5. Discussion
5.1. Interrelationship of dominant design, innovative performance, product innovation, and process innovation Our results indicate that innovative performance in a technological class is negatively influenced by dominant design in that class. In other words, there is a negative relationship between dominant design and the degree of innovation, i.e. innovative performance decreases because there is a dominant design in place. This supports the observation in Fig. 4, which indicates that classes which have a dominant design for more than seven years, have a very low patenting frequency. As Blind (2013) argues, dominant designs can have positive effects (e.g. increase of effi- ciency, helping to reach a critical mass for production), as well as negative effects (e.g. market concentration or the premature selection of technologies). Even though we use a patent-based measure for the dominant design, the existence of a dominant design is strongly influenced by market acceptance of the technology. For this, the case of HD-DVD vs. Blu-Ray may serve as a recent example, where not the best available technology won the battle for dominance, but the one with the more competitive design and stronger ties with strategic partners (Soh, 2010). However, although innovative performance decreases with the event of a dominant design, this behavior is cyclical as dominant designs emerge and disappear: In industries with dominant design there is still innovative activity, but the type of innovation changes temporarily. While overall innovation performance is negatively affected by dominant design, process innovation is positively influenced by the occurrence of a dominant design in a technological class. With this result, we can empirically support the basic assumption of Utterback and Abernathy (1975) that the rate of innovation changes over time from product to process innovation. Process innovation increases while the total innovation performance decreases when there is a dominant design. Across all classes, however, process innovation remains a minor share of total innovation, so that the intersection of product and process innovation shown in Fig. 1 is never reached. Still, it may be possible that the intersection will be reached in certain industries, so that process innovation becomes more frequent than product innovation. As prior literature suggests, the increase in process innovation may lead to a higher level of efficiency and a decrease in the number of competitors (Utterback and Abernathy, 1975; Suarez and Utterback, 1995).
