This paper presents the results of a comprehensive experimental programme aimed at studying the interaction of local and overall flexural buckling in cold-formed steel (CFS) plain and lipped channels under axial compression. The results were further used to verify the accuracy of the current design procedures in Eurocode 3, as well as to evaluate the effectiveness of a previously proposed optimisation methodology. A total of 36 axial compression tests on CFS channels with three different lengths (1 m, 1.5 m and 2 m) and four different cross-sections were conducted under a concentrically applied load and pin-ended boundary conditions. The initial geometric imperfections of the specimens were measured using a specially designed set-up with laser displacement transducers. Material tests were also carried out to determine the tensile properties of the flat parts of the cross-sections, as well as the cold-worked corner regions. A comparison between the experimental results and the Eurocode 3 predictions showed that the effective width approach combined with the P–M interaction equation proposed in Eurocode 3 to take into account the shift of the effective centroid consistently provided safe results. However, the Eurocode 3 procedures were also quite conservative in predicting the capacity pertaining to local-global interaction buckling, especially for plain channels. Furthermore, the experimental data confirmed the results of an optimisation study and demonstrated that the optimised CFS columns exhibited a capacity which was up to 26% higher than the standard channel with the same amount of material taken as a starting point.