[noduerme]

This is a thoughtful answer. I'm usually asked to ballpark estimate a project before starting, and of course I've already sketched it and scoped it by that point. So I usually give an estimate that includes some cushion for areas that might present more difficulty, or if I see the client is a little fuzzy on some functionality that might need a deeper dive once we get there. Then my hours usually come in 15-20% lower than the estimate, if there aren't a lot of additional demands. And usually if there's something extra a client requires, I can think of a cheaper/faster way or a more polished/complete way to accomplish it. This makes clients happy and saves me time. But absolutely, I try to overestimate a bit so as to never blow past the client's budget.