Speaking self-perceived communication competence (SPCC) is a construct with many potential implications for foreign language learning, but one that has been little studied. SPCC itself is a major predictive factor in willingness to communicate, a construct which has been widely conceptualised and researched. This study (N = 103) used a repeated measures ANOVA to investigate SPCC and its correlation with actual L2 speaking proficiency over the course of a year; there was no significant correlation. Qualitative data was then treated with grounded theory to establish why SPCC was inaccurate and to provide pointers as to how SPCC accuracy might be improved. The findings are discussed with reference to the literature in an attempt to establish a deeper understanding of SPCC, particularly in the Japanese context, its formulation and its implications for foreign language learning.