Youth mentoring is a popular and well-respected volunteer activity in the USA. Although there is support for the efficacy and cost-effectiveness of community-based mentoring (CBM), the research on school-based mentoring (SBM)--the most popular form of youth mentoring--is mixed. Most of the SBM studies find inconsistent or small effects. The reason for the difference in the success of CBM and SBM may be the unique context, structures and goals of SBM relative to CBM. In an effort to guide SBM programmes towards larger and more consistent positive results, the authors review context-specific considerations of providing mentoring in schools. We present several evidence-based practices that are compatible with SBM and should strengthen the effects of this popular intervention. To illustrate the importance of these considerations, we provide a case example of redeveloping a failed SBM programme into a promising programme by reviewing two iterative randomized evaluations. Finally, we encourage school-based mental health experts' involvement in the redevelopment of SBM.