Liquid biopsies have emerged as a useful addition to tissue biopsies in the field of molecular pathology. Previous literature has shown lower laboratory performances when a new technique or method for variant analysis is introduced. The aim of this study was to evaluate the differences in variant analysis between tissue and plasma samples after its introduction as a new sample type for molecular analysis. Data from a pilot external quality assessment scheme for the detection of molecular variants in plasma samples and external quality assessment schemes on tissue samples were collected. Subsequently, laboratory performances and error rates on sample level were compared between matrices for variants present in both scheme types. Results show lower overall performances (65.6%, n=276 compared to 89.2%, n=1,607) and higher error rates (21.0% to 43.5%, n=138, compared to 8.7% to 16.7%, n=234 to 689) for the detection of variants in plasma compared to tissue. In the plasma samples, performance decreased for variants with an allele frequency of 1% (56.5%, n=138) when compared to 5% (74.6%, n=138). Our analysis confirms that the implementation of new methods for detecting circulating cell-free tumor DNA in cell-free plasma is associated with poor performance. It is important to apply optimal detection methods as well as to extensively validate new methods for cell-free tumor DNA testing before treatment decisions are made.