Sequential sorting facilities are a key step in the courier, express, and parcel delivery industry. In these facilities, staff are assigned to work areas (WAs) to sequentially process different commodities as they move through the facility. When setting the staff levels at these WAs, the shift manager needs to balance different objectives, such as the overall number of staff, the cost of unsorted mail, and how frequently the shift levels change. However, existing literature on staffing these facilities (particularly in the field of mail delivery) focuses on longer timescales, assumes simpler operational constraints, and generally assumes deterministic mail volumes. In this thesis, we develop novel deterministic and stochastic models to staff these facilities for a mail sorting centre. We also propose a framework for general problem-based scenario reduction to use with the stochastic model. The deterministic model is a time-expanded network design model, using staff numbers to increase throughput capacities between WAs. To account for the uncertainty of commodity volumes, we also propose a novel stochastic model. This model is a stochastic programming model where the workplan is the first stage decision, the mail volumes are stochastic, and how the mail is routed over time is the second-stage decision. To solve the stochastic model (and other similar models) more efficiently, we propose a framework to generalise several problem-based scenario reduction methods. We show the applicability of the framework by performing numerical tests using different combinations of candidate solutions and scenario reduction techniques on three different test problems, including the stochastic mail centre staffing problem.