Heterogeneity in adoption – A critical review of theory

Abstract

This paper compares the NBD model with the Bass model as a means of predicting new product diffusion. A previous ANZMAC paper of ours revealed some strengths of the Bass model in this application. This paper reveals a particular strength of the NBD, namely its ability to cope with heterogeneity in individual’s adoption rates. The two models were originally conceived to function in different environments and neither should be considered a special case of the other. In this paper, we are comparing both models in static markets condition and the role of heterogeneity in predicting the rate of adoption. Both the Bass model and NBD, on the surface, in static markets, appear to generate the same result but they do not. Colombo (2003) and Habel et al. (2003) both suggested that penetration curve of the Bass model approximates the curve of the NBD model under the condition where the innovation parameter (p) is high and the imitation parameter (q) is low or equal to zero. The NBD model can generate results which Bass can not because the models differ in their treatment of heterogeneity. Heterogeneity accounts for the difference among a population in their propensity to purchase and that eventually affects the speed of adoption. The conditions under which this occurs are not exceptions. In practice, this anomaly may not be so obvious because the Bass model is usually applied to dynamic markets. Nevertheless, the model has a substantial limitation in not considering heterogeneity.