Eric Flint, famously, argued that sales, or more subtly, as I understand it, that persistent sales over time combined with "sell-through" are the primary market indicators of quality. And wouldn't we all agree that the market is the best determinant of quality...?
Eric sort of has a point, though I tend to feel that goodness is quite orthogonal to sales. A lot of fluff is sold in large amounts, while some indisputably good books are not exactly best sellers (eg Knuth's Art of Computer Programming, which is almost universally considered to be both good and influential, has amazon sales rank of about 10,000 - low, though higher than I expected).
It also begs the question of what is "good". Good for what?
And influential is also orthogonal to "good". For example "Mein Kampf" sold well, was very influential, but was not good.
Ultimately this is a matter of taste. If you think a book is good, then it is good. For you. This is also a time dependent issue - what you think is good when young may seem shallow or trite when you are older; conversely books thought bad when young may seem good and influential later. Annoying that.
Finally, "good for what"? Knuth's book is good and important, but I wouldn't take it to bed for casual reading, most nights. Larry Bond's "Enemy Within" is good for casual reading, and today is more important than when it was written, but is not absolutely important, nor is it exactly Nobel Prize material. Paddington Bear is good, important and sells well, but again it is of limited appeal to me at most times, despite rating high in all three criteria. Conversely, "Independent People" is good, influential, won the Nobel Prize, and has amazon sales rank of about 110,000. Not a best seller (in the US), but it does make good bed time reading.
However, probably the least reliable metric of the goodness or importance of a book is what the author thinks of it.