We did an audience survey a couple of years ago and got a lot of useful information from it.  It was just a page inserted in the program and during the curtain speech before the show they mentioned the survey and that people could take the pens in the lobby as a thank you gift.  Response rate was a very good, 30% of the audience turned in the surveys.  We found out which of our advertising methods was reaching people and which weren't.  As a result we dropped some places and placed a lot more emphasis on others.  I use the stats from that survey all the time in speaking gigs about the theatre and on promo pieces we do.  For example we learned that 42% of our audience comes from out of town to see our shows (a great economic impact stat to toss out).

The important thing with those types of surveys is to keep them short and to the point.  Multiple choice, scale of 1 to 10, or yes/no questions only.

Try to avoid open ended questions like "what do you like about our theatre?" and leaving a blank space for a response.  Most people won't fill out those kind of questions. They want to be able to quickly answer the questions and move on.  Also with open ended questions it is harder to accurately evaluate the results.  What you may think is a clear cut programing question, someone else may take as the opportunity to go on a rant about a rude usher two years ago.

You can include a question at the end of a survey that's simply, "Any other comments you wish to share with us?"  Even then be careful when putting the results together because you'll get some response that are way out of left field.

Just remember the key is keeping the survey short and asking specific questions you want to learn about.

Every few years we also try and get together several currently active and previously active volunteers/actors together to have an open brainstorming session.  These can be very successful, but all depends on who attends.