DNA markers are increasingly used for studying the taxonomic composition of environmental samples that contain DNA from different species. Here, I demonstrate that some of the currently used identification techniques based on K2P are flawed and reveal that DNA barcode species coverage for Metazoa in open-access databases remains poor. I then test based on existing and newly generated barcodes whether metabarcoding or metagenomics should be used for studying the DNA in primate fecal samples. I conclude that the latter is preferable because it simultaneously provides information on diet, host genetics, and parasite loads. In addition, newly developed read-based identification strategies reveal that metagenomic data provide more taxonomic precision for plant species than the barcodes used in metabarcoding. However, metabarcoding may be useful for diet quantification given that read counts are correlated between metabarcoding and metagenomic datasets for the same samples.