Online self-help programs such as the Acceptance and Commitment Therapy (ACT) Guide address significant barriers to receiving therapy but suffer from low adherence rates. Peer-delivered coaching, using undergraduate students as coaches, is an innovative alternative to traditional coaching methods which may improve adherence in a more scalable format. To test the efficacy of peer-support coaching for college students using ACT Guide in a naturalistic setting, we conducted a pragmatic randomized controlled trial with three conditions (phone coaching, text message coaching, and a no support control group; N = 230). Participants were block randomized, and program adherence (i.e., number of modules completed) was the primary outcome. Participants completed significantly more ACT Guide modules in phone (M = 7.1, SD = 4.9) and text (M = 5.7, SD = 5) coaching than the no support control condition (M = 1.6, SD = 3.3, p < .001). Participants who received phone coaching experienced significant improvements across almost all measured outcomes as compared to the control group, while participants who received text coaching improved only on some outcomes compared to control (ps < .05). These results support peer-support coaching, particularly when delivered through phone calls, as an efficacious method for improving adherence to and outcomes from online self-help programs.