This study followed a pre-test/post-test/delayed post-test design to investigate the long-term impact of metapragmatic instruction on learners’ ability to modify email requests. Twenty-five Spanish university students with an upper-intermediate proficiency level in English participated in the study. Over a two-week period, they received six hours of instruction on request modifiers by applying a form-function-context mapping framework (Taguchi 2011). Target request modifiers in written Discourse Completion Tasks (DCTs) were analysed for frequency and variety. Learners’ self-evaluations of email appropriateness were also used to determine their degree of confidence when producing request modifiers. The within-group comparisons revealed that learners made a significant progress in their use of request modifiers and confidence level immediately after instruction and two months later. These comparisons, however, also indicated that some types of request modifiers were more amenable to instruction than others.