We send surveys after each performance with a random select and a survey exclusion segment in th extraction. We use a appeal and exclude anyone who has gotten the survey in the last five weeks. With that, a sub could get a survey more than once in a year, but at least it would be spread out.