In recent years, induction thermography has been proposed to detect and characterize shallow and open cracks on steel components. The specimen is inspected by an infrared camera that records the heating and cooling behavior after a short induced electrical current pulse. The presence of a crack changes the distribution of the induced eddy currents with a consequent more significant local heating around the region of interest than the related sound one. This work investigates the influence of the testing parameters with an experimental approach, considering the analysis of a ferromagnetic master specimen with imposed simulated open cracks of different depths. The influence of the relative position between the coil and simulated crack has been evaluated, considering both numerical and experimental results. Two different coils have been adopted to optimize the experimental tests in terms of inspection time and improve the signal-to-noise ratio. It has been demonstrated that the post-processing of the raw thermal data is necessary to have useful information about defect detection for a pulse duration of 10 ms with only 20 A. A linear relationship exists between the phase contrast and the nominal defect depth.