The electricity industry is currently undergoing changes in a transitioning period characterised by Energy 3D: Digitalisation, Decentralisation, and Decarbonisation. Smart meters are the vital infrastructure necessary to digitalise the energy system as well as enable advancements in decentralisation and decarbonisation. As of today, more than 500 million smart meters have been installed worldwide, with that number expected to rise to several billion installations over the decade. Smart meters enable electricity load to be measured with half-hourly granularity, providing an opportunity for demand-side management innovations that are likely to be advantageous for both utility companies and customers. Among these innovations, time-of- use (TOU) tariffs are widely considered to be the most promising solution for optimising energy consumption in the residential sector, however actual use is still limited. The objective of this thesis is to investigate opportunities and problems related to TOU tariffs utilising smart meter data at the national level. The authors have identified four major research gaps which need to be filled in order to expand commercial applications of TOU tariffs. These gaps are the described and addressed in the following chapters: the "TOU load adaptation forecasting problem", the "TOU winner detection problem", the "TOU public dataset problem", and the "excess generation forecasting problem". This thesis demonstrates three modelling approaches and one new TOU dataset (CAMSL). A significant contribution to the field is through the discover of new summary statistical features (statistical moments) and assesses the capacity of these to encapsulate other more widely used explanatory variables of demand response. The thesis is concluded by discussing future works and policy implications, such as the necessity of the more tailored modelling works and public live-stream of smart meter data, which could accelerate the roll-out of the demand side management at the residential sector.