This paper presents results from an experimental evaluation on the pre-and postbuckling behavior of 12 steel wide-flange cantilever columns under axial load and lateral drift demands. The influence of several loading and geometric parameters, including the cross-sectional local web and flange slenderness ratios, applied axial load, and lateral and axial loading history on the performance of these columns is thoroughly examined. The test data indicate that crosssectional local buckling is highly asymmetric in steel columns under variable axial load. A relatively high compressive axial load can significantly compromise the steel column seismic stability and ductility but this also depends on the imposed lateral loading history. The AISC axial load-bending moment interaction equation provides accurate estimates of a steel column's yield resistance. However, the same equation underestimates by at least 30% the column's peak resistance regardless of the loading scenario. Measurements of column flange deformation, axial shortening, flexural resistance and lateral drift are combined in a single graphical format aiding the process of assessing steel column repairability after earthquakes. The test data suggest that current practice-oriented nonlinear component modeling guidelines (PEER/ATC 2010) may not provide sufficient accuracy in establishing both the monotonic and first-cycle envelope curves of steel columns. It is also shown that high-fidelity continuum finite element models shall consider geometric imperfections of proper magnitude in addition to the steel material inelasticity to properly simulate the inelastic buckling of wide-flange steel columns and generalize the findings of physical tests. Issues arising due to similitude are also discussed to properly limit steel column instability modes in future studies.