Identifying determinants of broadband diffusion by stage and by service

This study, based on a Gompertz diffusion model, examines the determinants of diffusion for total broadband, DSL, and cable modem services by applying the Arellano-Bond GMM dynamic panel estimation with the data for OECD countries from 1997 to 2009. For an overall diffusion, the determinants of the services are commonly the same. This study further examines the determinants in three diffusion stages segmented based on the Innovation Diffusion Theory proposed by Rogers. For the same service, the estimation results indicate that the determinants differ by diffusion stage. The reason is very possibly due to the different adopter attributes in each stage. Except for the early majority stage, the determinants in each stage are also different between DSL, cable, and total broadband services. Therefore, governments or service providers should tailor appropriate policies or strategies by stage and by type of service instead of adopting one-sizes-fit-all strategies.