It was a colossal failure because it failed at a colossal level.
It promised to deliver an undeliverable promise, which was it was going to use human and computer polls to match the “top two teams” in a championship game.
The very concept of that is absurd and for reasons already discussed here: it’s impossible to do such a thing using such tools for a league of that size.
Hell, the BCS would be inadequate for the NFL, and you think it was appropriate for the FBS?
Human polls are worthless. They are a consensus of arguments from imagined authority. Human polls are assumed improbable claims from credulous sources based on insufficient evidence, which by definition renders anyone who takes them seriously a fool.
Computer polls have the same problem: the assign arbitrary weights to criterion that cannot possibly be measured accurately given the size of the league coupled with the limited schedule. It’d be like if you went to Port Aransas beach and someone asked you “which grain of sand do you think is the biggest?” I mean I suppose there’s a definite answer to the question but discovering it would be an exercise in futility on a practical level.
So you’ve got two rankings systems with absolutely horrid flaws and defects, and we’re going to mix them together to come up with two fucking teams in a league of 130? That’s not a goal; that’s a fantasy. It’s a particularly stupid fantasy given that the template for a far more suitable alternative has been available now for literally 100 years.
And by the way, they didn’t expand the BCS, either. As I said, the BCS is a compendium of polls derived by coaches who don’t pay attention to anything, bad comedy writers in sports media and computer nerds. The CFP is a selection committee of people associated with the P5. It’s also a stupid concept, but it’s also different.