Transition loss in cosmetic manufacturing refers to hidden inefficiencies that occur between production activities, such as idle time, repeated steps, or uneven durations, which often go unnoticed by traditional monitoring systems. Although subtle, these delays can gradually reduce overall efficiency and output. This study proposes a hybrid approach that combines process mining and statistical process control (SPC) to detect, measure, and understand these inefficiencies using real production event logs. Process mining helps map actual work sequences, revealing patterns such as self-loops and time gaps that indicate potential loss. SPC provides a statistical view to check for variation and stability in recurring tasks. The results uncovered more than 408 hours of undocumented idle time, repeated machine cleaning, and signs of unstable performance in certain production lines - especially in the liquid and tube areas, affecting transitions for product sizes of 50 ml and 100 ml product sizes. The findings help PT XYZ better understand where time is lost and how to improve it. They can use this insight to refine workflows, focus improvements, and apply smarter monitoring going forward. This study offers practical input for ongoing optimization in cosmetic production.