There is an increasing interest in developing interventions for HIV and STD prevention that can be delivered on the Internet. However, we know little about what it takes to identify, recruit and retain participants in interventions so that we can test their efficacy and effectiveness. Objectives for this investigation were to evaluate rates of recruitment and retention in an Internet-based randomized controlled trial (RCT) to increase sexually transmitted disease (STD) prevention among men who have sex with men (MSM). The Smart Sex Quest study was a RCT conducted online. Eligible participants were MSM, at least 18 years old and US residents. After completing a baseline risk assessment, participants were exposed to tailored or control messages and asked to return to the site at three months for a follow-up interview. From January 2002 through June 2003, 3,625 persons logged on as potential study participants; of these, 563 were not eligible, while 1,286 left the site without filling out a baseline survey. Complete baseline data were available for 1,776 participants, all of whom were eligible to complete a follow-up. Complete follow-up data were available for 270 (15.2%) participants. While the Internet is a valuable tool for conducting research, conducting this longitudinal research online was severely affected by a loss to follow-up, and analyzing outcome data was hampered by significant differences between those who did and did not complete the study. Alternate ways to recruit for and evaluate online trials must be considered.