I think the thing is that it's very rare for an idea to be directly translated into a product in it's original form. There are changes and refinements along the way, and it's not unusual for the idea to be changed quite a bit from the original thought during the process, especially when iterative development techniques are used.
Rather than starting out as definitely good, definitely bad, or definitely average, most ideas are potentially good, and what happens to them after conception can result in either an excellent, a good-but-not-exceptional, or a bad product.
I also think this concept of a unique, non-derivative idea is an exceptionally rare thing, or perhaps something that doesn't really exist at all. Creators may not always intentionally use a derivative process to create their ideas, but once you can see the idea you can generally always describe it as "this existing idea, but with these changes", or "this existing idea, without this", or perhaps "these two existing ideas combined". Building on previous work is a good thing, and whilst it's also good to break the mould a bit with something new and less derivative, the foundations in existing products and ideas is almost always there.