This case study summarises researchers' experiences using Twitter as a research tool. Specifically, that involved using Twitter for the identification, invitation and incentivisation of survey participants in a study of alcohol consumption among Australian Twitter users. In this ‘auto-ethnographic research war story’, the authors describe their discovery of four unexpected features of the Twitter ecosystem. These are rorting, service provision issues, risk-averse behaviour and Twitter users' community-based behaviour. The authors also describe the efforts they made to remedy the roadblocks they encountered, and how they sought to preserve the integrity of their research. This case will be of benefit to social media and marketing practitioners as well as academic researchers and research students. It is offered here in the hope that they can avoid similar issues in their own work.