The assessment of the shear strength of reinforced concrete beams with and without shear reinforcement is investigated in this paper on the basis of the contribution of the various potential shear-transfer actions. To that aim, a comprehensive experimental programme is presented, aimed at representing realistic conditions in reinforced concrete members. The programme comprised 15 cantilevers and 15 continuous beams with and without stirrups. In addition, some of the tests were designed to fail in shear after yielding of the flexural reinforcement, allowing to investigate on the structural implications of this phenomenon. The tests were instrumented by means of surface measurements using Digital Image Correlation (DIC), tracking in a detailed manner the development and kinematics of the critical shear crack leading to failure. Based on these measurements, and by accounting for suitable constitutive models, the contribution of the shear-transfer actions is estimated at different levels of load. The comparison of the experimental results with the calculated shear strength (sum of the contribution of the various shear-transfer actions at failure) shows consistent agreement for the various cases investigated. On this basis, several conclusions about the dominant shear-transfer actions at shear failure are presented, particularly on the role of the shear reinforcement and on the implications of the flexural reinforcement yielding on the overall response.