Optimisation models for environmentally compatible energy planning based on the concept of economic equilibria share a common flaw that stems from their neoclassic roots: the hypothesis of a perfect information and the hypothesis of perfect economic rationality. One of the possible ways how to circumvent this issue, the method we propose first, consists in soft-linking data from sociological surveys that determine technical coefficients for the energy system optimisation model, creating a hybrid approach of coupling a deductive engineering model with typical inductive methods of social sciences. Consumer behaviour is described via usual technological attributes and used in virtual process technologies, keeping the model and data format compatible with all the tools and user interfaces already developed for energy planning models, including International Energy Agency (IEA) supported software platforms. The approach, called Social MARKAL and demonstrated as an example on the technology of lighting bulbs, can be extended to all demand sectors and to all models based on the concept of economic equilibria. The second proposed approach proposed goes even further by applying the methodology called Conjoint Analysis to find the "Willingness To Pay". The method can be applied on immaterial goods like insurance or air ticket pricing and here it is applied on energy services considered as a service, far from usual method of considering it as investment. Both methods eliminate the systematic error on the demand side where the efficiency of demand-side management measures is over-optimistic, which may lead to inaccurate decisions and poor policies. The improved model is thus better suited to build long term policies that are not solely based on technology progress but also taking into account social change. The most important innovation is that the energy consumer behaviour has been brought to the same mathematical optimisation platform and evaluated together and as a complement to technology change.