This paper outlines an application of belief functions to forecasting the demand for a new service in a new category, based on new technology. Forecasting demand for a new product or service is always difficult. It is more so when the product category itself is new, and so unfamiliar to potential consumers, and the quality of service of the product is dependent upon a new technology whose actual performance quality is not known in advance. In such a situation, market research is often unreliable, and so the beliefs of key stakeholders regarding the true values of underlying variables typically vary considerably. Belief functions provide a means of representing and combining these varied beliefs which is more expressive than traditional point probability estimates.