Remark on the phase shift in the Kuzmak-Whitham ansatz