This is a good idea, but here again the survey questions would be set up by the current set of volunteers. And the surveyees would only give output to these select questions, not input from their own experiences. So the two approaches can be useful, but I still privilege a minimal intervention from the organisers to let the community develop spontaneously.