At the risk of trying to walk the middle of the road [invoking Karate Kid "squish just like a grape" imagery]...
My (obviously amateur, obviously biased) observation is that a really good speaker must do two things: First, it must do something so well as to enable the listener to suspend disbelief and get lost in the music. That "something" can be natural timbre, dynamics, inner detail, imaging, coherence, sense of scale, sense of immersion, whatever. But a really good speaker will do
something (and ideally multiple
somethings) well enough that the listener is "mesmerized and [can't] wait to hear the next track". This is the easy part.
The hard part is, the speaker must not then do anything so poorly that it distracts from the illusion just created. Great long-term enjoyment and satisfaction is reserved for "well designed, competent" loudspeakers which are free from audibly significant colorations and anomalies.
@Floyd Toole said it best:
"The highest rated loudspeaker is the least flawed, not the most virtuous."
But imo a really good speaker will have at least one mesmerizing virtue.