Is feedback delivered by an expert sufficient to improve performance? In two studies, we tested, following failure, the influence of group membership (ingroup/outgroup) and source expertise (high/low) on the effectiveness of attributional feedback on performance. Results revealed a significant interactive effect, showing an increase of performance only when the source was an expert ingroup member (Study 1). This interaction was replicated on performance and success expectations in Study 2, which were significantly higher for high compared to low expertise ingroup sources. These data suggest that sharing a common identity with those you lead may help convert expert performance advice into real performance benefits.