New techniques for managing, specifying, and analyzing requirements in software engineering projects are frequently presented by consultants and agile trainers. However, the effectiveness of these techniques is not evaluated in a rigorous manner, leaving practitioners with the question ``Will it work in our company?'' In this paper, we investigate the performance of user story refinement techniques, specifically Example Mapping (EM). This is a Three Amigos session technique, in which people from different disciplines work collaboratively in a time-boxed workshop that aims to refine, or clarify, a user story with the use of examples. The creation of such examples is intended not only to obtain a more precise specification, but also and mostly to achieve shared understanding on the user story to develop among the team members. We investigate the performance of EM via two longitudinal case studies. To enable a rigorous validation of EM, we first define the Refinement Evaluation Tool (RET), a survey-based measurement instrument that extends the method evaluation model with questions that cover the shared understanding dimension. The results from our case studies show that EM performs well under certain conditions: the user stories should not be too small sized. We also investigated the learning effect for EM; our data indicates that two sessions are generally necessary for the team members to use the technique effectively.