In a usual country level MARKAL, the role of the demand has been neglected for a long time. Over decades, it remained inelastic (Abilock et al. 1979), then piecewise constant (Loulou & Lavigne 1996), until it was connected to rest of the economy (A. Manne, MARKAL-MACRO 1992). The optimal configuration consisted of competing technologies, and technologies only. Demand-side management measures were modelled through sets of measures described as technologies that were with zero or low cost and therefore always used to their full potential. Putting bounds on their penetration was subjective and depended upon the modeller and his good sense. Another important disadvantage of this approach was that the model could become remote-controlled through bounds on demand-side management. This influence increases with the growing role of the demand sector when the regional scale or sectoral scope of the model is reduced (Fragnière & Haurie 1996). On a city level, typically there will be not much energy extraction or transformation but principally demand devices. Our approach uses sociological surveys to produce technical coefficient equivalents used in virtual process technologies (information campaigns in favour of energy savings and of technology switch) that can be put into direct competition with tangible technologies (such as residential lighting or space heating). An information campaign has usual technology attributes such as investment cost, duration, efficiency (yield). This information campaign may be selected in the optimal solution depending on its competitiveness. Indeed, actual information campaigns can trigger behavioural change (Fragnière et al. 2010). It is thus important to include this element of behavioural change into long-term energy model. To build these social parameters, we rely on sociological surveys. The results of the surveys are transformed to be included into demand side of the long-term energy planning models. In this paper, we show how we transform the results, especially ordinal and nominal data from quantitative surveys into parameters that are adapted to a long-term energy modelling tool such as MARKAL or TIMES. To illustrate this contribution, we will present the results of surveys that have been conducted in Romania.