Multi-hop cluster hierarchy has been presented as an organization for large wireless sensor networks (WSNs) that can provide scalable routing, data aggregation, and querying. In this paper, we revisit the fundamental problem of maintenance of such a hierarchy. To this end, we observe that, due to tightly coupling their operation with the topology of the hierarchy, existing state-of-the-art cluster hierarchy maintenance protocols may not necessarily be efficient. Based on our observations, we make a case for a novel gossip-based hierarchy maintenance protocol that decouples its operation from the hierarchy topology. Through experiments with actual embedded implementations we have developed, we confirm that our protocol can outperform the existing state-of-the-art solutions by a few factors in terms of both the energy consumption and the latency of bootstrapping and recovering the hierarchy.