This dissertation analyzes customer-experienced delays in manual order picking systems, a core yet costly aspect of warehouse operations. Despite automation advances, manual picking remains prevalent due to its flexibility and low upfront costs. The study models order picking as queueing systems, treating arriving orders as customers, and investigates how warehouse design choices impact delays.nnTwo systems are examined: batch-picking and milkrun. In batch-picking, pickers collect items for multiple orders in one go, following optimized routes. The thesis compares four routing policies and derives exact results for the picking time’s mean and variance, helping estimate customer delays. It also explores how batch size and warehouse layout affect performance, offering exact distributions under specific conditions.nnMilkrun systems, where pickers continuously walk through the warehouse collecting items, are analyzed using polling models from queueing theory. For one picker and random storage, the study provides closed-form expressions for average delays. It extends this to more complex layouts and multiple pickers, revealing that picker clustering worsens performance. Although dispersive policies reduce clustering, they unexpectedly increase average delays, highlighting the complexity of system design.nnOverall, the research provides valuable insights into how operational and layout decisions influence efficiency and customer satisfaction in manual order picking environments.