Sidney Dekker's _The Field Guide to Understanding 'Human Error'_ has a bit about this (which he calls drift). One thing I liked in that section is a sort of anti-Murphy's Law: What can go wrong will usually go right, and then we'll assume that it will go right again and again, even if we borrow more from our safety margin.
That said, it's a trade-off, and if you're a startup, it's usually far better to do it dirty today than perfect next year.
That said, it's a trade-off, and if you're a startup, it's usually far better to do it dirty today than perfect next year.