International research has demonstrated that lesbian, gay, bisexual and transgender (LGBT) youth have elevated rates of suicide and self-harm. What is missing from the evidence base, however, is qualitative research investigating LGBT youth perspectives. This is a sensitive subject area presenting ethical, methodological and epistemological challenges, especially in relation to over-sampling the 'visible' sections of a hidden population, retrospective reporting, and capturing complex emotions. We report on our use of qualitative online methodology to examine Internet forums where LGBT youth discuss self-harming. We found that this methodology can address some research dilemmas by generating: (a) diverse samples in terms of sexuality and gender identities; (b) a different type of data, immediate and unmediated by researchers; and (c) complex psychosocial emotional data. We argue that this online data can enhance our understanding of the links among hard-to-reach youth, suicide, self-harm, sexuality and gender, which is crucial to developing effective and appropriate suicide prevention strategies and mental health policies.