First, let's talk about optimization: we can compare it to finding the shortest road. An optimization problem is like finding the shortest way from school to home - finding any way that will get you home is the solution to the problem, but only finding the shortest way is the optimal solution.
A greedy algorithm looks only locally - you know the shortest way through the park, so you take the shortcut rather than going through the main park avenue. This often actually gives you the best solution - if your home is behind the park.
but if your home is behind the part, but a little bit to the left, you might be better off skipping the part to the left already - and this is what a greedy algorithm cannot help you with as it only looks locally.
So, depending on the problem it can work well or not-so-well.
