I wouldn't fully trust either reviewers or users, as both have their own pitfalls. Then again, I think scores are by and large useless and mostly serve to validate people's biases and opinions, but that's another discussion.
All this being said, Metacritic user scores are indeed some of the most useless. They often get bombed, then fanbois counter-bomb so any game that's controversial hovers around 5 because, naturally, it's the average between the 10s given by blind fans (and Andromeda is no 10, much as I personally like it) and the 0s given for all manner of reasons (and the game is no 0, much as some might hate its very existence).
Inquisition is also a good example. Poor showing on Metacritic user scores, but when it comes to player-granted GOTY awards it ran away with them in its year. Yes, yes, it wasn't a year with too many amazing AAA games, but that's still a significant disconnect.
Andromeda's obviously not going to repeat this feat, even if it didn't have animation and dialog problems it wouldn't stand a chance against Breath of the Wild. But I doubt the general public would really rate it so low.
In the end, scores and manufactured internet outrage aren't worthy of the time I sometimes put in them. The best way to judge a game is to see it played for yourself, by people who have no affiliations to the developers. Not official videos, no IGN-esque doctored reviews, and no judging a game based on 4chan gifs and memes. As much raw and uncut gameplay footage as I can find is my policy, hasn't failed me in a long time, and it hasn't failed me for Andromeda either. It's the one thing I actually like about Twitch.