Usability testing is one of the most used methods to define the level of usability of a software product. However, there is always uncertainty to determine the best method that complements user testing in a depth usability assessment. Nowadays, the concern of many software developers is to identify an appropriate methodology, whose evaluation methods could be capable of measuring all usability aspects of a user interface. For this reason, we conducted a heuristic evaluation as a preliminary step to the implementation of a usability test, in order to determine in which extent these methods complement each other, and establish the gaps which are covered by each of them. The heuristic evaluation was performed by five specialists in the field of Human-Computer Interaction, who identified a total of fifty-nine usability problems in a transactional WEB site. Subsequently, a usability test was conducted with the participation of eight postgraduate students of a master's program in Informatics Engineering. The results show that most of the usability problems which were detected during the usability testing, had already been identified by the heuristic evaluation. Nevertheless, there were significant differences in the importance that was given to each problem. Usability experts emphasized in aspects that were not relevant to end users.