Suppose you want to know satisfaction of a large population about a service (for instance 20,000 subscribers to a newsletter). You don't need to survey the whole population. Depending on involvement's level of the population, you will expect from 10% to 40% of completed answers. So to get some 1,000 respondents, in the worst case, you don't need to question the whole population: you will extract a sample, which is representative of the whole population on one or two criterions (company's size, localization, age...). By now you have to extract and manage samples outside QuestionPro, and then to inject them in dedicated mailing-lists. My suggestion is to offer the possibility of injecting the whole population with criterions (as custom variables) and of extracting samples automatically (with marking of the people). By this way, it would be easier to launch the same survey or different surveys toward different samples.
Idea No. 6