Information Technology Reference
In-Depth Information
patience, price sensitivity and needs; among those, heterogeneity in the propensity to
adopt is a common approach to incorporate the heterogeneity of individuals [11].
It is supposed these differences allow the product takes place over a market, and
their distribution in a specific social system determines the shape of the pattern of
diffusion [10, 14].
3
The Double-S Phenomenon
Commonly, diffusion models predict a monotonic increase in sales after the launching
of product and up to the peak of growth. Despite that, evidence shows that in some
markets, a sudden decrease in sales may follow an initial rise. This special point in
diffusion curves is referred as a slowdown phenomenon [19] or “saddle” point
(concept proposed by Goldenberg, Libai and Muller [4].
After a literature review, Tellis and Chandrasekaran [5] point out three possible
explanations for this phenomenon, two of which can be considered as external
explanations and one of them as an internal one: economic contraction and important
technological advance (external) and discontinuity in the transition between the early
and late markets (internal). The authors find, after using a discrete-time split
population survival model for several products in several countries, that all the
reasons influence the S-shape phenomenon in some extent.
In a previous work, Golder and Tellis [20] argue that this phenomenon is due to
informational cascades, which happen when people collectively adopt a behavior with
increasing momentum, and declining individual evaluation of the merits of such
behavior after, because of their tendency to infer information from the behavior of
prior adopters.
Nonetheless, although their argument lies in individual behavior, they propose a
hazard model to determine the impact of explanatory variables such as price declines,
income declines, and market penetration on the time to saddle point; whereby their
hypothesis is not completely tested at the individual level (as in the Tellis and
Chandrasekaran [5] study).
At individual level, Goldenberg, Libai, and Muller [4] use cellular automata to
describe the process by which internal communication breaks down between the early
adopters and early majority, which could lead to a saddle point in the diffusion curve.
Nevertheless, cellular automata models do not allow modeling socioeconomic
characteristics of population, so conclusions from them are very limited.
Peres et al. [10] point out that one of the further research line on innovation
diffusion is still to explain saddle formation and incorporate the findings into the
diffusion framework.
4
Methodology
In order to analyze the emergence of the double-S phenomenon in the innovation
diffusion process, it was developed an agent based simulation model, which
represents the diffusion of two identical innovations in a market. The model was
