It sounds to me like you want to combine conjoint share data with market size data to get volumetric estimates. Models for market sizing range from more theoretical to more empirical and from very simple to very complex.
Simpler models when you don't have a lot of history for your product include taking historical growth curves from like markets. Not all "like" markets behave alike, however, so using an average growth curve from several like markets is usually a better idea.
When you do have some years of history for your product category folks sometimes fit a simple s-shaped curve (e.g. a logistic curve) to their historical data and use that, along with some assumed maximum penetration that sets the top of the curve, to estimate the shape of the future portion of the curve. I have seen even this simplistic approach work extremely well.
As Bryan notes, diffusion of innovation models or "Bass models" are what you want to look for. Sage publications has a nice short book by Mahajan and Peterson called "Models for Innovation Diffusion."
