Diffusion processes in markets for network effect goods - Determinants, simulation model, and market classification

While IT markets become critically important for the world economy, existin g approaches are insufficient to model their specific properties like posit ive network effects. Economic theory of positive network externalities is f ocussing rather on the installed base of a given product than on structural properties of the personal network which influences the individuals' decis ions, geographical and sociological network analysis covers many structural properties bur does not adequately model the dynamics of diffusion process es itself when strong externalities exist. Our paper integrates both approa ches into a simulation model of the actual diffusion process and identifies determinants predicting its result. While heterogeneity of preferences, hi gh product prices and a decentralized, regional or sparse structure of the network prevent concentration, homogeneous preferences, low prices, high co nnectivity, a random "global" topology or a centralized structure of the ne twork promote concentration cowards a single product. The simulation result s are substantiated by an empirical analysis of the German and US markers f or EDI solutions, ERP systems, and office communication software.