Two of the most popular vibration testing methods for nonlinear structures are control-based continuation and phase-locked-loop testing. In this paper, they are directly compared on the same benchmark system, for the first time, to demonstrate their general capabilities and to discuss practical implementation aspects. The considered system, which is specifically designed for this study, is a slightly arched beam clamped at both ends via bolted joints. It exhibits a pronounced softening–hardening behavior as well as an increasing damping characteristic due to the frictional clamping. Both methods are implemented to identify periodic responses at steady-state constituting the phase-resonant backbone curve and nonlinear frequency response curves. To ensure coherent results, the repetition variability is thoroughly assessed via an uncertainty analysis. It is concluded that the methods are in excellent agreement, taking into account the inherent repetition variability of the system.