There is a good piece on "hook testing" by a major market programmer in this weeks Gavin.
He states something that has been a given in every format but Smooth Jazz for quite a while: that a song needs at least 100-150 spins before it is "testable" because until a degree of familiarity is established respondents won't be able to recognize the hook (that 10-15 second clip) well enough to give a valid response. They further state that the first music tests should not be used as a sole criteria for playlist decisions because familiarity and acceptance scores may be deceptively low. So what does this tell you about hook testing new releases or music that is new to the audience in a market that has a Smooth Jazz startup? Check out the whole piece, it's pretty short and concise: http://www.gavin.com