I have seen the real-time PPM graph. It's pretty freaky to watch. Interestingly, outside of stopsets driving folks off pretty reliably, tune-ins and tune-outs seem largely random. My show is on at night, so over half of my tuneouts are people just turning the radio off as they get home. But the people switching to another station are literally going all over the market, and they do it in random places, not necessarily at the beginning of a jock break or a song they don't like. What it demonstrates to me is that many listeners have a broader interest than we tend to think... they are not necessarily "rock people" or "country people" or "talk people".
I will agree that PPM and ratings are very important and deserve to be a prominent aspect of a radio station's strategy, but I don't think PPM is "everything". Any salesperson or programmer who says that ratings are "everything" is failing to grasp the big picture. A good radio station performs for its audience, and therefore its clients. In spite of the fact that PPM measures actual listening (or more accurately, actual hearing), the sample is still so small that it does not begin to approach a scientific demonstration of "truth". This is as true in the PPM world as it was with diaries. When Arbitron is floating 15-20 thousand meters in Austin, that's when the result gets closer to real meaning, imho.
