-up model for each single customer. A method for analyzing individual consumption series is suggested, where different periods in time is used to divide and identify different parts of the electricity load; base load, heat load and household loads. A review of previous work is presented, and suggestions how to link the load analysis to construction parameters for an individual building is proposed.