We study the productivity and practices of individual information workers at a midsize executive recruiting firm using detailed task-level data. Specifically, we analyze data on (1) project level accounting measures of revenues, completion rates, and individual compensation, (2) self-reported individual perceptions of information technology skills, use and information sharing, and (3) direct observation of all e-mail activity over a period of six months. We develop a multistage model of production activities at the firm, including a multitask incentive model, and analyze the correlations among key technology, work practice, output and compensation measures. We find that a) information technology use is positively correlated with increase revenues at the individual level; b) increase revenues are correlated with increased compensation in a simple, direct way, c) certain intermediate measures of performance improve with IT use (e.g. multitasking), while others worsen (e.g. task duration), and d) compensation is positively correlated with observable performance metrics but can be negatively correlated with effort on non-observable tasks. These data shed light on the interactions among technology use, work practices, compensation and performance among a group of project-based information workers and are consistent with simple models of queuing and multitasking.