By conducting simulations, we show that the structure of the personal networks significantly influences the diffusion processes in network effect markets like the software market. Varying connectivity (number of links to other participants), the heterogeneity of preferences, and the price, we analyze networks with two different topologies. The first topology is determined by linking consumers with their closest neighbors ("close topology"). For the second topology the links are determined randomly (" random topology"). For both topologies, the concentration of market share strongly correlates with the connectivity. The "random topology" leads to total market concentration even with very small connectivity. In contrast, diffusion within networks with close topology results in larger diversity of products; here, we find total market concentration only for very high connectivity.