Nonparametric production analysis supplements observed input-output combinations with a number of assumptions to model the production process. The doctoral thesis starts with a brief ‘crash course’ in nonparametric production analysis in Chapter 1. Notwithstanding its simple basic premises, this black box analysis is quite powerful and has low data requirements. However, it can be made even more powerful by opening up this black box. The common thread in this doctoral thesis is to go beyond this black box modeling. ‘Beyond black box modeling’ here refers to two interpretations. The first and conventional interpretation refers to more realistic models of production processes by, for example, explicitly modeling the different subprocesses and their links (cfr. Chapter 2) or by modeling intertemporal links between processes over time (cfr. Chapter 4). Apart from this conventional interpretation, it also refers to the idea of looking beyond mere efficiency scores or productivity measures produced by these black box models. An equally important analysis is tracing the underlying factors of these results (cfr. Chapter 2 and Chapter 3) and learning from (dominating) peers (cfr. Chapter 5).