Many researchers include “progress indicators” in web questionnaires. The rationale is that if respondents know how much remains they will be more likely to complete the task. Previous research has shown that progress indicators often do not increase and sometimes even hurt completion rates, but in some circumstances they do help. Our hypothesis in the current experiment is that the effect of progress indicators depends on how long respondents believe the task will take and on how long it actually takes. We explore this by varying three factors in a web survey: (a) the task duration promised in the survey invitation; (b) the actual length of the questionnaire; and (c) whether or not a progress indicator is displayed. We found that the effect of progress indicators depend on respondents' expectations and the degree to which they were realized; the presence of a progress indicator led to fewer breakoffs when respondents expected a short task based on the invitation and when the questionnaire was indeed short than when they expected the task to be longer. We discuss how all three factors work together in influencing respondents' question-by-question decisions to continue the survey or to break off.