Diffusion modelling and industry dynamics in mobile telecoms data services.

The market growth of content and data services in the mobile telecoms industry entails the development of complex industry networks or 'ecosystems'. This research stems from the emerging Location-Based Services (LBS) sector, which has suffered the slow-start dynamics that often belie optimistic forecasts of new technologies. An exploratory and generic industry-level model is developed using the system dynamics simulation modelling approach to explain the start-up problem. This serves as a dynamic theory for the emerging LBS industry market where service diffusion is based on the growth of the installed base of an embedded product. Supply-demand interactions occur through complementary bandwagon effects in which benefits accrue from services that complement the main product. Econometric estimation is compared with system dynamics calibration to test a range of diffusion models on historic data, revealing some evidence of bandwagon effects. The model is proposed as a generic structure for market growth in mobile data services in general and as a tentative theory to explain the phenomenon of technology 'hype' cycles. The model leads to broad policy proposals on how to alleviate the start-up problem. Launching products without attending to service quality can mean that consumers become disillusioned whereas ensuring a high quality user experience requires a collaborative and long-term view of ecosystem development. This implies that major players such as handset manufacturers or mobile operators must play a central co-ordinating role. Validation of exploratory models of emerging markets is based more in terms of plausibility and usefulness, given the lack of historic data. A validation framework in the three dimensions of model content, process and outcome provides a holistic and semi-quantitative profile of the utility of generic models. A validation profile is proposed that can be used in a formative role to clarify model purpose and aid planning in model-based interventions