Every project should have a go / no go decision right at the start, that includes at the very least the following criteria:

Do it right, or don’t do it at all.

If you don’t know what you want, then don’t expect anyone else to either.

Which leads back to the first point.