Longitudinal cohort designs (with three or more measurement occasions) are invaluable to investigate between- and within-individual variation in outcomes. However, traditional longitudinal designs require a lengthy implementation and data collection period and impose a substantial burden on participants and investigators. We discuss alternative longitudinal designs, including planned missing data designs and retrospective cohort studies with secondary data, which require a shorter period for data accrual and reduce participant burden while maintaining statistical power. We also discuss analysis strategies to maximize data use and produce unbiased estimates of treatment effectiveness, including models for recurrent or multistate events and time-varying covariates.