IIRC the most valuable feature is the ability for the engraver to quickly and deterministically escape the set of behaviors that the software provides to ease the act of engraving exactly what they want to see.
Score and Finale have been the gold standards because there is a way to escape the default behavior, to draw whatever they want on the score. In Finale's case, these ways are all explicitly documented so that even a relatively new user can (slowly) find each escape valve. Finale even has a shitty, black-and-white Microsoft paint for the user to paint whatever articulation, shape, or even notehead they need to accommodate whatever notational style or concept is needed.
That's all to say-- the power user is going to live in the escape valves for a substantial amount of time. (How else could you explain people still using Score when Finale had been around for decades?)
Historically, competitors to Finale/Score simply did not have the adequate number of escape valves. In those cases good interface design was irrelevant because professional engravers couldn't get the control they needed to engrave scores, which meant they couldn't use those programs to make money. Either that, or the software would do a really good job at, say, jazz engraving, and then imply that it was good at all style of notation when it wasn't.
Somebody at Musescore should do a kickstarter to cross-reference the current Finale documentation against the escape-valves present in Musescore. In the domain of engraving, is Musescore feature complete with all of Finale's escape valves? If so, it's worth advertising that because it's a really important achievement.
(Note: I'm reticent to mention specific engraving problems for fear that a neophyte will take my 10 seconds of examples as a complete set and declare Musescore feature complete. But if someone wants examples and promises not to do that, I'll give them.)
Edit: clarification