This paper describes an experimental investigation into the relationship of delay to user performance in 3D object placement. In the first experiment, 10 participants performed a placement task as delay and placement difficulty varied. Results were analyzed with statistical analyses of variance. Placement errors and times rose as delay and difficulty increased. An interaction indicated that controlling delay is particularly important when placement is already difficult. In vehicle control tasks, previewing forecasts required input, and is a useful way of compensating for delay. The second experiment examined the relationship of previewing to delay and difficulty in placement. With regard to delay and difficulty, the 15 participants replicated the first experiment's results. Previewing reduced placement times directly, and mitigated the effects of delay and difficulty. A three-way interaction indicated that when previewing is used, more delay can be tolerated even in difficult placement tasks.