Oh gawd, you have no idea how many decisions we have to make--and almost all are tradeoffs. Go down the rabbit hole of "everything has to be perfect" and you end up with an amp like the Ragnarok--huge, hot, expensive, complex, and years late. Not that there's anything wrong with that...but it's the reality. (For the record, I love Ragnarok--it is every bit as insane as Yggdrasil on the digital side.)
As a for-example, there was a new product I was working on over the summer, with an exciting new topology (but not as exciting as what's coming in 2016). There were at least a dozen major design decisions that would affect its performance. The temptation, of course, is to take the high road in all cases. But when you're looking at a (very) cost-constrained product, with some serious limitations on space and heat, there are tradeoffs that have to be made. I took the high road in a couple of those cases, and ended up with a fairly insane level of performance for the product. Could it have been even better? Yes, at 3x the cost and in 6x the space. But that wasn't in the cards.
(And, I think you'll see that new product before the end if the year)
So...you make your choices...and hope, as you say, that more of them are the better ones and less are the worse.