Attendance at professional conferences is well studied, but few studies have analyzed presentation attendance within a conference. Understanding why some presentations attract larger audiences than others can be useful for both presenters and conference organizers. To this end, we applied a random forest model to data from 1,676 presentations at the 2015 Annual Meeting of the American Fisheries Society in Portland, Oregon. Our results suggested that presentation time and location, presenter travel distance and h-index (a metric of scholarly impact), and the lengths of titles and abstracts were important drivers of attendance. Based on these results, we suggest ways for organizers to maximize attendance potential and for presenters to attract large audiences. Finally, we suggest ways to improve data collection and analysis, and we highlight how insight into presentation attendance can benefit the planning and evaluation of future meetings.