To investigate the influence of macrofaunal species on recovery in soft-sediment habitats, a multi-site defaunation experiment was conducted in intertidal sandflats in Mahurangi Harbour, New Zealand. Paired treatment and control plots were monitored for 394 d after defaunation, and the recovery trajectories of individual species and their relationships with environmental factors were evaluated over time. Recruitment events were not apparent drivers of recovery, as we observed a massive recruitment of some species by Day 203, but the abundance of these individuals largely decreased by the end of the experiment, regardless of location in the estuary or hydrodynamics. Multiple regression models revealed highly variable responses, but suggested that several factors including sediment type and post-settlement species interactions contributed to species persistence in recovering plots. Juveniles generally settled at sites where adults from the same taxa occurred in the ambient community, suggesting local settlement patterns. Populations of functionally important surface deposit feeders and suspension feeders, including the large bivalves Austrovenus stutchburyi and Macomona liliana, failed to recover at most sites. The latter is likely one of the major drivers of divergence between disturbed and control plots in this experiment. The generally slow recovery dynamics, the degree of divergence of recovering communities and the failure of some key functions to recover have important implications for the response trajectories of coastal benthic communities subjected to increasingly frequent disturbance.