This paper uses both experimental and numerical approaches to revisit the concept of current transfer length (CTL) in second-generation high-temperature superconductor coated conductors with a current flow diverter (CFD) architecture. The CFD architecture has been implemented on eight commercial coated conductor samples from THEVA. In order to measure the 2D current distribution in the silver stabilizer layer of the samples, we first used a custom-made array of 120 voltage taps to measure the surface potential distribution. Then, the so-called ‘static’ CTL () was extracted using a semi-analytical model that fitted well the experimental data. As defined in this paper, the static CTL on a 2D domain is a generalization of the definition commonly used in literature. In addition, we used a 3D finite element model to simulate the normal zone (NZ) propagation in our CFD samples, in order to quantify their ‘dynamic’ CTL (), a new concept introduced in this paper and defined as the CTL observed during the propagation of a quenched region. The results show that, for a CFD architecture, is always larger than , whereas when the interfacial resistance between the stabilizer and the superconductor layers is the same everywhere. We proved that the cause of these different behaviors is related to the shape of the NZ, which is curved for the CFD architecture, and rectangular otherwise. Finally, we showed that the normal zone propagation velocity (NZPV) is proportional to , not with , which suggests that the dynamic CTL is the most general definition of the CTL and should always be used when current crowding and non-uniform heat generation occurs around a NZ.