Incremental Nonlinear Dynamic Inversion (INDI) has become a popular control strategy for unmanned aerial vehicles due to its disturbance rejection capabilities and minimal model reliance. However, its standard formulation neglects actuator dynamics, leading to undesired coupling in systems with heterogeneous actuator characteristics such as in the Variable Skew Quad Plane (VSQP). This paper analyzes the limitations of INDI in such scenarios and demonstrates how Actuator Nonlinear Dynamic Inversion (ANDI) solves them. The closed-loop transfer function analysis shows how ANDI eliminates cross-axis coupling by directly incorporating actuator dynamics into the control allocation process. Additionally, ANDI is compared to a modified INDI approach that employs lead-lag filters to homogenize actuator behavior. While this approach is effective in decoupling, it unnecessarily slows down system response and worsens saturation handling. Simulation and flight test results using the VSQP validate the theoretical findings, confirming that ANDI offers improved control accuracy on coupled axes without compromising performance on decoupled axes.