Concept for development project management by aid of predictive analytics

Manufacturing companies in high wage countries strive towards shortened development and innovation cycles at decreased costs in order to strengthen their competitive advantage. These goals can be achieved by efficient development projects. However, approaches aiming at designing efficient development processes such as the value stream analysis only analyze development projects retrospectively as well as periodically and therefore do not continuously improve the efficiency of the respective projects themselves. Therefore, a concept is needed to anticipate deviations from the target process and thus inefficiencies within development projects by aid of predictive analytics. To derive a predictive analytics model, neural networks are applied to identify the impact of deviation indicators on the efficiency dimensions time, costs and quality of an activity. Upon reversion, it is possible to monitor the deviation indicators and use the respective indicator values as input for the neural networks. Based on the identified impact of the indicator on the efficiency dimensions, the neural network is able to predict the final values of an activity in terms of time, cost and quality. By comparing the predicted values with the defined target values, the deviation can be determined and preventive measures can be implemented to eliminate inefficiencies.