The purpose of talent identification (TI) is the earliest possible selection of auspicious athletes with the goal of systematically maximizing their potential. The literature proposes excellent reviews on various facets of talent research on different scientific issues such as sports sciences or genetics. However, the approaches of conventional and genetic testing have only been discussed separately by and for the respective groups of interest. In this article, we combine the discoveries of these disciplines into a single review to provide a comprehensive overview and elucidate the prevailing limitations. Fundamental problems in TI reside in the difficulties of defining the construct ‘talent’ or groups of different performance levels that represent the target variable of testing. Conventional and genetic testing reveal a number of methodological and technical limitations, and parallels are summarised in terms of the test designs, the point in time of testing, psychological skills or traits and unknown interactions between different variables. In conclusion, many deficiencies in the current talent research have gained attention. Alternative solutions include the talent development approach, while genetic testing is re-emphasised as a tool for risk stratification in sport participation. Future research needs to clearly define the group of interest and comprehensively implement all methodological improvement suggestions.