From my understanding of personas, one of the main purposes is to find out how your users behave and what their mental models of the system might be. I had always thought that, like in usability testing, tapping people that already have a knowledge of your system will introduce unpredictable biases that are difficult to control for. I'm sure we all know the axiom that people can't really tell you what they want. I have heard that people do build personas off of people that currently use their system, however.
So my question is two-fold (and a good answer will be as well):
- Is doing persona research on potential users good practice?
- If yes, how do the research and creation methodologies differ? If no, what would you consider the best way of getting high-quality, accurate feedback from current users?
(Note: I know that the answer won't be cut-and-dry. UX work never is. :) )