Abstract
This article proposes a method that overcomes a number of problems associated with new product diffusion models noted in the marketing literature. We illustrate the methodology in the context of better understanding global variances in new product adoption. Building on existing diffusion models and sample matching principles from international consumer research, we suggest a "staged estimation procedure." The procedure provides both sensible and robust estimates and remains usable even if the diffusion process is in its earliest stage in most or all countries. In an empirical illustration covering 184 countries on five continents, we use cellular diffusion data to gain insights on how exogenous/endogenous country characteristics affect country-level diffusion patterns.