Calculating return on investment of training using process variation

Organizations have relied on training to increase the performance of their workforce. Also, software process improvement models suggest that training is an effective tool for institutionalizing a development process. Training evaluation becomes important for understanding the improvements resulting from the investments in training. Like other production process, the software development process is subject to natural and special causes of variation, and process improvement models recommend its statistical management. Return on investment (ROI) has already been proposed as an effective measure to evaluate training interventions. Nevertheless, when applying ROI in production environments, practitioners have not taken into consideration the effects of variation in production processes. This paper presents a method for calculating ROI that considers process variation; we argue that ROI results should be understood in accordance to statistical management guidance. The proposed method has been piloted at a software factory. The results of the case study are reported. These results show how to calculate ROI by taking into account the variation of a production process.