We show that task juggling, i.e., the spreading of effort across too many active projects, decreases the performance of workers, raising the chances of low throughput, long duration of projects and exploding backlogs. Individual speed of job completion cannot be explained only in terms of effort, ability and experience: work scheduling is a crucial "input" that cannot be omitted from the production function of individual workers. We provide a simple theoretical model to study the effects of increased task juggling on the duration of projects. Using a sample of Italian judges we show that those who are induced for exogenous reasons to work in a more parallel fashion on many trials at the same time, take longer to complete similar portfolios of cases. The exogenous variation that identifies this causal effect is constructed exploiting the lottery that assigns cases to judges together with the procedural prescription requiring judges to hold the first hearing of a case no later than 60 days from filing.