Holy crap Bolt. We actually agree again! Research killed radio. The entire industry.
Then you are both wrong.
Research lets good PDs be better.
One time I did an amateur music test for an AC station that had CHR formatics and I had, for fun, jotted down my score to about the first 100 songs out of 600 in the test.
Now, that station was #1 in 18-49 women and #2 in men. It was highly successful. The test was not to fix anything... it was to stay ahead. But we thought we were so good that it hurt!
To my surprise and major embarrassment, my scores were off by more than 25% on over 2/3 of the songs. There were a lot of songs I liked that the audience hated. And more that I enjoyed that did not make the listener's collective passing grade. With the test's results, we became better and stronger. And we learned what kinds of songs were just poison.
Later, with a low rated (last place FM in what is a top 20 market) station in a rhythmic format, we did a low budget test and cut 900 songs off the playlist and reduced currents form 70 to about 20. Next book: #1.
Another example: I was on the team at KLVE in LA in'95. With a test and call out, went from about 15th to #1. And it was very #1, not just a close win. Knowing what not to play, the list went from nearly 1000 songs to 300. And the station stayed #1 until our own second FM, KSCA; knocked us to #2 with the AC format.
And the best example of building one of America's great stations around more research than any other station has ever done was Jerry Lee's FM in Philadelphia. Music research. Perceptual research. Commercial spot research. And the moment the station was sold, they cut the research and lost half the audience. Brilliant.
But research is only as good as those who interpret and implement it. In many cases, all a music test does is the equivalent of giving a machine gun to a chimpanzee and then expecting something good to happen.
The Edsel: unresearched product.
The Mustang: highly researched product.
Same company, different decades.