The "millions" who listen, most of them are not satisfied with what they hear. They continue to listen, because they have no choice. They rather hear from media that's free. Same goes for people who insist to stand in line for very long periods to get what they want, but they hate standing in line or sitting in traffic.
People listen to the radio, because they are fixated, they're used to it, so they listen, like a never-ending habit. They are tired of many of the songs they hear but they leave the radio on, regardless.
If people heard something "new" they'd be curious and regain the interest. People like change.
People hate change.
One of the biggest causes of a successful radio station losing audience is a new PD who decides they have to "refresh" the station... which usually means changing a lot of details ranging from liners to even the music flow and blend. This is generally followed by a marked decrease in listening, both in TSL and cume.
Changes in a station that is doing well have to be casual and progressive, not abrupt and intrusive.
For example, when a station loses audience due to a change in a morning show, research with those listeners who go away reveals "it had changed so I took that as a moment to see what is on the other stations" to be among the common responses, followed by "I didn't like the new show as much as the old one and just kinda' quit listening".
In general, those who react to change say, "I loved it just the way it was. it was my favorite radio station, but it's just not the same any more". What comes to mind in those cases is "New Coke"; change, but not for the better.
People listen to radio because they like it. Particular stations reflect their taste and they enjoy the announcers and other things the station offers. Even too big a change in music, such as those occasional music tests that reveal shifts in feelings about many songs, can be disruptive and a good station phases the changes in by slowly replacing the lowest testing songs with a few it did not play before.
In marketing in general we see things such as new logos or changes in a brand's product to require extensive testing. Recently, a high-end laundry product changed its aroma thinking they were modernizing the smell and many people thought that the smell was "stronger" and quit buying it. A logo change can imply a change in the product itself, which may make consumers try competing products instead.
Of recent, studies have determined that one of the issues with getting a car buyer to change brands is that "that other car has so many features I don't understand but the new one of my current brand is like the old one, just improved". That is what happens when technology moves faster than the average car retention span. The result: many people keep their old car longer, and also tend to stay with the same brand, too.
And when you say that listeners are "tired of the same songs" you are missing the point of research: to find out which songs are "hits" today and to keep playing them while shedding songs that no longer test well and replacing them with ones that do. Since the library is not static since neither is the target age group members, there are always songs on the horizon that will replace ones that are no longer liked enough to play.
