Significance Although efforts have been made to determine the structure of talin and the way it interacts with integrins through the “head” domain, our work shows now that many of the previous mechanistic models based on the talin adapter are likely to be misleading as they are constructed on a crystal structure representing an improperly folded talin head domain. In this work, we identified the problem with the current talin head model and proposed a FERM-folded talin head. By analyzing these structural features of the FERM-folded talin head in a cellular context, involving also the kindlin adapter, we are making a critical and unprecedented contribution to the understanding and regulation of cell-matrix adhesions.