Why was this an issue in the first place? Are there any meaningful metrics or data showing 3 minutes to be a problem at all?
I prefer 3 minutes and agree with similar opinions reasoning above. While I rarely consume the full 3 minutes I often find myself needing to go over the 2 minute mark, even if slightly, especially in those delicate moves that often happen.
Capping it to 2 minutes will probably mean I stop playing CCL.
If the issue that needs addressing is AFK then that should be handled in a different way, like accelerating timers after a certain cumulative time (maybe across a couple turns) with no detected client activity or similar.
I think I must insist that those proposing the change show some actual data set that demonstrates conclusively that there is a problem at all with 3 minutes in the first place. Otherwise the change would not have any actual basis to justify it, other than a straw poll with a few hundred votes, and not really very representative (also straw polls can be easily manipulated with multiple IP based votes etc).