The design of thin-walled shell structures such as storage tanks is governed by the buckling strength verification under external pressure. By means of the buckling strength verification it is decided, if the shell wall requires bigger wall thickness or if – alternatively – circumferential ring stiffeners are required. In EC3-1-6, two different methods are offered for the buckling strength verification: a) elastic buckling stress design acc. to 8.5, where formulae are given in Annex D; b) plastic reference design acc. to 8.6, where formulae are given in Annex B; c) global numerical design acc. to 8.7, where LBA and MNA calculations are used; d) global numerical geometrically and materially nonlinear analysis with imperfections included (GMNIA) acc. to 8.8; In methods a) to c), the critical load (bifurcation) of the perfect shell is addressed, which is subsequently reduced by an empirical imperfection knock-down factor. Thus, design aims towards a prebuckling state. In method d), design aims towards a postbuckling state, where big deformations are present and membrane action has developed. For a tank with typical dimensions it has been shown in [1], that the GMNIA design resistance may be almost twice as much as the formula based design resistance. According to the EC design concept, the ultimate limit state (ULS) is characterized by the maximum load that a structure can bear. Thus it is unsatisfactory, that a designer without access to a powerful FEA package ends up at only half of the real capacity of the structure. In the present paper we evaluate data from various GMNIA calculations from industry. We propose an analytical model, by which the stable postbuckling state of a cylindrical shell under external pressure can be captured. [1] Knoedel, P., Wacker, T., Ummenhofer, T.: Generic substitute imperfection patterns for the GMNIA of tanks and silos made of steel or aluminium. Paper 124, Proc., 15th NSCC, 26-28 June 2024 Luleå, Sweden.