Análise dos fatores preditivos de performance em projetos de inovação por meio de regressões múltiplas

Abstract

This study aims to quantify the capacity that criteria used during the selection and prioritization stage of innovation projects have in predicting the future performance of projects. For the development of the study, the existing works on the subject were analyzed, through a systematic review of the literature, which later supported the construction of the methodological procedures used. The predictive capacity was quantified first by means of simple regression models, and later by means of multiple regression models, which had as input data from twenty-one evaluations performed during the screening of new ideas, and as outputs data from the ideas’ performance evaluations, which were done after their development and subsequent market introduction. The screening data used fall into four categories: impact, viability, economic and risk. The performance metrics were developed and collected during the execution of this work, having been considered four different performance metrics: financial, knowledge acquisition, brand enhancement and ecosystem strengthening. The results obtained evidenced that, except for knowledge acquisition, the performance of an innovation project can be partially predicted through the criteria used during the screening stage. The study also found that only a minority of the criteria used is necessary for the construction of the multiple regression models with the best predictive capacity. Moreover, the study presents quantitative evidence of the relationship between the radicality of an idea and its expected performance.