I think the number of "Thumbs up" a map has become rather useless. Most of these votes are from a time when CTD had a bigger number of players.
Example: Flyingorb (68 thumbs up, 12 neutral, 42 thumbs down) has a total of 122 votes. It's a map that has existed back then when CTD had hundreds of players.
Iceflame, for example, has a total of 36 votes. Despite being several years old.
I think the ThumbsUp/ThumbsDown is not a meaningful metric on how popular a map is.
The number of how often a map has been played, however, is more relevant. But only for the most popular maps, like Richtungswechsel, 4stops, Quadcore, or Darkvector. The number of games speaks a clear language, those maps are very popular.
However, maps with a small number of games are harder to compare. The luck factor plays a bigger role here. If a map gets randomly selected at the random map mode just a few times more often than another, it doesn't say much about popularity.
Another thought: Once a game goes from fun to testing, is has pretty much zero chance to ever come back. (Unless a player wants it back really bad and decides to play it 20 times by himself.)
I dont think a test month would annoy the player base. Most good/regular players play Random: Tournament only, so they won't be affected anyway. And most of the amateur players probably won't even notice