In benthic studies, the identification of organisms at the species level is known to be the best source for ecological and biological information even if time-consuming and expensive. However, taxonomic sufficiency (TS) has been proposed as a short-cut method for quantifying changes in biological assemblages in environmental monitoring. In this paper, we set out to determine whether and how the taxonomic complexity of a benthic assemblage influences the results of TS at two different long-term ecological research (LTER) sites in the Po delta region (north-eastern Italy). Specifically, we investigated whether TS can be used to detect natural and human-driven patterns of variation in benthic assemblages from lagoonal soft bottoms. The first benthic dataset was collected from 1996 to 2015 in a “choked” lagoon, the Valli di Comacchio, a lagoon characterised by long water residence times and heavy eutrophication, while the second was collected from 2004 to 2010 in a “leaky” lagoon, the Sacca di Goro, a coastal area with human pressure limited to aquaculture. Univariate and multivariate statistical analyses were used to assess differences in the taxonomic structure of benthic assemblages and to test TS on the two different datasets. TS seemed to work from species to family level at both sites, despite a higher natural variability of environmental conditions combined with multiple anthropogenic stressors. Therefore, TS at the family level may represent effective taxonomic surrogates across a range of environmental contexts in lagoon environments. Since the structure of the community and the magnitude of changes could influence the efficiency of taxonomic surrogates and data transformations in long-term monitoring, we also suggest periodic analyses at finer taxonomic levels in order to check the efficiency of the application of taxonomic substitutes in routine monitoring programmes in lagoon systems.