This paper focuses on a two-stage coordination strategy for distributed storage systems located at the end-user level. A first day-ahead commitment stage lies on the Alternating Direction Method of Multipliers adapted to a decentralized problem in which the followers and leaders exchange prices (Lagrangian)/quantities information over iterations. Then a three-step coordination is proposed for real-time control to guarantee a fast convergence while trying to remain as close as possible to the profiles scheduled in the look-ahead stage. Specific attention is attached to the endusers' objective function and their willingness to respond to coordinator signals. A layout of up to 100 households is simulated with the performances assessed regarding the trade-off between loss of revenue and global objective improvements. The impacts of load and solar generation forecasts errors are also investigated along with the effect of storage parameters. Obtained results highlight the need for a trade-off between the targeted objective in the look-ahead phase, and the confidence that it can be fulfilled in real-time.