Remember that time when Metrics told Austin that Mercs/Mando's were fine and then, like months later, Metrics was all like "oops, I counted trash mobs as part of the Mando/Merc DPS #'s in Ops lol"...I bet they have a good laugh over that still. That was funny.
And for folks mocking the metrics, the above passage is the key. Using metrics is much better than not using metrics but the system is still vulnerable to garbage in -> garbage out.
E.g. for PvP, how about looking at class population and performance for rated warzones? How about for the top rated teams? If less than 1% of competitive matches between elite teams has a commando in it, that is an interesting metric -- it tells you that the best players in the game have decided that the commando is crap for competitive WZ.
I know BW likes to keep their cards close to their chest but they could do some open ended discussions and surveys where they ask players "what metrics do you think would be most valuable for balancing XYZ"?
Metrics are invaluable but you have to look at the right thing. This includes looking at different parts of the player population -- most notably looking at how classes play in the hands of the best players.