Exploring the Relevance of Two-Part Models in Innovation Research: Towards a Better Understanding of Innovation Sales

Abstract

In empirical studies, firms’ innovation is often measured as the proportion of sales that can be attributed to new products or services. In a review of 20 years of research on innovation, we find this variable to be used in 141 studies providing a useful common ground to compare and contrast findings. We argue that two-part models are attractive when modelling innovation as they separate the decision to innovate from the amount of sales generated by innovation. Our review shows that only a minority of studies employ two-part models, most often in sub-optimal ways. We develop a two-part modelling framework that is suitable for innovation research and different from the conventional methods prevalent in econometrics. Furthermore, we provide a worked example of how it can be implemented. This paper contributes new knowledge about two-part models for innovation research and hopes to lay the ground for new theorising in the future.