In this article we compare the benefits for game design and development relative to the use of three Game User Research (GUR) methodologies (user interviews, game metrics, and psychophysiology) to assist in shaping levels for a 2-D platformer game. We illustrate how these methodologies help level designers make more informed decisions in an otherwise qualitative design process. GUR data sources were combined in pairs to evaluate their usefulness in small-scale commercial game development scenarios, as commonly used in the casual game industry. Based on the improvements suggested by each data source, three levels of a Super Mario clone were modified and the success of these changes was measured. Based on the results we conclude that user interviews provide the clearest indications for improvement among the considered methodologies while metrics and biometrics add different types of information that cannot be obtained otherwise. These findings can be applied to the development of 2-D games; we discuss how other types of games may differ from this. Finally, we investigate differences in the use of GUR methodologies in a follow-up study for a commercial game with children as players.