Army training device maintenance tends to be based on a preventative maintenance schedule supplied by the trainer manufacturer and planned based on achieving standards of operational availability. These schedules do not take into account operational use. The Army can reduce costs by tying the maintenance of training systems to the use of the systems. Although there are many ways in which usage can be tracked, it is likely that some combinations of metrics are more closely correlated to the need for corrective action maintenance. Selection of the most appropriate measure or set of measures often depends on the design of the training device. The correct combination of metrics is best determined through modeling of historical data containing the metrics in combination with the failure information. A specific comparison for two different types of Army trainers is presented. The use of data visualization in support of analytics is explored. Figures presented show the ability of graphics to reduce large amounts of data to displays that are readily interpreted by humans. This data is analyzed to provide recommendations for ways to measure utilization and how it can be used to predict failures. Data analytics provide justification for changes in policy and procedure to reduce costs associated with simulator maintenance.