Deviations in a global supply chain directly affect a retailer’s on-time delivery performance, causing availability problems and lower customer satisfaction. If the variation can be anticipated and more accurate lead-times estimated, proactive measures can be taken to decrease the impact. Existing estimation approaches use machine learning algorithms based on historical data to determine a lead-time value for the future. However, those approaches can only handle knowledge available in a machine-readable form, while expert knowledge about the domain is not considered during the actual prediction. Therefore, this thesis describes three novel approaches used for delivery time predictions that combine a machine learning model with human input. The proposed logic covers two phases: learning based on actual delivery data and capturing human knowledge to cover exceptional situations not reflected in historical data. The proposed models and the resulting estimates were evaluated using deliveries from a retail company. This thesis shows that the pure machine learning model delivers better results than a combination of humans and machines. On the one hand, it is due to the difficulty of incorporating the complexity of human knowledge into the algorithm in a suitable way. On the other hand, the effect is caused by the human tendency to generalize and exaggerate. Although the pure machine learning model delivers superior estimation accuracy than the human-machine combination, the systematic analysis of the results presents insights for future development in this area.