The electrical conductivity of mixtures of conductive and insulating materials is reviewed. In general, the conductivity of such mixtures increases drastically at a certain concentration of the conductive component, the so-called percolation concentration. Among the parameters influencing the percolation concentration, the filler distribution, filler shape, filler/matrix interactions and the processing technique are the most important ones. On the basis of these parameters, different models have been proposed aimed at the prediction of the conductivity or the percolation concentration. It will be shown here that statistical, geometric or thermodynamic models explain the conductivity behaviour of specific mixtures on the basis of insufficient assumptions. However, the conductivity seems to be predictable with the help of structure-oriented models.