We examine the theoretical underpinnings of the Bass innovation diffusion model and show that, contrary to the present understanding in the literature, it does not describe the behavioural theory of an innovatorimitator dichotomy within the adopting population as was intended. Rather, this model describes the behaviour of a homogeneous population in which it is assumed that product adoption occurs as a result of exposure to a news source or from the binary interaction between adopters and nonadopters. We provide an alternative framework in which the innovator-imitator explanation of diffusion is explicitly incorporated. In the so-called two-compartment model the innovators and imitators constitute two distinct but interacting subpopulations. However, the added structural complexity cannot be supported by the available diffusion data, resulting in unsatisfactory model estimation results. This suggests that the innovator-imitator population dichotomy assumed to underlie the present understanding of the innovation diffusion process may be inappropriate. Replacing it with the assumption of a homogeneous population could have potentially important implications for new product advertising strategies.