Quality assurance of external beam (radio)therapy (EBT) requires tools with specific characteristics. A radiochromic film dubbed "Gafchromic EBT" (G-EBT) that is particularly suited for external beam therapy because of its features was introduced in 2004. Its characteristics, especially the high spatial resolution, make it suitable for measurement of dose distributions in radiotherapy, especially intensity-modulated radiation therapy (IMRT). While several aspects of the film characteristics have been previously reported separately, we present a comprehensive evaluation centered on practical IMRT verification, leading to an optimized protocol. Therefore the constancy within one batch, the relationship between optical density (OD) and dose (dose range between 1.4 Gy and 8.4 Gy) and the dose rate dependence for four dose rates (55, 108, 217, 441 MU/min) were investigated. In addition to these characteristics, energy dependence between two energies (50kV and 6 MV), tissue equivalency, post irradiation coloration over one month, pressure and temperature sensitivity were evaluated. We then optimized the protocol using the G-EBT films, in combination with an EPSON-Expression 1680 pro flatbed scanner, for IMRT QA, while either striving to keep the compound error as small as possible or trying to reduce evaluation time. As a basis for this protocol optimization, the characteristics of the scanner (such as inhomogeneity of the scanning field) and its software (such as consequences of extracting only the red color channel) had to be determined first. The interaction of film and scanner (variation of the OD depending on the scanning direction or the scanning resolution) was assessed as well. Using the optimized protocol for IMRT QA, the compound error could be reduced to approximately 2% for a quality-driven approach and maximum 5.5% for an approach attempting to reduce procedure time. While the quality-driven approach provides appropriate accuracy for individual patient QA, the procedure-time driven approach can only be used for preliminary measurements.