Have you already considered a survey? I won't bore you with my suggested questions, but besides the basic PDF or print question you could ask people about their propensity to buy both if they have a significant discount (which involves even more number crunching). Including questions of past purchases may also provide you with a mild measure of the quality of the answer. Combining the last with a reward (screen saver?) for respondent if they use their GiTP account, then you could get a good measure of quality in your data. I said I won't go much into suggested questions because surveys seem to open more questions than they answer.
I know, surveys also mean work, but this one could be a worthwhile investment in risk-reduction.