A method of practical use was set up to determine the connectivity in a porous material, modelling the physical system as a lattice, whose coordination number is assumed to be an index of connectivity itself. This task was approached through the theory of percolation and input data were provided by two different experimental techniques, that is, adsorption/desorption of nitrogen and mercury porosimetry. The overall procedure is based on the calculation of probability f(P) of occupation of the porous channels and of probability F(P) of percolation. In the framework of the above--mentioned lattice model, the average coordination number Z is calculated through the best fitting of a universal curve to the values found for F(P) and f(P), adopting as fitting parameter the ratio L between the characteristic linear dimension of the whole lattice and the characteristic linear dimension of each of its cells. The procedure described was implemented through a numerical code and applied to three commercial alumina. A simple empirical relationship was found between Z and the percolation threshold, showing an excellent coefficient of statistical correlation. The three products proved different in connectivity, allowing subtle distinctions from each other, despite their hysteresis cycles in the adsorption/desorption process appeared quite similar from a qualitative standpoint.