Over the past decade, we have witnessed a bloom in the field of evolutive protein engineering which is fueled by advances in molecular biology techniques and high-throughput screening technology. Directed protein evolution is a powerful algorithm using iterative cycles of random mutagenesis and screening for tailoring protein properties to our needs in industrial applications and for elucidating proteins' structure function relationships. This review summarizes, categorizes and discusses advantages and disadvantages of random mutagenesis methods used for generating genetic diversity. These random mutagenesis methods have been classified into four main categories depending on the method employed for nucleotide substitutions: enzyme based methods (Category I), synthetic chemistry based methods (Category II), whole cell methods (Category III) and combined methods (Category I-II, I-III and II-III). The basic principle of each method is discussed and varied mutagenic conditions are summarized in Tables and compared (benchmarked) to each other in terms of: mutational bias, controllable mutation frequency, ability to generate consecutive nucleotide substitutions and subset diversity, dependency on gene length, technical simplicity/robustness and cost-effectiveness. The latter comparison shows how highly-biased and limited current diversity creating methods are. Based on these limitations, strategies for generating diverse mutant libraries are proposed and discussed (RaMuS-Flowchart; KISS principle). We hope that this review provides, especially for researchers just entering the field of directed evolution, a guide for developing successful directed evolution strategies by selecting complementary methods for generating diverse mutant libraries.