I guess it depends on which stage of your website development you are in. When I'm bootstrapping it and I'm not even sure whether the idea will be a success, I'm not willing to invest in anything more than good enough. When I have some empirical evidence that the idea will fly (e.g. a/b tests, first sale closed), only then will I get in touch with a great designer to do custom work. As a plus, I will probably have discovered what I really want from the designer by that time and will be able to drive the design in a better way.
