The ability to rapidly obtain significant and accurate information from extensive data records is a key factor for companies’ success in today's competitive environment. Different machine learning algorithms can be used to extract information from data. However, to enable their application appropriate data structures must be defined. In addition, the quality of data must be ensured to allow appropriate decisions to be made based on the resulting information. Condition-Based Maintenance (CBM) decisions usually result from the analysis of the combination of data monitored on equipment with events data, such as failures and preventive maintenance interventions. Thus, to enable CBM implementation, data from equipment maintenance history should be properly organized and systematized. This paper presents a study performed in a manufacturing plant with several production lines. A structure to properly organize the failure records data and an overall data structure, including data events and monitored data, were defined to enable the application of CBM. The information obtained based on the data structure for the failure records allowed prioritizing the failure modes of a machine for CBM implementation.