Marketing-Induced Fluctuations and Baseline Shifts in Product Adoption: A Change-Point Diffusion Analysis

A major chunk of historical innovation diffusion models resembles a smooth S-shaped curve. In real life, this rate of diffusion keeps fluctuating due to various internal and external factors. Although studies have been proposed to incorporate the impact of measurable factors like price, promotion, and advertising-related variables, the timing process related to innovation adoption still continues to exhibit dynamic behavior. Hence, we have developed a diffusion model that is able to capture the non-stationarity of the diffusion process with respect to the time points at which a behavioral change in the diffusion process is observed, studied as the changepoint concept. The model is applied to the datasets of breakthrough innovations such as television, refrigerators, and clothes-drying machines. Incorporating changepoints into the generalized diffusion modeling framework markedly enhanced model performance, reducing mean squared errors by 58-99% and improving R² values notably, thereby substantially increasing estimation accuracy.