Electronic submission to a conference is a process that is known to evolve nonlinearly in time, with a dramatic increase when approaching the deadline. A model has recently been proposed by Alfi et al. (Nature Physics, 2007) for such a process, and the question of its universality has been raised. This problem is revisited here from a data analysis and modeling point of view, on the basis of a larger data set. A new model is proposed that better describes the total evolution of the process (including saturation) and allows for a running prediction of the total number of submissions.