High-throughput sequencing technologies are changing the way in which diversity is studied at all scales and has the greatest potential to facilitate studies of taxa that are intractable to other methods. Insect ecology is one such field, with great abundance and diversity combining with incomplete taxonomic knowledge to hamper studies of diversity at large spatial and temporal scales. A new high-throughput method has recently been proposed to address such issues within a self-contained phylogenetic framework that is linkable with existing biological knowledge via Linnaean taxonomy. This method, 'mitochondrial metagenomics' (MMG), has already been the subject of a number of proof-of-concept studies, frequently focussed on Coleoptera. These studies are unified here with additional similar datasets for the first time to draw together the lessons to be learnt from the results obtained to date and infer the immediate methodological questions that remain to be answered. Particular attention is paid to the prospect of bulk sequencing of mixed specimens and the associated bioinformatics challenges. Consideration is given to mitochondrial phylogeny reconstruction with the prospect of rapidly increasing taxon sampling and the potential for phylogeny-based taxonomy assignment for otherwise uncharacterised communities. Mitochondrial metagenomics is then applied to a landscape-level assessment of the response of the leaf litter beetle communities to habitat differences, taking a combined compositional and phylogenetic perspective. Finally, the results are synthesised for a perspective on the remaining methodological impediments to the further development of MMG, and the future prospects for synthetic analyses of diversity are considered.