A theoretical and empirical comparison of innovation diffusion models applying data from the software industry

This paper presents a theoretical and empirical comparison of quantitative innovation diffusion models with data from the software industry. Looking at some of the most widely used aggregate models, such as the Generalised Bass (1994) or the Kalish (1985) model, the inclusion of decision variables is compared to scientific groundings from innovation marketing, focusing on the analysis of their mathematical equations. Eligible attributes such as the reduction and the carry-through effect are addressed and performance simulations of the mapping function are run, using commonly observed pricing strategies. Adoption and marketing data of two innovative software products are applied to calibrate the models and to evaluate their forecasting precision by comparing the results with true data within a period of ten month after initial release. It becomes evident that the predominant S-shaped models are not always suitable. A market analysis shows the complexity of market structures and the involved requirements for these models. Critical branch related aspects, such as the reputation of the company, online distribution channels, and piracy are discussed and disclose future research spots in the estimation of diffusion shapes of innovative products.