Genotyping of bacteria through typing of loci containing a variable number of tandem repeats (VNTR) might become the gold standard for many pathogens. The development of genome sequencing has shown that such sequences were present in every species analyzed, and that polymorphism exists in at least a fraction of them. The length of these repetitions can vary from a single nucleotide to a few hundreds. This has implications for both the techniques used to measure the repeat number and the level of variability. In addition, tandem repeats can be part of coding regions or be intergenic and may play a direct role in the adaptation to the environment, thus having different observed evolution rates. For these reasons the choice of VNTR when setting a multiple-loci VNTR analysis (MLVA) assay is important. Although reasonable discrimination can be achieved with the typing of six to eight markers, in particular in species with high genomic diversity, it may be necessary to type 20 to 40 markers in monomorphic species or if an evolutionary meaningful assay is needed. Homoplasy (in the present context, two alleles containing the same repeat copy number in spite of a different history) is then compensated by the analysis of multiple markers. Finally, even if the underlying principles are relatively simple, quality standards must be implemented before this approach is widely accepted, and technology issues must be resolved to further lower the typing costs.