To date, research has mostly focused on the impact of energy efficiency on the total electricity demand but not on the electricity demand profiles. To address this gap, we estimate the impact of energy efficiency measures and policies such as minimum energy performance standards on the peak load by developing a bottom-up model that generates Swiss household hourly electricity demand profiles per appliance based on time use data. The model estimates that evening appliance peak demand can be reduced by 38% when the appliances are replaced by the highest energy efficiency label available on market. We find that changing light bulbs to LED would have the same peak reduction as switching cooking or wet appliances to off-peak periods throughout the year. We also show that the evening appliance peak demand could reduce in 2035 by 24% thanks to the improvement of the energy performance of the stock. Cooking appliances, the least favourable appliances to be involved in demand response, is expected to be the highest contributors to the evening peak in 2035. Our findings show that policy makers should pay due attention to energy efficiency improvement not only for reducing electricity demand but also in order to reduce peak load.